Affiliate Marketing Undervalues The Link

Two days I posted about Gretchen Rubin's new book, The Happiness Project. When I linked to the book in the original post (and here too), I included an amazon affiliate code.

And here is the data two days later:

Happiness project

So, in two days, that blog post generated 535 views of the Amazon page and 40 purchases. The affiliate fees associated with those 40 purchases add up to $6.50.

But those 535 views are pretty valuable. Those 535 clicks translated into a total of 118 orders in the past two days, including a Kindle. The total affiliate fees associated with those 535 clicks were $25.20.

But even including all the commerce that was generated from that link, that $25.20 is a cost per click of roughly 5 cents. I think that's low for a bunch of reasons.

First, let's take the 535 views of Gretchen's book. Yes, only 40 of them actually bought it right then and there. But surely more of them will convert over time into buyers. Maybe when they see Gretchen on TV. Maybe when they get hooked on her blog. Maybe when they see the book on Amazon's bestseller list.

And of course, the same is true of all the other things those 535 visitors/visits saw and did on Amazon.

comScore once did a panel-based survey of people who saw a banner ad. Very few of them actually clicked on the banner ad and transacted. But many who saw the banner ad eventually searched on the item they initially saw in the banner and transacted later. comScore has also observed that many products that are initially found and/or researched online end up being purchased offline. I wish I could find both pieces of comScore research. If I can find them, I'll come back and link to both (got one of them now).

The point is that my blog post drove a lot of value to Amazon that is not totally captured by the 40 purchases of Gretchen's book or even the 118 transactions that were done by those visitors in the past two days. The value of that link, in my opinion, is significantly greater than $25.20 and as a result bloggers and other users of affiliate services are getting under compensated for the value they are providing.

UPDATE: Gian Fulgoni, co-founder and Chairman of comScore left a great comment which you can see if you click on the comment link and scroll down to the end. Here is part of it where he shares links to the research I cited above:

Fred, you're correct that past comScore research has shown that search and display ads cause latent buying and offline buying that are not reflected in the click:

Essentially, it all comes back to the fact that the click does not reflect the "view thru" impact of ads (i.e. the impact of ads that are not clicked) — and that even clicks don't accurately measure all the latent buying that occurs (because of cookie deletion).


Comments (Archived):

  1. billerickson

    After I clicked through, I added it to my wish list. Every few months I go through and buy about 10 books off my wish list. Does amazon give you credit for that purchase?

    1. fredwilson

      good question. i don’t know the answer.

      1. Ed

        No, they will not pay on those old return visits. They’re actually one of the stingiest about cookie duration.

      2. AJ Kohn

        No, you will not get credit for those purchases. Amazon has one of the worst affiliate programs. A person clicking on your link must purchase that product in the same SESSION for you to get credit for the sale. Other affiliate programs provide you a 30 to 45 day window.

      3. chachra

        No more credit once 24 hours have passed OR purchase occurred OR another affiliate link was clicked. The first 2 are not standard with other affiliate programs. Session lasts for 15-60 days … no clauses. 3rd bit is pretty standard, since that makes assigning credit easier.From Amazon’s operating agreement:”The session begins when the customer clicks through a Special Link on your site to the Amazon Site and ends upon the first to occur of the following events: (a) 24 hours elapses from the customer’s initial click-through, (b) the customer places an order for a Product that is not a Digital Product, or (c) the customer follows a third party link to the Amazon Site that is formatted with an Associate’s tag.”

    2. Geoff

      I did the same – maybe Amazon should pay Fred x cents for every wish list its added too!

    3. David Noël

      I did something similar: clicked-through to Amazon US to check it out but needed to add it to my wishlist in the German Amazon and just ordered it today.I suspect I wasn’t the only non-US user clicking through and eventually buying. If I recall it right, Fred gives all affiliate proceeds to charity and I’d love to have contributed.Affiliate programs should support that.

  2. Driftaway Coffee

    so this goes back to brand awareness increased through your blog post. it’s the holy grail of marketing measurement. nobody can algorithmically figure out it’s value. agree with your argument though…

    1. fredwilson

      that was the underlying reason for both sets of comScore research i cited

    2. Darren Herman

      not to be snippy but it’s not awareness – it’s purchase intent in this case.

  3. Darren Herman

    This was my original response to your post:There is where attribution comes in – both post click and post view. Most marketers assign a time restraint (i.e. 1 day, 14 day, 30 day) and will count all orders/sales within a time period based on if someone “clicked thru” or “viewed thru” a link. So being that you are only 2 days into the link to Amazon, maybe check back in 30 days and see what your total volume is.Then I re-read your post and see that I missed the point you were trying to make.Ideally, you as Fred Wilson should be getting compensated beyond your initial book you are trying to sell for Gretchen by Amazon. You are providing Amazon volume b/c you are bringing them engaged traffic. This *should* be measured in the amount of sales volume in your affiliate window.

    1. fredwilson

      actually i am getting compensated for all the transactions the people who came via my link do. but as Bill points out above, some of those transactions don’t happen in session

      1. Darren Herman

        Did some digging:We will pay you (in accordance with Sections 5 and 6 below) referral fees on certain Product sales to third parties. For a Product sale to be eligible to earn a referral fee, the customer must click-through a Special Link on your site to the Amazon Site and during a single session, add the Product to his or her shopping cart, purchase the Product with digital fulfillment (e.g., digital download or streaming) in the case of Digital Products, or purchase the Product via our 1-Click feature. The session begins when the customer clicks through a Special Link on your site to the Amazon Site and ends upon the first to occur of the following events: (a) 24 hours elapses from the customer’s initial click-through, (b) the customer places an order for a Product that is not a Digital Product, or (c) the customer follows a third party link to the Amazon Site that is formatted with an Associate’s tag. We will only pay referral fees on eligible Products after order, payment, and shipping or, in the case of Digital Products, digital fulfillment have occurred.Didn’t realize that Amazon only compensates you based on single-session Click. Ouch. Yes, you have a very strong point you are making. You might want to put the above paragraph in your blog post so people can see the agreement.

        1. fredwilson

          good suggestion

  4. Ed

    I agree completely, especially in the case of Amazon.I send them a lot of traffic with thoughtful tweets (good for author exposure too).Yet the amount they’ve paid me in return is a pittance. In particular, they will not pay an affiliate for sales of Kindle selections.I am dumbfounded by this, because their cost is extraordinarily low.But they’re shooting themselves in the foot; the motivation for me [and others] to continue gifting them real estate, is waning.

  5. Adrian Bye

    you’ve hit on a very big topic here. yes, thats absolutely true, and amazon is underpaying its affiliates. i’m sure they’d like to pay more but when fraud is taken into account this may be the best they can do.other affiliate programs will pay out orders of magnitude higher than amazon.p.s. i bought the kindle version of the book 🙂

  6. javery

    The problem is that while I agree it is under-valued Amazon holds all the cards on this. Sure, you could decide its not worth your time to use an affiliate link but then Amazon is really just making more money than before. The problem is that Amazon has become the preferred place for so many people to buy books that they are still the best option for affiliate marketers. If affiliates linked to BN instead of Amazon most people would still buy it from Amazon and the affiliate would make even less.In this case I think the affiliate fee isn’t derived from the value that Amazon gets from the link, it’s based on the fact that even with that low value linking to Amazon will still make you more than linking to the same item on any other store on the web. There is no reason Amazon should pay more. It’s one of the many reasons I am long on Amazon.

    1. David Kadavy

      My thoughts exactly. They sort of have us all by the balls because readers will automatically go to Amazon.I’ve heard some people who sell print advertising say that if you were able to track it as CPA, the cost would be astronomically high; but, the companies that still use print are willing to pay more – by virtue of the fact that they *don’t* know how effective it is. Maybe some of that money will eventually erode to produce higher CPA’s on affiliate networks.

  7. jonathanmendez

    all web publishers are getting way under compensated for the value they are providing advertisers and merchants. it is their audience and their content generating the interest & intent. i’ve made the case here that they are getting only 10-20% of the true media value. we need new systems that remove the performance greed and will make things more equitable. we should look to search advertising for how to do it. without change to our shortsighted approaches not only will the publishers will lose in the long run but the web’s advertisers, merchants and audience will too.

    1. Darren Herman

      For those who do not know Jonathan Mendez, he’s one of the best out there. Good stuff Jon.

    2. fredwilson

      i just read your post. if publishers are only getting 10-20% of the realvalue of their media, there is a big opportunity to fix that

      1. LIAD

        big opportunity from publishers perspective but zero incentive from the merchants.anything they do to improve the end-to-end tracking and value capturing of affiliate schemes takes revenue directly out their pockets.they are happy with the status-quo and have no reason whatsoever to innovate/embrace new technologies which enable publishers to better track and get paid for the holistic value they the end of the day you would have published that link regardless of whether there was a chance of you earning $20 revenue from it. – thus the money you made was a nice bonus rather than the end in itselfon the whole i think that people are happy that the internet offers mechanisms to track and capture 20% of the value they create, the 80% value they create but dont get paid for is hypothetical from their perspective and what they never had they dont miss

        1. ShanaC

          Not necessarily:Marshall Clark on Chris’s Blog mentioned the idea on Chris’s Blog of the Panoramic View of the Web. In practice we do this in stores and malls when someone wants to measure how much is being bought. They flood a store and do a headcount of every possible area. This needs to be done now. If you have metrics on content creation, you know how valuable it is, because you can tell how many people are looking at that page. Even if no one immediately is buying anything, you also can pick up valuable statistics over time (especially if someone wants to start placing in eyetracking) why they are looking at what they are looking at versus not. That’s the online equivalent of flooding the store when it comes to ads. For merchants, knowing those statistics would allow them to float the rate paid. You have an awesome site that pushes a lot of stuff because of the way it is presented, guess what, you will make money because of a floating rate. Because the statistics are now floated and their are now usability guidelines to go with why those sites work (in comparison to those which do not), you could figure out what sort of real content does actually belong on a site if you want to affiliate link to say a coffee pot. Amazon in the end could provide helpful content adjustment advice for those who are not pushing enough stuff, much like McD’s franchises get their signs from headquarters. Win-Win for everyone. (I mean you can ignore the advice from Amazon, no one is going to make you make your website look a certain way because Amazon says so, there are just many more affliate links out there, and the spacial differences of a website versus a McD’s are very different, but the analogy holds)Further, we’ll never know fully what people will and won’t buy and when. It took me over three months of staring at this one dress to buy it (and I’m very happy with it, and I get lots of complements on said dress, I shouldn’t have waited the three months…) We have some understand of how lead generation works, about how people search and look at stuff on the internet, if you want to make more money on the content side, you should take a look at the material on the internet, about how people look at pages, search for material, and then buy, but not enough to accurately predict why I waited 3 months and someone else would have made the exact same purchase immediately.As the Great Retailer John Wannamaker says “Half the money I spend on advertising is wasted; the trouble is I don’t know which half.” Never forget that when designing a retail, marketing, and merchandising system, even an online one.

          1. Mark Essel

            “If you cannot measure it, you cannot improve it.”Totally concur with your opinion on improving knowledge of user behavior on the Net.

          2. awaldstein

            Ah…but in my mind what we are measuring has changed and that is the confused complexity of social metrics. Pre social web, analytic data was useable and justifiable to drive transactrions, likewise the ROI religion. Social data is now the analytics of communities.

