The Top Of The Funnel

Whether you have a e-commerce business, a SAAS business, a media business, a marketplace business, or some other business model, you are going to start thinking about customer acquisition at some point.

And there are a lot of options out there for acquiring customers; direct sales, indirect sales, channel, search engine marketing, social media, email, display, etc, etc.

But the best option, if you can pull it off, is to own an organic customer acquisition channel that is large and that sustains itself.

At USV, we have investments in a bunch of companies that have very large organic and sustainable top of the funnel customer flows. Many of these companies use a number of customer acquisition techniques, but they start with the organic channel and optimize it with their product development efforts.

Here are a few examples:

Codecademy – Codecademy offers a number of subscription learning services to people who want to learn to code. But because it has been offering free curriculum for learning to code for six years now, it was the first organic result I got this morning when I typed “learn to code” into Google:

Quizlet – Quizlet offers over 200mm study sets on the web and mobile to people who want to study and master something. No matter what you want to learn, you can find a study set on Quizlet to learn it. Quizlet is the “Wikipedia of studying” and because these study sets are free on the web and mobile, they have a huge organic flow of new users every month. Quizlet offers two subscription offerings to students and one to teachers and this organic flow is the primary customer acquisition channel for these offerings.

SoundCloud – SoundCloud is the first place most musicians go to post their music on the Internet. There are upwards of 200mm tracks on SoundCloud, the vast majority of which are free for anyone to listen to. This content is a huge attraction for listeners on web and mobile. SoundCloud has three subscription offerings, two for listeners and one for creators. The organic channel is the primary acquisition mechanism for these subscription offerings.

Kickstarter – When a project creator launches a Kickstarter project, they share it with as many people as they can through email, social, blogging, etc. This brings millions of backers to Kickstarter every month. Most of those backers arrive to consider backing a specific project and move on. But enough of them stick around to see what else is going on that Kickstarter has been able to build a large and sustainable business without any need for paid marketing channels.

Etsy – Etsy is now a public company and is no longer a USV portfolio company but I am the Chairman and remain actively involved with Etsy. Etsy is similar to Kickstarter in that sellers who have shops on Etsy are actively promoting their shops through various channels. Most of the buyers who arrive on Etsy that way purchase from the seller who brought them but some stay and shop from other sellers too. Etsy explained in it’s IPO filing that the vast majority of it’s traffic was organic. That is slowly changing but in Etsy’s early years, all of the traffic was organic.

I could keep going but I think you get the point. One of the things I look for in an investment is a free and sustainable flow of customers. This big top of the funnel may not be the only way a management team will choose to build their business but it makes a great foundation to build on and the LTV/CAC is infinite.

#entrepreneurship#management

Comments (Archived):

  1. awaldstein

    Community is the most ineffable and powerful force there is.That is what I read here and agree with wholeheartedly.

    1. fredwilson

      Yupppp

  2. PhilipSugar

    I would say this defines your investment thesis. Understand you can build companies in other ways, and I have and many other successful companies have.But this I think defines the core of USV Investment Thesis. Well said.

    1. LE

      Understand you can build companies in other ways, and I have and many other successful companies haveI think Fred is biased as many investors who have not been raised in traditional business in so many words. It’s a different world there for sure.I think also the saying ‘live by the sword, die by the sword’ applies.For example Fred’s example with codecademy includes search results and the result is based on a site ‘learntocode.me’ which codecademy doesn’t even operate. Want to see me get Laurence Bradford the (female) owner of that site to feature another company in the number 1 position? How much money do you think that would take? Sure I realize this is only 1 example and there are dozens of other things like this that they do for the funnel of leads (like any business) but the point I want to make is I’d rather have a formula that I control and pay for than one I get for free and have no control over.An example of this back in the day was a) yellow page ads (very expensive but worked) and b) newspaper ads and c) tv ads etc. High barrier to entry (payment and YP number of years got you better positions) and much harder to compete. You pay your money you get your results. d) Guerilla marketing. e) personal relationships that you reward to keep business flowing

      1. Vasudev Ram

        >Guerilla marketingJust googled for Guerilla marketing and this is one interesting link I found:http://www.creativeguerrill

        1. LE

          Not sure that site (or that page) is really in the spirit of what I see as guerilla marketing. For one thing the few examples (quick look) are all companies with large budget. So big deal and wow they are being creative. I’ve done guerilla marketing so I feel I am in a position to give a definition of it by the way. Not something I read (as always) something I do and have done (without input from anyone I might add).Guerilla mktg is really more what you (Vasudev) could do or what a small shop owner or small company can do on a shoe string. Of course so can a big company as well.Here is an example.There is a local cupcake bakery that opened that I went into the first week. They have these great cupcakes but nobody knows about them. They can’t afford to advertise and/or do direct mail whatever (which could be in some context guerilla marketing for sure). So I gave them this idea:I said “bake up cupcakes and go into the office buildings and give the receptionists free cupcakes for the office. Make sure there is something so everyone knows where they came from’. That to me is guerilla marketing. Not some stupid corporate fancy campaign that they sit for months working up. Something you can do just like that. Nobody is going to turn down free cupcakes or free food.Make sense? Guerilla marketing is just being clever and devious often. It’s not free but it can be very effective..

          1. Vasudev Ram

            Makes sense. But some of those bigco’s marketing campaigns in the link I gave, are good too. They probably spent a lot of money creating them though.

        2. ShanaC

          If you are looking for a great example of actual Guerilla marketing, Left Hand Brewery is a great model. Basically, when the company was very small, they gave out stickers of a Left Hand image, and left them in places where beer drinkers drank beer. Because the initial version of the sticker didn’t really have any context (It was literally an outline of a hand, no brand name, no context that it had anything to do with beer), it became a very searchable thing – which directly led back to beet. Plus, while they were doing this campaign, they limited their distribution – so the beer became very underground and elite. The mixture gave them huge buzz.That, and if you like milk stout, they had a very good one (I don’t, but that’s what they are famous for)

          1. Vasudev Ram

            Thanks, Shana.

      2. PhilipSugar

        We agree 100%

    2. Vasudev Ram

      What are some of those other ways – that you have used?

      1. PhilipSugar

        I have sold to enterprises and government for the last 25 years. You are going to spend a ton of money on customer acquisition. That is the bottom line.Have you heard the saying: “Do you know why it costs the military $500 for a hammer? Because it costs $450 to sell them one”It’s true. For the military you have to document where and how every single part was sourced, you have to agree that they can audit your books, you have to keep your books in a certain way.For large enterprises they want you to have SOC 1 and SOC 2 certifications. That costs a ton of money.When you sell to the consumer you do what you want.I am not saying that’s wrong. Frankly if you care about what you do all of those things that cost tons of money do just that cost tons of money. Not improve your product, but the reason you have to do them is the buyer cares more about not looking bad than spending money. And there are people that either don’t care or don’t know (same in my book) what they are doing.I am not going to call out names, and I certainly am NOT calling out any of the companies mentioned. But I certainly KNOW that there are many companies that play fast and loose with controls.No, I am going to call some out.Duck Duck Go is the result of Google playing fast and loose, see their FAQ’sA contractor can disable the President of the U.S.’s account on twitter? Let’s not go into politics, but you have so little controls?

        1. LE

          Well in the past we bid on GPO contracts and won a few. Here is the way it goes and at least part of the reason for the $500 hammer that exists in