          3. ShanaC

            I disagree with this, I think what is happening is that communities are now visible. You can now “metricize” them. (OK so what if I made up a word)It still doesn’t explain why people are where they are, why they like doing what they are doing, ect. And those are the critical questions of the funnel that everyone keeps talking about. Once you get at that why, you’ll get at that when, what and how. And that is a field that requires major study.Hugely investable field, but I would not want to go at it cold. Definitely something you want a team for, there are a lot of observations to be made, a lot of people to talk to, and things always change even as they stay the same (Nothing is ever new under the sun.)

          4. awaldstein

            Thnx Shana.We disagree on this one but whether they are new or now visible doesn’t impact the need understand social performance metrics. Business owners plan, execute, evaluate and reexecute. Having metrics, other than transactions is key.What excites me is that I don’t think they exist yet and as they get developed this will uncover not only keys to community building but also keys to the commerce models that will grow out of them.Thnx for the pushback.

          5. ShanaC

            Anytime. That’s what a good conversation is for. Part of what I am saying is a little impractical (perfectly standardized cookies for measurement purposes and eyetracking?) but hey, I can throw it out there.

          6. COMRADITY

            Totally agree with this.

          7. LIAD

            Great commentGreat quote

          8. ShanaC

            Thank you. I love that quote too. There is something so true about it…

          9. COMRADITY

            Importantly, in Wannamaker’s day, he only wasted 1/2 of his advertising. Today, the fragmentation of the audience by the abundance of media choices means that Wannamaker would be wasting more like 80% of his advertising.The math is pretty intuitive. In the days of mass media, it was possible to reach up to 90% of the population by “railroading” spots on all three networks in primetime. You still wasted 50% of the investment because a message can’t be relevant to everyone and distribution/out of stocks limit availability.Today, big box stores and technology have improved availability. But with so many choices, the audience is fragmented and it is impossible to reach 90% of the population. In fact, the more you spend, the more frequency you get because you are most likely to hit the same “heavy media users” over and over again. So, the elephant in the room that no one discusses, is just how much of the media investment is wasted.

        2. Scott Centurino

          It might be pedantic and obvious to state, but not all merchants have zero incentive. Just the established, larger ones (i.e. Amazon). A smaller, hungrier one (such as DaveinHackensack’s) is absolutely incented to pay for a broader range of funnel activities, specifically to compete with the larger player. However, that requires a significant commitment/risk by the linker since the ultimate conversion by the consumer is likely to be significantly lower (though perhaps not in the case of the extremely narrow Amazon program).I’ve often wondered whether a meta-affiliate system might be able to solve this problem. Sufficient success could potentially lead to a shift in the stingier programs, or their participation in the system. In fact, such a system could support additional funnel activities by providing advertising services as well.As with advertising and other intermediary systems, the issue is an inherent chicken-and-egg problem building scale (required for the economics to work). Merchants require scale of consumers (proxied by the content providers), content providers require merchants consumers will buy from (although paying for pre-conversion activities could help).

          1. ShanaC

            Essentially it sounds like you need an internet mall, as oddly dated as that sounds. Amazon in theory is supposed to fill that niche, it doesn’t sound like it is though.

      2. Mark Essel

        I can smell the opportunity. We need to shift the system so that the problem is reframed. 1) As a content creator and affiliate you would like to be compensated for downstream purchase actions directly or indirectly related to your link and commentary2) Those downstream (as Liad stated below) have no financial incentive to change the status quo3) To improve the analysis of downstream actions, whether they happen a day later or a few months later, real time data needs to be collected from sites/services which opt in to a Universal Tracker (protocol/standard). 4) Simple statistical models that correlate user purchase actions with multiple relevant product referrals can fairly distribute affiliate revenue. 5) Fully decoupled manufacturing processes from single brand drop shipping (Amazon and other big online brands, are ripe for disruption), will allow the market to fairly assess the value of an affiliate sale. 6) Quantization of sales influence, versus the various steps of design, manufacture, product warehousing, drop shipping, or other segments of product to consumer flow will be optimally rewarded to benefit not only the consumer, but all parties involved.

        1. Darren Herman

          In theory that works and that’s why we have “cookies.” The flash cookie was supposed to solve much of this. But there is ton of governmental pressure around privacy and the use of cookies for tracking and targeting is going to come under pressure potentially in 2010.

      3. jonathanmendez

        I’ve just posted some hard data and thoughts I’ve yet to share publicly in reply

        1. Stuart Watson

          Jon,Happy New Year. It’s been awhile and glad to see things are going well. I wanted to highlight a couple of things and ask some questions.The numbers are dead on…. for advertiser working off leads. Fred’s #s are for retail which are based on a commission %. Amazon’s affiliate program is not the best on many levels (as stated in the comments). So: Apples vs. Oranges. And if you take a look at the top 50 online advertisers (by media value) –, Nextag buys a lot of media (on a CPM) and get’s compensated on a CPC/CPA on the back side. While at Performics during the 2001 CPM bubble burst, we were telling publishers to put CPA ads in as default ads. If they had done that, they might not have turned to ad networks. What I’m trying to say: Publishers have had numerous opportunities to figure out how to monetize the audiences they aggregate. One very large publisher saw the light. Media General ( bought a coupon site Deal Taker ( question: Was this ad network buy a remarketing campaign or an acquisition campaign? I’m betting it was remarketing. Remarketing campaigns always work better (8.8 conversion rate) and cost less ($0.52). If it is an acquisition campaign, then yes the ad networks take a significant % of the revenue. A well known fact – The heart of the issue is attribution. It has been for awhile. Darren pointed to a few vendors working on the problem and here is another: Tagman ( Paul Cook, the CEO, goes as far as suggesting that with CPA partners advertisers should “award percentages of commission to multiple channels” – Boy will that make the publishers mad. I want to blog on that one because it’s easy to knock holes in that solution.Final questions:Do ad exchanges help solve the problem?– To understand the ad exchange/ad network landscape, read Pubmatic’s Ad Revenue report – does having the attribution data and building out a model enable an advertiser (or publishers) to act in real-time to help all parties (advertiser, publisher, consumer) win?Do traditional media measures (cost per point) come back into play?Which god do you believe in… view-thru or click-thru? You can’t have both.Lots of opportunity for new solutions and looking forward to feedback.Thanks Fred and Jon.

          1. jonathanmendez

            Stuart – great catching up with you here. Forgive my brevity (late at night and headed-off early tomorrow) but I wanted to reply to your thoughtful questions/comments.The buy in questions was networks.Was not re-marketing (as you know conversion rate is my specialty :)Attribution is large issue as Chris Dixon points out above but I think that is an buy-side issue with forward facing solutions (and a myriad of problems) and shouldn’t be lumped into the larger publisher issue (IMO) of media value and equitable distribution of it.In my opinion exchanges make this problem worse (seeing that now…go ask any performance advertiser and find out how excited they are about them).Indeed, lots of opportunities for new solutions. In my estimate there’s a good $6-$8B of media that should be shifted to the pubs (that calculation for another time/post).

  8. Melanie Notkin

    I agree. That’s certainly why I don’t participate in Cost Per Click or Cost Per Acquisition models…and I gave up on most affiliate programs because they bring in pennies and unless you are doing one off links (which I’m not), they are more work that they are worth.I’ve been told many times that while a consumer did not purchase an item because of a sponsored twitter conversation there and then, they did make a purchase at a retail location and verbally credited my telling them about it. Imagine if I was only compensated for direct purchases… and a tiny percentages thereof. I’m paid for promotion, not commission on a sale. Perhaps affiliate marketing should work in that direction.

  9. Alex Iskold

    Fred,Having seen affiliate monetization in action for the past few years, I completely agree with you. There is only 1 real winner – Amazon, capitalizing on the sum of the long tail.In addition to the problems that you described I will add a few more:1) The percentage that you earn resets monthly. Say you drove thousands of purchases the month before and was raised to 8% rate. The next month it all resets back, so while, as you are saying, the value was delivered in the past, Amazon erases that past from memory.2) To your point about measuring value per click. Amazon does not even display clicks you drove, it only displays transactions you generated. They make clicks appear to be totally useless, while themselves putting ads in pages and monetizing them.

    1. fredwilson

      i agree Alexthey do report on the number of click i generated

      1. Alex Iskold

        Hey Fred,I am assuming you are looking at the Orders Report.The click report that they provide is incomplete and confusing. They do show total clicks, but it is unclear for example, how many actual clicks each link you put up provided.This is because they show the total number of clicks and then they show clicks on things that people actually bought. So if your link was clicked and then something else got purchased – there is no way to see. Or if your link was clicked and then nothing was purchased, there is no way to see that either.So from my personal experience, has millions of objects and tracking what is actually going on using that system is impossible. (Unless I am missing something here…)

        1. fredwilson

          you may well be right Alexi just assumed they all came from my Happiness Project link because that’swhere the action has been on my blog in the past couple daysthe report is indeed confusing

    1. fredwilson

      you are right, there is a bigger point here. it’s the one that jonathanmendez posted about

  10. Caterina Fake

    This is why all the money goes to Google and whichever sites can position themselves in the last moment before a purchase is made, at the end of that process. A user sees something on your site. They have a vague recollection of the title, and search for it. Google takes them to Amazon, your cookie is overwritten and Google gets the kickback. Same thing for coupon sites. And research has also shown that there are specific time cycles for various purchases. I don’t remember the exact numbers, but someone will do research about a vacation they are taking an average of, say, 90 days before they actually book a hotel, and all the referring sites have been lost, overwritten and uncredited during that period.

    1. Darren Herman

      … and the dirty little secret about AOL IM… if that’s open, it’s dropping cookies…. and AOL gets the credit 😉

    2. fredwilson

      hi caterinanice to see you here in the comments!this is so true. i wonder how all the value that is being provided earlierin the process gets captured and paid?

      1. Caterina Fake

        That’s the problem, the “introducing” sites are not being tracked or compensated at all. But companies that are able to figure out which those introducing sites are, and spend their advertising dollars on those sites will have a significant advantage over sites that are spending most of their SEM dollars with Google not knowing the original referrer was AVC. They might find that advertising on cNet — or AVC– brings in more actual new leads than Google.

        1. ShanaC

          This should be trackable. I mean you should know on any given day where people are on the internet. This is like knowing approximately how many people are looking at vogue or a billboard at time square or a tv show. This is how you measure value. Why is there no universal system to measure that?

        2. Giles Bowkett

          Hi Caterina and Fred, I think you’re both missing some very important points and using some very flawed logic. I ranted about it on my blog: http://gilesbowkett.blogspo

    3. ericfranchi

      100% correct. All of the paid (and earned) media touchpoints get thrown out the window the moment that the user finally does a Google search to book/buy/subscribe, or as Darren alludes to below, goes directly to the marketer’s site and the last view-based cookie (usually via AOL’s fires.Big problem, and big opportunity for those who can figure it out!Eric Franchi

  11. Ian Rosenwach

    This speaks to the question of attribution in online advertising. The ideal ad system would compensate a Publisher for every impression, click, and purchase their site delivers. Each one of those actions has a different value. In affiliate marketing the purchase (conversion) is the compensated action. AdSense is the click, and ad networks are primarily either the impression or click. There are many fragmented players whose systems don’t communicate so it’s hard to track the life cycle of a single customer across multiple sites. Google’s search engine gets a lot of credit, while there may be many actions that lead up to a search. Maybe someone saw a billboard ad, and then an ad before a movie, a friend mentioned a product, they saw it in an offline store, and then searched online. Each step brought the person one step closer to the purchase, but they are all disconnected so it’s hard to spread the advertisers compensation equitably across each channel. Who knows, maybe the customers who purchased through your link were already planning on buying the book or a kindle, but your link happened to remind them and make it convenient. Affiliate Marketing is a CPA channel. Advertisers view it as such and publishers tend to do best when their traffic is near the bottom of the conversion funnel – close to buying something. I agree you drove valuable impressions to Amazon, but if they started compensating on impressions they would need to lower the CPA payments. It is an important advertising question that I don’t think is solved yet…good post.

  12. Nick Giglia

    It makes sense, Fred. I clicked on the link, and poked around with the book. However, I also just got a Kindle, so in the future, if I choose to purchase the book, I will likely do it straight off the Kindle. Even if I didn’t purchase right there, as you said, I now have awareness for the book (and an answer to why Meetup kept recommending I join things called “The Happiness Project”) and will likely purchase in the future. If I like it, I will probably influence others to buy the book, as will other people here.This is another symptom of the over-reliance on “quantifiable” data, when the value you’ve actually driven is much higher than the simple number of orders. Yet another abstraction brought about by the web…

    1. ShanaC

      That’s another mess and a half. New devices, new problems. I mean on some level, this is all studyable, but we’re going to have to throw part of the “everything must be quantifiable” into case studies and behavior watching. It’s going to be a confusing few years…

  13. jmarbach

    In the Amazon affiliate program, the affiliates cookie id lasts 24hrs (correct me if I’m wrong). So, I think you’re right, affiliates are being under compensated for the sales that they influence through their initial links. Do you think extending the cookie lifespan to 30 days or longer would be more fair in terms of compensation?

    1. fredwilson

      it would sure help

  14. chris dixon

    hey fredi think there are 2 problems. the first is this:…the second is even among intent harvester (affiliate links), only the last click gets paid, even though usually there are many affiliate clicks before a purchase.

    1. fredwilson

      jonathan mendez linked to a post of his higher up in this thread that makesthe same point you make in the “content sites are getting ripped off” postthis is a big issue for the web. i’m going to make it a priority to blogmore about it

      1. chris dixon

        yeah, i learned a lot about this issue from Jonathan’s blog and talking to him. its a huge issue on the web. its important to distinguish the two problems though: problem 1) intent generation vs harvesting, 2) even among harvesters need much better attribution (e.g. average of 3 *affiliate* clicks per camera purchase yet only last one gets paid – hence coupon guys and cookie stuffers are getting all the money, even among paid links).

        1. Alicia Navarro

          Thanks Fred and Chris, for raising and championing what is a major challenge in online publishing, and one that needs to be resolved if we want to continue to get quality content for free.One solution that will make a difference is a change in attribution models. This has been a hotly discussed topic at all the affiliate conferences in 2009, and its an initiative we plan to champion this year. Currently, content sites are not earning what they should from their content, because commissions are kept low by merchants due to the dominance of voucher/coupon and cashback sites. We believe the optimal solution is a change in the ‘last-click’ model to a distribution between first- and last-click, or last- and an interesting concept called ‘best-click’, defined as the content site that wrote the most valuable content about the item. There are many issues with both approaches, but both will go some way to dismantle the stranglehold on affiliate marketing that voucher and cashback sites have. This is something merchants would love, so we are encouraging them to absorb the short term pain changing attribution models would entail, in exchange for the long term survival of quality editorial content that discusses products in a broader context.Its a challenging initiative, but I do believe 2010 is the year for these changes to start to happen.

          1. COMRADITY

            Speaking of ‘best click’ – here’s an idea.What if the fee were related to the importance of the referrer to the total sales/day. For example, for many books in the longtail to succeed, they need the momentum of those first “keystone” referrers to get the ball rolling. Wouldn’t it be great if they were the most richly rewarded? One way would be to reward based on the proportion of daily sales they generate. This assumes the reasons for their high daily sales proportion would be a) high conversion rates due to credibility/trust for the longtail niche they represent and b) no one else knows about the book. I imagine someone is going to figure out how to “game” this – but I’m sure there are ways to protect against false orders.IN the case of “blockbuster” books, referrers wouldn’t get so much. Because no single referrer would be responsible for a significant portion of sales.But who cares – you want to encourage referrers to focus on the little known books – who cares what they think about the blockbusters.K—

          2. fredwilson

            Thanks for stopping by and commenting AliciaChanging attribution models seems like a very necessary step

      2. rosshill

        Just to throw it out there – should there be a financial connection between a suggestion and a purchase? Jason Kottke’s micropatronage drive a few years ago explored an alternate model for bloggers. His content is valuable and people who found it valuable contributed to him directly.

    2. Mark Essel

      Was just about to link to ya Chris, Wooha no need (that’s a big deal in the flakey iPhone world of cut and paste).It’s great seeing an important systemic error getting attention. Let’s invent some brilliant tracking solutions to renew the affiliate concept and maybe make a few bucks along the way.What we need are smarter links. First off users should see something like a shortened URL. We can do all sorts of fancy link tracking with shortened urls. We need a way of identifying user action with high confidence that was at least partially the result of earlier content sharing.When three of my super human filters or friends all praise a service or product who should get the affiliate cut? Right now only the last one does, unless I choose to go directly to the source or puchase from another friends affiliate link.On average the link value has the greatest value when it leads me to relevant and quality information, maybe even free of charge. How about a regular donation (user decided each month) that let’s us show appreciation for quality links and sharing. It’s clear that affiliate link value is under rewarded, driving us to search and invent more motivating models. Victus Media’s personal ads are initially designed around affiliate revenue (via Amazon), but we’re more than happy to consider more effective alternatives. Users could even tip us with a mini-payment if they appreciated a quality shared link, or go as far as signing up for a monthly or various per click fees to advertise their links through are matching system.

  15. Vivek Sharma

    Great post Fred. There are many great points here and affiliate marketing is a space with a lot of juice still left in it. An important point that hasn’t been stated is that Amazon is slowly but surely becoming the Walmart of the web. This is bad news for the thousands of smaller e-commerce sites. These smaller shops currently have to partner with Amazon, a competitor, in a Faustian bargain that dilutes their brand in the long run. For just this reason, my partner have started a company to help the little guy level the playing field with Amazon.

    1. AgeOfSophizm

      But first we have to take out the other 800 pound gorilla (i.e. Best Buy/Barnes and Noble etc.) which is well underway. Although you may look at Amazon as the Wal Mart of the internet, at least it is doing the job of disintermediating the big box retailers, who IMHO are worthless.

  16. Chris Dessi

    This is an interesting post, and it touches why I left the world of “pay for performance” – CPC, CPA etc in favor of the social media world. There is a fatal flaw in the pay for performance model – fraud. When people are incented to generate a conversion event (either click or purchase), they will do almost anything to generate the conversion. This is where subversive marketing tactics come into play (continuity programs that have lots of fine print) etc. Each Ad Network can be different regarding compensation structure, but generally cookie duration needs to be short because then they lose the reign on fraud detection. There is a never ending struggle between the advertiser and the publisher. The larger more respected companies with affiliate programs have minimal payout models like your Amazon example. The publishers that make the real money in this scenario are the search guys that have great relationships with numerous Ad Networks and are paid out special CPA’s (due to their stellar performance in the past) within 7 days of a pixel firing. The contractual obligations for the advertiser (risk) usually states that there are no chargebacks or deductions. These Ad Networks will take no margin, or even lose money to keep the search publisher to they can keep the advertiser from the competition. I can chat for days on this one. But you’ve just skimmed the surface on the grumbling underbelly of the internet. Great post.

  17. Jonathan Wegener

    I think Javery is onto something when he says “while I agree it is under-valued Amazon holds all the cards on this.” At the end of the day, there’s a market here. The affiliate fees that e-commerce sites are willing to pay will settle (or have settled) on the optimal payout rate that reflects the value being delivered by the inbound links, right? It’s a free market..I see this as similar to how Eric Schmidt suggests that Google doesn’t need to worry about click fraud because the market sorts it all out: “Eventually, the price that the advertiser is willing to pay for the conversion will decline, because the advertiser will realize that these are bad clicks, in other words, the value of the ad declines, so over some amount of time, the system is in-fact, self-correcting. In fact, there is a perfect economic solution which is to let it happen.” (

    1. fredwilson

      i agree that the market will sort it outsome enterprising entrepreneur will figure out how to allow me to sellGretchen’s book directly and cut out Amazon entirely

      1. Muhammadatt

        This is probably true, but isn’t it only a partial answer? I read the issue as the much larger/harder problem of trying to credit you with the total impact of your post — selling Gretchen’s book directly does not account for the folks that see it in a bookstore, remember it from your post, and then buy it on impulse. Or for the people who mention it to a loved one as a book they want to read, who then buys it for them as a gift, etc. Are you looking for a new model of affiliate sales or a new model of marketing measurement? Both?

        1. fredwilson

          better measurement would be the right place to start i think

          1. ShanaC

            OK what should we measure?

      2. joeagliozzo

        That probably only works if you are the exclusive seller of Gretchen’s book, which is not likely.For example if I wanted to buy Gretchen’s book and if you sold it directly from your site, I am still more likely to still go and see if it’s available on Amazon and buy it there because (a) I know there is very little risk of fraud at Amazon and (b) I have an Amazon prime account and get free expedited shipping.I think the only way this works at scale is if you match the convenience and price of Amazon and also have enough “loyalty” from readers that they are motivated to buy through your channel to support your site. Certainly possible (and maybe there is a private label ecommerce and fulfillment business that only sells through blogs?), but I think LIAD above was right in saying that many/most content providers don’t know what they are missing or don’t care enough to put up with the hassles of a different system.

      3. Lancelot_dL

        The issue is not only purchases of Gretchen’s book. In fact, Amazon paid you 4 times the affiliate rate for the books purchased due to other purchases by people coming from your link. The competition would have to be a competition to Amazon itself offering many items, and paying you by pay per click for driving traffic to their site. The amount of pay per click can perhaps be computed by the chances of future purchases for people coming from your links or other methods. This isn’t easy because of Amazon prime (which most heavy book buyers have) which joeagliozzo and others mentioned, but also Amazon’s prices and their general monopoly position which will only increase. It’s good to have you on the case.

    2. MattCope

      Jonathan and Javery are on point.If there are two gas stations across the street from each other, and one charges $2.59/gallon, and the other charges $2.99 a gallon, the first will see significantly more traffic.But Amazon is effectively the only gas station on the block.So, there’s not really a competitive market in affiliate book marketing. Yet.

  18. Dave Pinsen

    “bloggers and other users of affiliate services are getting under compensated for the value they are providing.”In the case of, judging by your example, that certainly seems to be the case. But that’s not true of all affiliate marketing programs. Consider the example of a subscription-based site I run, which currently has recurring referral commissions of 15% (I am thinking of bumping that up to 20%). Forty purchases through my affiliate badge at the current rate would have generated $84 in commissions for you, and these commissions would have been recurring, i.e., if the same 40 individuals remained paying members of the site next month, you would earn another $84, and so on.

    1. fredwilson

      now we are talking!

      1. Alex Iskold

        Could it be that Amazon rates are low historically because they started with books and their margins on books are too small? I mean I don’t think they can possibly afford to pay 15% to an affiliate, I doubt they make that much?

        1. Jerry Vandesic

          I don’t believe that the rates from Amazon have been historically low. I started as an Amazon affiliate at the very beginning, and the inital rate maxed out at 15% for books. This was the case for the first few years, but then they started ratcheting the rate down to basically 6-7%.

      2. Dave Pinsen

        At 20% (which I just had it changed to) it would be $112 on 40 purchases. Also, I let my affiliates sell the product at a 5% discount to what users would get going directly through my site — that way your readers have an incentive to buy through your affiliate link.

      3. Will Hambly

        Fred,Help me understand: There were 40 sales of the Happiness Project from your blog based on the cookie, which amounts to total sales of $608 for Amazon (15.20 x 40 = 608). You claim you received only $6.50 in referral fees. $6.50 divided by $608 is slightly more than 1%, putting you below the lowest payout tier of 4% offered by Amazon Associates. Please clarify the data, because it seems that you may be miscalculating.(You will see this post below too b/c I mistakenly replied to Dave)

        1. fredwilson

          here are the latest numbers:since jan 1, that link has sold 63 books for total revenue to amazon of $806and i have received a total of $32.27 in fees

    2. Will Hambly

      Help me understand: There were 40 sales of the Happiness Project from your blog based on the cookie, which amounts to total sales of $608 for Amazon (15.20 x 40 = 608). You claim you received only $6.50 in referral fees. $6.50 divided by $608 is slightly more than 1%, putting you below the lowest payout tier of 4% offered by Amazon Associates. Please clarify the data, because it seems that you may be miscalculating.

      1. Dave Pinsen

        Will,You may want to reply to Fred with this comment, since your question is directed toward him. That way, Fred will get the e-mail from Disqus notifying him of your comment instead of me.

  19. Joe Lazarus

    Here’s a PDF of key findings from one of the Comscore studies you mentioned in your post, which was done in partnership with Yahoo…'s an issue not just for affiliates, but for internal marketing teams as well, who often undervalue brand marketing and other advertising ROI by attributing conversions to the last click into their site instead of giving credit to impressions further up the purchase funnel. As Caterina suggested above, it’s especially problematic for high-consideration purchases like autos, mortgages, travel, and other big ticket items.

    1. fredwilson

      thanks Joe. i’ll add that link to my post

  20. AlanPearlstein

    Affiliates need to get credit for “view through” transactions also (transactions that occur after viewing an ad, as opposed to clicking on the ad). We tend to see three view through transactions for every click based transaction from an advertisement – I’m sure it would be similar for an affiliate ad. Every time I go to an ecommerce site, I visit a coupon site before I make the purchase. So the coupon site (an affiliate) gets the click credit for the sale, but I usually get to the site from a different ad (which might be from another affiliate), and that ad gets no credit for the sale, when, in actuality, the original ad drove the sale for the company.The issue is you don’t get credit for many of the immediate sales that you generate and you don’t get any credit for influencing future purchases (branding).

    1. Alex Iskold

      This sounds right to me, the only issue being click fraud.Right now, Amazon is dealing with this issue simply by saying if it transacted – its not a fraud, but if they paid per click then they have to deal with the whole fraud issue that Google is battling and its not really their core business now.

  21. Alex Bain

    Doesn’t this assume that the only form of compensation bloggers receive for their work is monetary? If you determined the true monetary value generated by an affiliate link, and the difference between that and the monetary compensation, couldn’t you argue that the difference is the monetary value of the goodwill a blogger received by pointing people toward books or other products that they liked enough to write about?Certainly all bloggers don’t seek that goodwill, but some seek it exclusively. Perhaps the average across the blogosphere is where we’ve landed.

    1. fredwilson

      true, certainly of mebut there are a lot of content creators who are doing it for a living

      1. Alex Bain

        Content creators (whether self-employed bloggers or newspaper journalists) will probably never again be fully compensated in money for their work as long as some talented writers derive some or all of their compensation from goodwill.I don’t see this as a problem. Just evolution and a market adjustment.There’s a thrill in finding/reaching an audience, and that’s payment enough for many. Most of the performers on American Idol are never monetarily compensated, but their work is the most popular content on television.

        1. COMRADITY

          The thrill of finding/reaching an audience is may have been enough when advertisers would pay a lot to reach them.Now that the content creator market leaders aren’t being paid enough by advertisers to create content, they are changing the game (e.g. Newscorp, Hearst re-transmission negotiations with MSO’s this week) in a way that can empower content creators all the way down the value chain. The more producers like Fox Network can command from distributors, the more content creators will be able to command. This is a market dynamic that will raise the bar, and create new value/wealth.

          1. Alex Bain

            I’m confused.I’m saying that content creators (e.g. bloggers, singers on Idol) enjoy reaching an audience, and the monetary value of that enjoyment is the difference between what they are being paid and what they “should” be paid.Even if many writers/singers/filmmakers rely on their art to pay rent, the fact that some don’t will generally make it difficult for the pros to receive 100% of their comp in dollars.Are you saying the opposite? That bloggers, singers, or other content creators will receive more monetary compensation for their work in the future?

          2. COMRADITY

            Yes. I’m saying that your point of view is true for the past, but the game is changing.The signal is this past week’s retransmission deals between FOX and Hearst and MSOs (cable operators like Cablevision & Time Warner Cable). Until now, FOX and the networks had been willing to “give away” their distribution for free in order to add to their “eyeball” count to increase ad revenues.Consumer purchases for media (including cable, ISP, and wireless access fees) surpassed advertising $ in 2003 and the gap continues to widen.The networks are aiming for a share of consumer purchases by seeking retransmission fees from MSO’s. Cable networks, like Hearsts Food Network, are also negotiating for increases.Content’s importance to creating value, wealth and power is increasing.I think this is a very good thing for everyone in the market.

  22. markslater

    i bought the book – about three web instances later. I went to felds book log – saw what he had bee reading (he reads alot!) found a book that i had been considering – then went to two other ‘web associates’ who’s readings i follow – bunched em all together and bought 4 books -so i transacted ‘because of you’ but not within the loop

  23. Batman

    Fascinating, as I’ve never thought about this before. Thanks for bringing the obvious to my attention. Unfortunately, I don’t think there’s going to be remedy anytime soon. What do you suggest?

  24. Andrew

    What’s your take on music affiliate marketing? Should music bloggers using affiliate links be paid more then 5% of every sale?I’m particularly curious because I noticed you didn’t post an iTunes iMix of your top songs of 2009, and instead opted to post a free streaming playlist. Also noticing that you only linked to amazon for album downloads. Why amazon?

    1. fredwilson

      i don’t buy music from apple. it goes back to getting screwed by their drmcrapi know that’s all gone but once burned twice shyso i buy all of my mp3s from amazon and emusici prefer a stream because you just hit play and listen

  25. Jules Pieri

    I think there is another dimension to this analysis…you generated about a 20% conversion of clicks to purchases from the traffic you drove (when you wrap in all transactions). Amazon gets some of the credit for this high rate, surely….people convert more effectively to a site they trust.But you get the balance of the credit for such a high conversion rate…you did not waste Amazon’s resources by sending them “junky” traffic. Similarly, on my site we consistently have 50% plus conversion to purchase rates when we (reluctantly) send traffic to Amazon. Obviously this is highly pre-sold, quality traffic. But we get the same crappy rates as a link farm. Substantially less in fact, since a link farm probably has higher volumes. I don’t know how to qualify “quality” but those exceptionally high conversion rates (yours and ours) seem a great proxy. You clearly have authority with your readers and that is worth money. You should get more of it.

  26. Greg Cohn

    Influential web publishers have always been underpaid for theIr influence. Considering amazon is an ecommerce site, it’s not surprising they’d only pay you for direct commerce generated — even if you got a longer-lived cookie, you’d still not get credit for all the additional awareness and buzz you can help generate. What if you could get paid by the publisher for that instead – say, by the follow or retweet?

  27. jkaljundi

    To add, the ComScore report “The Click Remains Irrelevant – Natural Born Clickers Return” shows it’s not just affiliate and other clickers returning and buying at some point in the future without advertisers affecting then not getting their share. It’s also non-clickers, people exposed to ads in many cases buying in the future.There is a lot of good research on this also in Microsoft Advertising Research Reports section: http://advertising.microsof…There are numerous good docs out there, but one is:…which says: “Research shows that conversion reporting based on the “last ad” standard fails to credit the multiple digital touchpoints in a consumer’s online history. The analysis digs deeper into the idea of multiple touchpoints that occur before a conversion, and looks specifically into touchpoints by “time.””Online advertising, like any marketing process, is a long tunnel. Looking at funnel models like AIDA (attention-interest-desire-action) show that you have to go through all these stages. What happens today is mostly the last direct response stage, where the money for last ad / last click often goes to search engines, comparison shopping or e-commerce sites. That does not mean that 90% of marketing was done before that, and it is incorrect to pay to only the last action point. Even worse, in many cases these last 10% of media spaces play lottery, who of them gets the money based on one of the last actions, when exactly the buyer is buying. That does not mean the previous actions – views or clicks, like on your Amazon affiliate link posting – were worthless and should not be compensated. That’s one of the biggest problems, and as such, opportunities in online business models and advertising today.

  28. Norbert Mayer-Wittmann

    Links are worhless — and link-based search engines are a hoax. (I’ve said that for YEARS + YEARS … 5 years ago I said about M$ “what do you want sell today” … today that applies to the big G ;)One of these and it won’t be long……;D

  29. Nick Johnson

    I completely agree! Affiliate marketing should bring more value. I’ve pretty stopped including affiliate links in my blog posts. I focus more on my products that I sell. I’ll still mention those books on Amazon that I liked an recommend, I just don’t count on making any big bucks from the conversions.

  30. Michael Lugassy

    As someone who run many affiliate programs at, I think longer, more secure cookies are definitely top priority. Valuing clicks or quality of visits like what eBay was doing on https://www.ebaypartnernetw… fails miserably. It lacks publisher transparency, not tied up to actual results and doesn’t solve the number one problem with cookie cutters.I believe this is why the biggest earners in CPA are PPC gurus. They are a wiz in closing the loop from intent (typing queries), to their own crafted landing page (engaging) and to the merchant checkout (action) without giving you too many options.In my opinion, a re-invention of the CPA model would start with better tracking, possibly through a centralized transaction server where the sale happens at the affiliate network and not on the merchant. This will allow future sales (and even multiple affiliate) get counted and accounted for.

  31. eran shir

    Great discussion on the problems of affiliate attribution. There is certainly a big issue with the current simplistic (and self-beneficial) model of last click/cookie drop attribution. As Darren mentioned, AOL/ are notoriously known for flooding the web with cookies just to get the CPA attribution, regardless of whether they advertised that product or not.However, the more important point is that the absolute rates affiliate programs pay are 5x-10x lower than what they should and could be. If you compare the negotiated deals big CPA networks (like strike with ecommerce sites vs. what the same sites offer to their affiliates, you see a huge gap, sometime up to an order of magnitude. Not to mention things like view through attribution which doesn’t even exist in most affiliate programs.In short, while the attribution problem is a real one, solving it will not help much until affiliate programs start paying fees closer to the real value of the transaction.

    1. joeagliozzo

      Eran, your observation about disparity between large and small networks says that there are more opportunities for content provider networks (or aggregators). By joining a network you get more bargaining power and obviously the money is there if the ecommerce sites will pay the larger traffic providers more $$. The network is entitled to a cut because they added value..

      1. eran shir

        There is certainly an opportunity in building something like groupon or priceline for affiliate marketers. That is, a ‘network’ that will negotiate realistic terms with merchants and enable any affiliate to get in on the better terms in an adhoc manner. The problem is that merchants will try to fight that hard, and since affiliates are far from being a cohesive group, they’re a great group to exercise ‘divide and conquer’ on. This is the well known game theoretic problem “tragedy of the commons”:… Web publishers suffer from this problem not only when it comes to affiliate attribution. Same mechanism drives down CPMs and will continue to.

  32. ShanaC

    See, I heard about Gretchen” Happiness Project’s Blog somewhere Else. I didn’t even realize there was a book before here. Which makes all of this problematic. How many times do I have to hear about it before I buy the book. A big problem now…

  33. Jan Schultink

    I did not even click the link, but might buy later…But in this case, is it all about dollars? You spread the word about a book that can make people happier. The author of this wonderful book will see the revenue somehow, and be encouraged to write more.

  34. Chris Messina

    I think this is an interesting subject, and one that I’ve not thought deeply about w/r/t “how much” I should be getting from my affiliate clicks. For about a year I’ve included Amazon affiliate links in some of my tweets — occasionally for products I’ve bought, or for good deals that I want to pass on to my readers — and have a substantial range in my conversion. Obviously the more stuff you sell, the higher referral rate you earn, but it’s true that I’ll never see any upside for purchases that happen after 24 hours.Still, I have a few provocations that might help me understand your perspective better:1. How much do you think you should have earned from all the business you sent to Amazon? You made $25.20, but seem to think you deserved more. If you had made a 20% commission, but were limited to the same time window, would that have been more satisfying? What if the affiliate rate was 1%, but lasted for 30 days? I’m curious what would feel like a “fair” amount of income given the traffic you sent to Amazon, so that we have a real number to pivot against.2. I’ve tried some other affiliate systems (not many, but a few) and Amazon’s is in so many ways superior and easier to use. They’re of course also very well integrated into the shopping experience, and so when I’m shopping on Amazon and see something that I want to tweet about, there’s now a button for that. I know of few other shopping or ecommerce sites that make it so easy to be rewarded for providing the site traffic, without going through some complicated sign up process. Shouldn’t that count for something?Put another way, had you wanted to link to, say, the Barnes and Noble book because they had a better affiliate rate (if they even have an affiliate system), would you have any idea how do to that?3. You mentioned wanting to sell the book directly — obviating Amazon as the middle man. In such a circumstance, would that also mean that you’d be willing to deal with fulfillment, credit card processing, and returns? What amount of your affiliate rate would you be willing to give up to cover those kinds of costs? I’m just trying to understand how this actually would play out if you were able to sell direct — especially given the huge array of products that Amazon carries beyond books.4. Amazon is also highly entrenched, and is a well known brand. When it comes to sending your readers off to someplace else, it seems that you’re stuck in a quandary: find a retailer that will actually pay you what you’re worth for the traffic you generate for them and accept whatever kind of user experience or brand unfamiliarity they may offer, or go with the known brand (Amazon), which people trust and probably already have already an account with (and can thus buy with 1-click). That is, would you rather send your traffic to “” (which no one’s heard) for a higher rate of commission, or would you rather accept a lower commission for increasing the likelihood of an actual sale?Just curious how you wade through these issues…

    1. Bart Epstein

      Chris’ points go wonderfully to the heart of the matter: Do any other reputable online booksellers offer better affiliate programs? If so, would you make more money pointing your readers there? If not, the market seems to have spoken.My suspicion, by the way, is that, for a huge segment of your readers, if you point us to some other bookseller we will click the link but make the purchase at because that is the site we know, love, and trust.I’m not sure about everyone else but I can report that they have done a fantastic job of keeping me happy over the years — even more so since I took the plunge last year and joined Amazon Prime. They are so reliable, efficient, and simple that is easy to take them for granted.

  35. David Esrati

    Fred- the real question is can we skip Amazon entirely- at least in the publishing field? Authors could have a digital file- and sell direct, compensating the referrer a much smaller percentage than the cut both the publisher and the retailer take- and paying the agent- you, much more.We’re moving to an economy of “trust agents” or an “expert economy” and their should be a system that rewards the key movers- because they generate real interest.Build that system- and publishing changes- including newspapers.

    1. JLM

      Very thoughtful comment. I tend to agree that one of the missed opportunities — perhaps more accurately: one of the evolving opportunities — is to get direct to the customer via the Internet. And to bypass as many steps as are possible which add layers of cost and consume time.Of course, Dell Computer did just that years ago starting with catalogs and then moving to the Internet.As Dell has recently done, gravitating back toward more traditional channels of distribution, I suspect the future lies somewhere in between which makes it a more difficult task for an author.The publishing business is arguably a “valued added” iterative process with editors and marketing folks who have enormous advantages which would be very difficult to duplicate effectively.I believe that Steven King has released some of his most recent books via the Internet.

    2. fredwilson

      i think you are right, selling directly and cutting out Amazon, may be the solution

      1. David Esrati

        Fred- that’s what happens when the greed and arrogance seeps into business. Watch as Wall Street has this lesson forced on it in the next few years.I believe we’re going to see epic changes- with a move to local investing.We’re also going to see the total self-destruction of the broadcast networks, cable bundling- and a shift to the favor of content producers- because why produce for someone else, when you have direct access.David Chase doesn’t need HBO to do more Sopranos. HBO needs David Chase though.The balance will shift.This doesn’t solve the problem of tangible goods- where Amazon makes things easy with their affiliate system. But, watch Google to come out with a universal system- that works for every site and pays affiliates- they are, after all- selling commodities (basically anything with a UPC code).

      2. BmoreWire

        I agree with this but this requires a set of tools in bloggers hands where relevant advertisers are offered to them and then they can link them in or choose the banners from a gallery ala Zemanta. PR meets Advertising. I wrote a similar post on this topic

        1. fredwilson

          open sky is a new startup that is working on this ideahttp://www.theopenskyprojec…

          1. BmoreWire

            I guess you just have to make sure it doesn’t end up feeling like Amway. Also, the concept needs to go beyond retail to keep content generators making good content. Need a solution get gets Auto, CPG, and Telco dollars. i.e. Ford launches a new model and loads a palate for auto bloggers with press releases quotes, images, specs and quote request modules. Blogger gets paid on reach, frequency, cpm. Which also means you need a white listing system for bloggers to get access to premium ad dollars.

      3. lindsaycampbell

        Aren’t we failing to recognize the crucial part of the distribution chain Amazon provides- the storing and shipping- that would make “direct” sales cumbersome and expensive. Part of the reason Amazon got 40 purchases off of your post is not only that people trust that your recommendation but that Amazon makes it infinitely easier (especially with 1-click) and cheaper to buy through them than to get the book by other means.

        1. COMRADITY

          Agree. Instead of thinking of Amazon of paying too little, I think of them as charging too much for the services they provide (transaction, fulfillment). Amazon charges a premium for these services because they collect the money and because they are taking advantage of the fragmentation of the marketplace. I’d rather they continue to collect the money, the opportunity is to overcome marketplace fragmentation.

        2. fredwilson

          yes, and i’d argue that the one click part is even more important than thestoring and shippingbut someone could build that and make it available to publishers andbloggers

  36. Robi Ganguly

    Fred, I think that there are several ways to look at this. One is the way you put it in this post: bloggers and many affiliate marketers are getting under compensated.Another way of looking at it turns it around and looks at it more positively: You have reached a point where you can write a post and generate $25.20 in direct revenue for that post, in addition to the numerous other actions that people take on your blog when viewing a post (most of which have some value, I’m sure).How does this compare to where you were a year ago, or 3 years ago?Corporations are still figuring out how to make themselves profitable on the web, much less attribute that value to every consumer’s action.Yes, you are right to be frustrated with Amazon’s tracking, but why aren’t we looking at all of the other companies who could be competing with Amazon in terms of payment terms, %’s etc? Why aren’t you asking your portfolio companies to tell you how much value your links to them create?I think instead of talking about getting paid “more” we’re best served by talking about getting paid “properly”. It seems like you used to be happy with any payments/revenue coming from this blog, at all, and I’m sure the numbers were much lower. Now, you’re asking yourself: am I getting full credit for the activity that I drive?Improving the attribution analytics that result in “proper” payment across every channel will help in getting there, but what we also need is a market dynamic to get to a more “proper” answer. Without competition for the right to pay you, we’re not going to get there.Ultimately, I think the fact that Amazon pays you $25 today, without much in the way of affiliate competition for the goods that they offer, is rather surprising.

  37. Josef

    Fred, that problem permeates the entire online advertising industry.Media agencies are a part of the problem. Click-throughs are easy to understand and easy to evaluate quickly, and as long as it is what media agencies (and therefore their clients) look at as one of the major criteria for a campaign’s success, the online advertising industry won’t evolve.The study you allude to was done by ComScore and sponsored by the Online Publishers Association (which means a lot of agencies probably wouldn’t pay attention because it would be seen as biased): work for a large online publisher and we have seen many studies, both internal and 3rd party that show how online advertising elevates brand recall, unaided awareness and message association. But the promise of being able to measure the ROI of every marketing dollar has swung the pendulum too far.Not every marketer will be able directly measure the ROI of their marketing dollars. If a brick and mortar retailer uses online advertising to elevate awareness of their brand, there is still no great way to directly tie the online advertising to in-store traffic. So agencies, publishers and advertisers try to concoct all kinds of clever tricks to get some level – any level – of interaction: sweepstakes, games, microsites.Sometimes they work and they’re worthwhile. But at some point, advertisers have to either see the value of online advertising beyond clicks, or it will be relegated to bottom-feeding CPM rates and Direct-response campaigns, both of which will eventually lower the quality of a lot of advertising-funded sites which we take for granted today.

  38. Chris Phenner

    Most comments here place blame for under-valuing publishers on (i) Amazon’s stingy affiliate policies and (ii) under-reported purchase intent measurement. I see comments from ‘DaveinHackensack’ and ‘David Esrati’ that I think are interesting because they allude to outcomes other than a book’s sale. In other words, perhaps an ‘affiliate book sale’ is a ‘bad outcome.’Consider if the Happiness Project focused its affiliate links on a higher-value outcome than a book’s sale, and then rewarded third parties for ‘demand creation’ (not book purchases). 5% of a book’s sale price if purchased within a 24-hour session period from an link is likely ‘aiming too narrow’ for capturing the right value. If you visit the Happiness Project’s web site, you’ll see a half-dozen calls-to-action (eg, purchase book, follow on Twitter, friend on Facebook, sign up for weekly/monthly emails). What’s unclear to me (but should be clear to Gretchen at Happiness Project) is which among those outcomes is of the highest value.’DaveinHackensack’ gets it right because he notes that subscription-based outcomes are far higher-yielding than one-time purchase events (whether from Amazon or or a directly-sold item, whether physical or digital). Gretchen could isolate all affiliate-related links through a funnel (say, shadowbox reg flow) that optimized for the best outcome for the Happiness Project. Of course, Fred would have to know about that link, but perhaps platform companies in USV’s portfolio (Adaptive Blue, Tumblr and Disqus come to mind) could play an enabling role in ‘triaging the affiliate outcome’ and offer Gretchen control over the user flow.I negotiate CPA vs revshare-based relationships for a living, and I am always disappointed when a self-described ‘long term partnership’ counterpart opts for CPA over revshare — it tips their hand that they are more concerned about nearer-term than longer-term financial outcomes. If publishers like Happiness Project are choosing a book’s sale as the focus of their affiliate marketing efforts, then I would say they (and are getting what they’re asking for.A directly-billed, recurring relationship between Happiness Project and [consumer] is the best outcome, and you can look no farther than Zynga’s valuation as the best example of this. I’m not saying Happiness Project and Zynga will ever enjoy the same audience sizes, but I’d say we complain less about AMZN’s lame affiliate tracking and focus more on what means by which Happiness Project can capture real value (and subsequent to that, remunerate affiliates).

  39. Mihai Badoiu

    There are separate issues:1. Is the rate the publisher gets for the complete transaction high enough?2. Should publisher get payed for impressions?3. Is there an issue with tracking attribution?On 1, the answer is rather simple: the rate is not low. Yes, you may have provided greater value, but Amazon is the place to go for buying books and you’re a rather small affiliate. There’s an implicit negotiation between Amazon and affiliates, and Amazon has more leverage. Most payments (prices, salaries, etc) are based on negotiation, rather than value added. They reflect the value added only when there’s serious competition. This is not the case here.On 2, in a perfect world the publisher should be payed for impressions. Say the click improves the brand, by making the user aware of the site. Then the user mentions the site to his friends who may purchase in the future. There is clearly some very small value here.On 3, Amazon definitely has the technology to track attribution much better. There’s the wish list or cart: I have added the book to my cart and then “saved for later” (different than wish list), but that doesn’t mean I’m going to buy it. Amazon can easily track this. I doubt they do. Then, maybe I bought books that are on the same subject. Amazon can track this because they have lots of data about related books. (who bought what, categories, etc) Also, with the Amazon cookie, they can track all my future actions within this area, up to the buying of books. Amazon can easily do all that. (I don’t know whether they do or not) But clearly, there is no technical issue in this case.PS I wonder whether Amazon get the same rate (of 0.013% based on $6.5 revenue on 40 clicks for a $12.5 book) from Google. Google is a much bigger fish, and they may have a lot more leverage. My guess is they get more, a lot more.

  40. Ilya Lichtenstein

    I posted this on HN, but I would like to reiterate that Amazon is a VERY small piece of the affiliate marketing world.Most affiliate marketing money comes from CPA, which pays closer to $40 per purchase. This averages out to about $2-$3 per click, which is how much my company as well as many other successful affiliates are making.I think affiliate marketing OVERvalues the link. When you’re making $2-$3 a click, you end up raising PPC bid prices significantly, which ends up forcing out anyone other than affiliates.Feel free to email me at ilya [at] if you want to learn about how the affiliate marketing industry really works.

  41. Michael Johnson

    Yes, Affiliate Marketing can make you a nice living

  42. Karl Rossmann

    After reading your post, I remain unconvinced. Your blog post is worth 5 cents per click. Your post compelled commerce totaling $25.20 for 118 transactions = 21 cents per transaction. Not bad for writing up a couple of thoughts off the top of your head. Honestly, I think this is a pretty good deal.

  43. eric_bingham

    Fred – it could have been a lot worse. Your conversion rate of 22% is significant. The shortcomings of the affiliate model (for publishers) are well known and have been explored nicely in the posts here. The interesting thing to me is how to harness the value of sites like yours with well informed and passionate users. (sites with passionate/targeted users) reportedly earns half its revenue from Google ads. That’s a step above affiliates in the monetization chain but requires significant tonnage. Anyone who can aggregate and measure actions from a large set of such sites has a real opportunity.

  44. Chris Pirillo

    Affiliate marketing usually sucks for the publishers. That’s the rule, NOT the exception.

  45. rafer

    Fred, I don’t think you and JonathanM are being reasonable. You can’t have it both ways. Either you build a direct sales channel (and therefore suffer high effort/harder scaling) or you automatically lose most of the EVA. Companies are acquired for channel, not for tech. This is why.

  46. Vijay Dondeti

    Fred, I think the other comScore report you are looking for is discussed here the way, great blog. Long time reader, first time commenter.)

  47. Prokofy

    I *think* what you’re saying is that these big companies that offer you a revenue share like this are low-balling you when it comes to what you bring to the table. And you’re right. All of us tilling in the field of Web 2.0, whether a big ranch owner like you or a share-cropper like me are getting shafted.

  48. paramendra

    You are so very right. Especially now that Amazon and Blogger have integrated and I intend to do affiliate marketing with a vengeance in 2010.But I misunderstood the blog title. I already had a comment in mind before I read the first sentence: ads make sure Google searches are free. From the title I thought you were saying it is wrong for bloggers to link text in their blog posts to affiliate marketing products.I am glad you do some pro-blogging. It keeps the entrepreneurial spirit on the edge and fresh.

  49. Jim Peterson

    Hi Fred:I subscribed to Gretchen’s blog based on your referral- I’ll be in the camp that buys the book later where you don’t get a commission, I’m sure.I also sell books to Amazon through a distributor that supplies them books in the construction category. They require to buy the books for 15% less than anyone else does (of us, at least)- which they pass on to the end user with lower prices….So it is rather circular…consumers get low prices…Amazon provides incredible service…so everybody has an Amazon account…To sell books on one’s site is is just easier and more helpful to the reader to link to Amazon has lots of volume and can demand low prices and good terms (short lasting cookies, etc.) …And on and on it goes!

  50. Curt Monash

    I’m a little confused on the numbers. https://affiliate-program.a… suggests affiliates get a minimum of 4% per purchase, yet your cut — $0.17 or so per copy — on the original book seems to have been around 1%.

  51. Thibaut Barrère

    Amazon used to be a good payer a few years back. Gradually the need for people to promote their products became less important and they have been lowering their commissions, if I recall well.It seems to be like some kind of rule of thumb: the program is interesting for the publisher as long as the advertiser is still trying to grow on its market. Amazon is already very big on its market, so the incentive to push people to sell out their stuff became lower.But there are still some interesting affiliate programs out there (eg: some niche software), many of them with more than 24h cookie for instance 🙂

  52. Jeff Greenfield

    Fred – It’s about time that Affiliates are paid on the same basis of Display partners — both on the ‘view-thru’ and the ‘click-thru’.The real question is how you properly attribute those sales.If allowed to run ‘wild’ like the current display market — the fraud would be tremendous.

  53. AndyY

    so you were not compensated for the visitors that didn’t purchase, and they may in the future. but forgetting practicalities of tracking this stuff, what’s the realistic value? it may not be so much if the average customer visits several times from several sources before the same way, you were perhaps over-compensated for the 40 that did actually purchase, since it’s likely that some of them had seen the product before, and perhaps even viewed the amazon page via other sources that could be entitled to a % of the commission.there must be a ceiling on the max commission payout for a purchase, so the interesting question is the relative value of each contributing referral. if a customer comes in via 3 separate sources before finally purchasing, should the commission be split equally? should the first and/or last receive more? should content publishers receive more/less than google ads when both appear in the history? etc.simple model: assume the average customer makes n visits (via n sources) for a purchase, buying on the n-th visit. suppose each of the n sources should be compensated equally and the commission is x% – then they each get (x/n)%.but the current system only credits the most recent source, giving them 100% of the commission. assuming random distributions etc, as a source you’re likely to be the last source before purchase 1/n of the time, but you’ll get n times your entitled commission (under this model) when you do. 1/n times n is 1, so over time the commission you’re still likely to get your fair commission when compensated only for being the most recent referring source, how the current system works!I think this result remains the same even if you decide the the % commissions should be divided up differently. so, measurement and referral credit is imperfect. but how far off from the theoretical ideal is it, in terms of percentage dollars? perhaps not so far. perhaps, on average, over time, no different!so while I agree that there’s reasons to add value above your 5c/click, the other reasons that suggest it may be lower than 5c/click could balance out, exactly!

  54. BmoreWire

    This is really the nature of affiliate. The purpose of affiliate is people put links in place that can’t afford a sales force and don’t know the value of their inventory. We did a lot of analysis of affiliate vs network vs search for advertisers and affiliate is almost as valuable as search…..basically because of two reasons. Affiliates pull ads that are relevant to their sites so it is more relevant than a standard banner also it’s way undervalued so you get a lot of bang for your buck which is why a lot of online advertisers use affiliate and search as the foundation to their online marketing plan because search works, and affiliate is the best deal around (that is if you can palate a bit of a cowboy’s marketplace where your ad could show up anywhere with little control.)

  55. Angel Djambazov

    A couple of notes about affiliate marketing from Amazon’s point of view:1) Amazon thinks of large shopping portals like Bizrate, Nextag, and as affiliates. Bizrate alone averaged 18 million uniques in November ’09 according to Compete. Within those 18 million is the ONLY non-search engine that ranks in the top 5 of their downstream. With that kind of traffic it is a difficult argument to get them to see 535 views as significant.2) Chris Dixon is right that affiliate marketing is a last cookie in system. It is also valuable to know as a publisher that of ALL the book merchants out there Amazon has the stingiest cookie length. Amazon’s cookie is “session only” which means if someone doesn’t buy in that session you don’t get credit. Places like Barnes&Noble, Borders, Powell’s range up to 45 days after the referral where you will still get credit for the sale. The point is to research your advertising partner and refer traffic to the one that will offer you the best deal in a) payout, b) referral length, c) tools and support.3) Finally if you are going to send traffic to Amazon I highly recommend using Amazon’s A-Store which provides customizable product widgets. It’s a good way to display the book/product you are talking about and refer your reader to the page of that product rather than just to Amazon’s homepage. The closer to checkout they are the more likely they will purchase.

    1. fredwilson

      thanks for that last suggestioni’ll give that a try next time

      1. deancollins

        The easiest way is to download the amazon ie toolbar and browse to the product – provides cut and paste link (i assume firefox has something similar).

      2. lindsaycampbell

        I also think mentioning that purchasing through your link drives money to your blog (which readers know goes to charitable causes) would be another way to encourage purchases that you get credit for. That kind of transparency is a good thing, especially for people like me who haven’t given much thought to affiliate programs.

  56. CollegiateLiving

    re: the affiliate fee isn’t derived from the value that Amazon gets from the link, it’s based on the fact that even with that low value linking to Amazon will still make you more than linking to the same item on any other store on the web. There is no reason Amazon should pay more. It’s one of the many reasons I am long on Amazon.eBay also does this as well…long term power sellers on Ebay can benefit from contiuous traffic from affilaites and page one google placements from auctioned items…can you say rock and a hard place?

  57. StevenDiamond

    I have come to the very same conclusion on my own blog. It’s time Amazon share the wealth. Power to the people!

  58. Fabio Seixas

    Fred, it is really hard to track all this other transactions, principally the ones that do not involve the click, in a way Amazon (or any other retailer) can bring up an accurate solution.But one way to compensate the affiliate for this is paying then a smaller amount for some of the future purchases the new costumers made at Amazon. This isn’t an accurate solution but can minimize this issue and even motivate the blogger to bring more premium costumers (the ones that made many future purchases) to Amazon.

  59. bdambrosioi

    The immediate sales are just a proxy for the much more difficult to measure value of the lead-gen.”Value” isn’t an arbitrary number – it is determined by the dynamic of the marketplace and supply-demand. The supply is large enough that the price Amazon has to pay for the volume it wants is, apparently, below Amazon’s limit price. If the supply was smaller, and Amazon thought it worthwhile, it would “value” (ie offer to pay) affiliate clicks higher.Are you arguing they should pay more just because it would be a fairer split of the gained revenue? Business doesn’t work that way. You could try to group affiliates and try to negotiate from a position of strength, but I suspect it would be very difficult to gain a large enough fraction of sites to have leverage.

  60. Financial Samurai

    Great analysis! Affiliate links to amazon and such really isn’t that great a proposition for bloggers. They win, we win much less.

  61. David Lanning

    Affiliate marketing (as opposed to selling your own product) does have several drawbacks.But, on the upside… affiliate marketing doesn’t require you to invest cash for inventory up front and risk not making the sale. Yes, I understand that building your readers and maintaining a blog is an investment. But, you don’t have a warehouse of books to worry about selling.The risk involved with affiliate marketing is less and consequently the rewards.

  62. Gerald Buckley

    In my previous gig as a science, technical, medical (STM) publisher… Amazon was a DISASTER of a channel for us. It was abysmally bad. They put the squeeze on their STM publishers to the point there is no profit margin left in the deal. Make THAT up on volume ;)Sure, we had our own web shingle. All worked our well. Good margins. Yada Yada. What worked INCREDIBLY well wass selling articles or chapters on a pay per view basis from every journal or book we’d published (back to 1917). THAT was (and remains) phenomenally lucrative.The issue in that space now seems to be the schism happening around Google Books and how they’re posturing with regard to copyrighted works and orphaned works. I’m not so sure they overstepped their bounds. But, hey… it’s got people talking seriously about the merits of copyright overhaul finally.Amazon’s NOT the only game in town. Lots of people do make money via Amazon’s marketplace. We chose to keep pricing parity across our channels and just couldn’t make it work in our case.

  63. fretbase

    Regular reader, first time commenter. :-). While I completely agree with your post, I don’t think Amazon is really reflective of the entire affiliate system.I am one of the publishers of – a website for guitar enthusiasts. Our primary source of revenue for our website is affiliate links and we have learned a lot about these programs through our experience with them and our commissions on thousands of items sold after our users clicked an affiliate link on our website. In short, Amazon’s policies are much less generous terms than many of the affiliate programs offered through LinkShare or Commission Junction in terms of when you get credit for the transaction.For example, through commission junction we promote purchases at Guitar Center. For those purchases, we get credit for the transaction (and our commission) not only for visits during the session the user has after clicking on our link but also for purchases resulting from visits by that same user up to 14 days after the click.The result is that we more heavily promote Guitar Center and Musician’s Friend for guitar gear purchases than we promote So yes, the market does really play a role in helping to sort this out.On a similar note, eBay recently reduced the affiliate fees paid to their advertisers and also lowered the transparency into how the fees are calculated. As a result, they will move down and others will move up.In short, I agree with you that Amazon under pays but they aren’t necessarily reflective of all affiliate advertisers and I think the market will sort this out.

  64. valto

    There should be “free affiliate marketing provider”. I have not been searching for affiliate marketing providers/networks for a while, or haven’t catch any news in past years to notice if there is someone doing this already.But some years back when I was looking for good affiliate provider (to sell own products), all good ones that I could find (that looked legit) were pretty costly to become a seller (setup fee) ie. to get own affiliate program setup.The natural business model would be that the affiliate network would only take cut in between and/or display relevant other advertising in between (if there is no cash value/transaction to take cut from).In addition the network could sell “market data & analytics” to offline retailers etc.Seller could freely setup and adjust the affiliate compensation model/pricing etc. and then offer it to marketers.

  65. gianfulgoni

    Fred, you’re correct that past comScore research has shown that search and display ads cause latent buying and offline buying that are not reflected in the click:…Essentially, it all comes back to the fact that the click does not reflect the “view thru” impact of ads (i.e. the impact of ads that are not clicked) — and that even clicks don’t accurately measure all the latent buying that occurs (because of cookie deletion).Interestingly, this is not at all surprising to anyone who has had experience measuring the effectiveness of advertising in traditional media. Because there are no clicks to measure or cookies to track in TV, print or radio advertising, researchers use consumer panels and analytical designs that compare the behavior of people exposed to an ad campaign with the behavior of a balanced control group of people who were not exposed. The buying difference between the groups (after adjusting for any pre-campaign differences that might have existed — and which are very small if the control group is carefully selected) can be attributed to the effects of the ad campaign. The same reliable analytical design is used in cinical trials by pharmaceutical companies to measure the efficacy of new drugs.At comScore, we use our panel and the test vs. control group approach also on behalf of many of our publisher and advertiser clients to measure the effectiveness of online search and display ad campaigns in lifting sales — both online and offline. We have also used it to measure the holisitc impact of referral programs. It really is the only way to accurately measure the complete impact of online ads and referral links in lifting sales of the advertised brand.Here’s a link to an example of the encouraging results we get when we use the approach to compare the effectiveness of online ad campaigns with TV advertising. It compares online and traditional advertising on an apples to apples basis:

    1. fredwilson

      thanks so much Giani am going to copy and paste some of this into my main post


      @gianfulgoni – What if we changed the channel?Most of the above comments start with the traditional premise that websites are publishers, affiliate links are ads, and Amazon is the retailer. This is anointing Amazon with a lot more value and power than they are due. Compared to “real world” retailers, Amazon isn’t a retailer at all. Retailing is about choosing the right products and merchandising them to help consumers make purchase decisions with confidence. Amazon dominates the value chain, despite not being a retailer, because the marketplace is highly fragmented – true retailers with the knowledge to merchandise, their vendors, and their market individuals are isolated from one another.If you change the channel and think of the longtail publisher as a specialty retailer, the problems and opportunities are different. In the real world, the specialty retailer’s asset is that he knows more about his niche than his customers do. He leverages this asset by using it to filter and select the best stuff and using it to merchandises the stuff in the store to increase the chance that people will feel confident about a purchase decision.An example: an avid fly fisherman has a retail store where he displays flies on a board that pictures the conditions for which a particular fly is optimum, based on his experience. This retailer takes a lot of risk – i.e., buying and maintaining inventory, pays rent. After all these costs, he nets 20% before marketing. Back in the day, his marketing is localized, through commissions from fishing guides, and direct marketing. After that his profits may be 10% of gross revenues, reflecting the value of his knowledge asset and the time he has spent developing it and communicating it to customers in the store.Now the specialty retailer eliminates his rent expense and goes online. He publishes content based on his knowledge and merchandises by embedding relevant affiliate products and services, eliminating his inventory cost. For example, he creates an interactive form in which the browser inputs conditions to reveal the optimum flies, with a link to buy them. Instead of buying and inventorying these flies, the link goes to a page on another website that executes the transaction and fulfills the order. He is paid a 10% affiliate fee.In this context, he has no margin left for marketing. Affiliate programs argue that “if you put it up there they will come” (which is a myth based on the days when mass media had limited choices) and that it is “free” to market on the web through social media. But of course this is not true. This guy is already spending most of his time advancing his knowledge to be the resource for the best information and the best flies and fishing equipment, and publishing it in his blog. He has no time left to learn and execute “free” social media.How do you convince Amazon and other transaction/fulfillment platforms to raise the affiliate fee? How do you enhance the clout these longtail websites have? By finding, networking and organizing the websites, their vendors, and their market’s individuals. The result is enhancing marketing effectiveness (both by eliminating waste and providing a relevant context in which communication is more compelling), real time metrics to improve performance, and critical mass to increase leverage at the bargaining table with Amazon.

  66. Jerome Camblain

    You’re right Fred as you also did not take into consideration the foreign readers of your blog who went to buy it on their local (national) amazon website that stores their account details (as I did).

    1. fredwilson

      oh great, another way to lose creditsomeone should make a list!happy new year Jerome

    2. David Noël

      Just saw your comment after posting mine.That’s exactly what I did too.

  67. kidmercury

    the real solution is for communities to have their own shopping cart and currency. part of the aff revenue is shared with the community. this wipes out all teh tracking and branding issues. also sets the stage for selling value added services (i.e. book club for happiness project).but once we start thinking of how to make this scalable we quickly run into the money supply issue. as there are no more venture capitalists, only venture fascists, this is a bit of a problem.anyway, the price of the freeconomics renaissance is peace, rebuilding the moeny supply so that war is no longer profitable is a part of hte price. i know i sound like a broken record but ignorance is the only opponent and this is the most effective means, as evidenced by how annoying it is.

  68. falicon

    Holy crap this sparked quite the response (sign of a real problem worth solving!)…a few quick points/questions to add to the fire:1. Isn’t this really in the realm of “brand marketing” and how to be properly compensated for promoting a brand?2. Maybe it has a bit more to do with intent…in which case, it’s probably more of an education process for users (bloggers) than anything else. If your intent was to promote the book, your best bet is to do a direct link to the publisher (which hopefully for the author is self-published so they can earn even more)…if your intent was to promote Amazon in combination with the book (which is what you really did), then yes there should be better means for affiliation3. If you flip the problem and think about it from the other side though…paying more than 5 cents per click quickly becomes a cost-prohibitive problem for all but the largest of sites/publishers (in fact, look at google adwords in any ‘popular’ category and you’ll quickly see that the small, mom/pop start-ups are quickly shoved out by larger brands willing to over-pay for those clicks to gain dominance).4. As a completely unrelated point…Disqus, just like Twitter, becomes a bit painful to read/keep up on for a hot topic (like this)…not sure how it could be better, but it *feels* like there could be a better way to summarize/group/present massive discussions like this…

    1. fredwilson

      i agree with all of your points but point 4 is an opportunity for disqusthat i’ve been eager to see them solve

  69. Andrew Wee

    IMO Amazon’s associates prog is a crappy example of an aff prog – poor 24 hr lifespan tracking cookies, non-existent tracking, low commission payouts. I can think of a number of other aff progs which will give better payouts and conversions per click.

  70. Lior

    Very well put. This is probably why affiliate marketing tactics have become so aggressive.

  71. Morgan Warstler

    Hmm. Fred, I don’t think this analysis holds much water…1. There are thousands of small permutations of of how I go about deciding to buy or not to buy and all that click tracking, and weighting of importance, is far too complicated – for a fleeting end result. Take Happiness as an example. Your credibility exists in a specific domain with me, that doesn’t necessarily carry over into how to be happy. The weatherman doesn’t convince me what car to buy. Online influence measurements bear this out (see stuff from Buzz Logic). So when you mention Happiness to me before some really happy chick who babysits my daughters does, who matters most in that equation? She does. Because she’s kinda freaky in her always being happy thing, and I now have to go read this book, to make sure she’s not a psychopath. She drove the sale. Figuring that out isn’t possible, just like pretending you have anything to do with other purchases I make from Amazon over the next two days. I know who Amazon is, I pay to be in their prime program, I buy shit there all the time. What did you really do in the equation, other than choose to sell the book link through Amazon because it decreases the sales friction – if you linked to some unknown sales entity, you’d make less sales.2. Just generally, examining real life, this reeks a bit of the guys who are always saying they are responsible for some deal you do because two years ago they mentioned that principal to you at a cocktail party. There’s some value there sure, but unless they specifically asked for a piece of a specific deal based on a specific introduction, it is outlandish.

    1. PhilipSugar

      I’ll give another real life example. Its like realizing that the hotel concierge is recommending a shitty tourist trap because they get a “cut”. How much was that info worth? The buck I slipped him/her. Certainly not the cut that makes the tourist trap concentrate paying for referrals versus paying for good food.So what has happened because you realize you aren’t going to get an unbiased opinion?? You pull up Zagat’s and now the concierge doesn’t even earn the buck.This story has been played out so many times with channels. Eventually the channel gets cut out.

  72. MikeSchinkel

    Shame, shame, as a VC you are ignoring The Golden Rule; he who has the gold makes the rules! :)Amazon owns the mechanism that makes it easy for you to add the link and get paid so they get the lion’s share of the value. It’s that very reason you fund many of the companies you do; because they have the ability to lockup value and gain an inequitable percentage of every transaction. As Kevin Kelley presciently said[1] back in 1998 “Them that’s got shall get.”Same is true with Google and AdWords. Do you think the blogger using AdWords gets the majority of revenue paid by the advertisers? If so, think again.That said, it seems like an ongoing opportunity ripe for disintermediation. As I once heard Michael Arrington say “The big opportunities are to turn a multi-billion dollar business into a a hundred million dollar business.” Something to be said for that.[1]…


      I hear the term “disintermediation” bantered around a lot.I get what Kevin Kelley says “them that’s got shall get”. But what is valuable to “get” is changing. When I have a website that has successfully promoted my book and an audience ready to buy and have a choice of Amazon or eBay or Google or Yahoo to provide the transaction/fulfillment/customer service – then I’m the “them that’s got” – even if the service is collecting the money. So if “disintermediation” is about more competition among the transaction/fulfillment/customer services I get that.But if “disintermediation” is about eliminating Amazon et al, I don’t think that will happen. No individual website with an asset to leverage is going to be able to get to a hundred million dollar business by creating all that infrastructure just for themselves.The void in the market is a new “intermediary” that helps individual websites connect with the most receptive people so they can do what they know how to do well instead of trying to figure all that out.

      1. MikeSchinkel

        What I meant by “disintermediation” is someone who will change the rules. Nobody is going to beat Amazon by playing the same game as Amazon. But if someone comes up with something new, and compelling, they can potentially disintermediate Amazon. That’s what “Disruptive Innovation” is, and it happens continuously especially around the web.15 years ago nobody could envision how anyone could challenge Microsoft. Today it’s obvious. Similarly paradigm shifts will probably occur that will have Amazon, eBay, and even Google in a future position that Microsoft is in today; with someone new eating their lunch. (It’s already happened to Yahoo. :)[1]

        1. PhilipSugar

          I think the disintermediation is not going to be Amazon that has to do the really hard work of having a warehouse stocked with items, filled with people, working complex machinery, and dealing with the public (like accepting returns) to insure you get one of the millions of times delivered fast.It will be with people that are trying to make money because they have info that Amazon might or might not have (i.e. who wants to buy the book)

  73. Dave Naffziger

    I think you make a great point about affiliate marketing in general, however I think your points aren’t valid by using Amazon as an example. The primary reason is that Amazon is the dominant vendor of books online. People inclined to purchase books online are far more likely to purchase them at Amazon.This means several things:- Your 535 views did very little to increase brand awareness or even brand affinity at Amazon. You didn’t even mention Amazon in the post.- Of the 40 buyers of the book, some percentage of users would have purchased from Amazon anyway, if you had linked them to some other website. Everyone has their payment details stored in Amazon, they know and trust the seller and their competitive prices.- In fact, fewer than 40 buyers in aggregate would have likely bought the book if you had linked them elsewhere.You can get better commission rates if you link elsewhere. But the final conversion rate of Site #2 will be far lower than the Amazon link (7.5%). The reason why so many affiliate marketers market Amazon products is that Amazon’s affiliate program simply pays the best per click (or most likely the best per on-site impression). I’d think that even if you sent users directly to the publisher or author, the conversion would still be terrible (think of all those crappy checkout processes at small merchants).What you really did is stimulate demand for the book. If you had used no affiliate link, Amazon would have still seen a boost in purchases. Perhaps, the real question is not what is the value of the affiliate link, but is there a better way to measure and compensate publishers that stimulate demand.

    1. Vivek Sharma

      Very interesting comment David. Good to run into you here. What do you about recommendations of non-book purchases? Amazon isn’t the leader here but they are on their way to becoming the online general store.

  74. rhberlin

    There are several good posts. I want to hold my one very short. I think we must first understand buying behavior. If we just take the difference between impulsive buying and extensive buying behavior, we must address that for extensive buying behavior it is really hard to figure out, who is responsible for the purchase. It could be that you did the whole research somewhere else and then go to a price comparing site… man it is the same in local shopping. For impulsive buying it’s different. You could address that this guy did really well by presenting the product, if the purchase comes from his site… so the problem is to rate the value added by the middlemen.A complicated but interesting model would be that the overall conversion is measured. The guys with the best conversion gets more percentage. Why? Because first if they go away it is bad for you and second they showed that they do a better job by presenting your product.

  75. Keenan

    Lot of good comments, here. A lot to learn, I will be back to get edumecated.All I know is I bought the book on my Kindle after reading your post on my Iphone. No way for Amazon to know that your post is the reason I bought the book.I really like the book, and will most likely post about it on my blog. Why shouldn’t you get credit for those purchases as well. Starts to feel a bit like MLM, but if the shoe fits.

  76. Paul Profitt

    I’m not sure if I’m repeating what somebody else has already said but. In the Affiliate Marketing circles that I frequent, it is a well know fact that Amazon pays crappy commissions. So I would attached the blame for your low commissions squarely upon their shoulders, and not generalize and blame Affiliate Marketing.You will get a much better commission rate if you use PayDot com or Clickbank, of course I appreciate the fact that they might not always have the product or service that will benefit readers of you’re BLOG…But the point is, that there are other alternatives available to you rather than Amazon.

  77. Steve Poland

    Fred, great post btw, as you know, I love affil (ebay-owned) pays on a CPC basis. I think this is maybe where things need to shift for Publishers — so instead of you linking to that book on, you link to a middleman aggregator (, pricegrabber, epinions) that shows a product page and lists the price of the product from multiple retailers. Those retailers then all pay a CPC, which they take roughly 40% of and then pay you. Books are currently $0.05 – $0.10, thus you’d get $0.02 – $0.04 per click — and of course, you’re putting your users to a landing page, in which they must then do another click so you get paid [as no cookie has dropped yet].This will give other etailers visibility, they will pay a CPC they want [market will set itself], and you should earn what is expected.Note: also takes into account a quality score.Second note: would have worked out better for you most likely. Yes, it’s a 24-hour cookie [and I do believe that cookie should stick for wish-list purchases], but you didn’t just earn commissions on that book — you earned on any purchases in 24-hours by those users of yours [including that Kindle, which I presume was the bulk of your $24 in commissions]. I almost want to say kind of works out in the end — it doesn’t have to pay on products you didn’t link to.

  78. larryadams

    Hi Fred, late to the party here, but I thought I’d add a few comments. I’m the product manager for Google Affiliate Network.Yes, there are alternatives to Amazon for books. Borders, Barnes & Noble, Abe Books (now owned by Amazon), Books-a-Million and I’m sure other smaller firms all have affiliate programs (disclosure, runs their affiliate program using Google Affiliate Network). Go to any site and do a control-f + ‘affiliate’ and you’ll usually find a link to join the site’s affiliate program.Evaluating the value of an affiliate link is different than evaluating the value of a display ad. Your link is an endorsement; this is a different kind of marketing than the skyscraper ad you’ve got running on the right rail of your site (which displayed an ad when I viewed the blog post in question). How many page views generated those dollars; what’s the effective CPM? Would that be competitive with a CPM from an ad network?It’s also not realistic to assume that an advertiser is going to pay you for every dime of value created by a referral. If they did, they wouldn’t see any profit. If Amazon didn’t have an affiliate program, would you have still linked to their site when mentioning the book?They are offering a commission to make it more likely that you’ll include a link to them instead of an alternative book seller. What if you used Kontera or some other alternative to create ads out of text in your post instead of using an affiliate link? Or what if you recommended a different book because you would have been paid a consideration? Figuring out the value of these other alternatives is really the way to determine if you’ve been underpaid for the traffic driven from this post.There are a lot of publishers out there that do exclusively affiliate because they’ve discovered they can make more on an effective CPM basis using affiliate links. High volume affiliates often request (and sometimes are granted) higher CPAs based on their numbers.As an aside, the price an advertiser is willing to pay for an affiliate conversion is dictated by just a few things:1. Margins (or lead value for non-retail transactions)2. What competitive advertisers are offering (i.e. what percentage does BN pay for a book purchase)3. (maybe, if the advertiser is sophisticated) Lifetime value of the consumer acquired4. (maybe, again for sophisticated advertisers) Offline multiplier-larry

    1. fredwilson

      thanks Larry. this is all very helpful

  79. AndreaF

    Fred, I understand that you, rightly, would like to get paid more for driving 535 visitors to Amazon and make them aware of the book. And I agree with you.However, I think it’s worth considering the following points as well:1 – There seems to be an agreement that Amazon’s affiliate program is one of the worst around so it’s more a problem specific to Amazon than a general one;2 – Most affiliate programs offer by default or can be nogotiated to an increasing %; i.e. the more you sell the more you get;3 – I am fairly certain that some of the people who purchased or will purchase the book following your recommendation on Amazon did so because they were already familiar and comfortable with Amazon; e.g. assume your link redirected to a less known retailer instead of Amazon; people would have more likely learnt about the book there and then gone to Amazon or Borders or B&N and bought the book; this adds complexity to the issue but also tells you that the book has been sold also because Amazon has a brand value and you are paying for it in the form of a smaller referral fee;4 – Your business is not to recommend books; if you were doing so on a regular basis I am sure your conversion rates and % fees would go up; potentially to the extreme that Amazon would end up advertising on your blog;5 – Many people have probably clicked on that link because they are interested in anything you, Fred Wilson, say or recommend not because of the book, out of curiosity or compulsion and are not necessarely valuable leads;6 – How much is Amazon actually getting for the sale of that book? versus the author and the publisher? Said that, I also think that the value of Amazon is well beyond that offered by its affiliate structure and although a better way of rewarding people like you who are opinion leaders must be found, the Amazon model, in general, is a valuable one and I can’t see a replacement for it anytime soon.A couple of ways in which rewards can be improved are in my view around the ‘Groupon’ model and through systems providing more relevance like ‘Hunch’.Something on the horizon is to look at the way SlicethePie and other start ups are approaching music; if that model develops for music and other types of content, it can give freedom to authors to publish anywhere they want. We are a few years away from that but not so much.Said all that, you should call Amazon and complain.

  80. James Cherkoff

    I think this is where MS was heading with it’s ‘Engagement Mapping’ idea with its central mantra that The ‘Last Click’ Model Is Dead. Not sure how far they’ve got with it. More here…

  81. PhilipSugar

    This is a great discussion I want to ask a really heretical question here:If the person going through the affiliate program was asked whether or not they wanted to pay the affiliate or get a cheaper price (and not have the affiliate know it was them) what would they pick? I think we know the answer….how much would I be willing to pay my friend who recommended a store?Yes, I will be told I will never link such an affiliate, but I’m trying to get at the root, which is how much does the person that is getting the recommendation feel its worth? Because I think the internet eventually is like all channels its like water moving to the lowest point which is what is the individual willing to pay, and that will then become what they pay which ends up being as close as possible to the cost to produce and distribute.Part of what the internet has done has made information which was scarce (and valuable) totally abundant (and worth very little) I was reminded of this when I worked for a large Japanese Trading company a long time ago. Their most valuable asset was their communication network (Telex) which became completely outmoded.Not sure but what if a search engine came out and said it was free or passed all the affiliate fees back…..I think that is what Bing is attempting.

  82. Will Hambly

    That’s about a 4% of sales, so it seems right. The Amazon Associates data was probably just not updated because they only recognize a sale once the item ships, not when it is ordered. Makes more sense now. Thanks.

  83. Sung Yoon Lim

    I think this is very true.

  84. sungism

    This needs to get monetized well.

  85. fredwilson

    so you are the person who bought the kindle i got affiliate credit for. wellthanks!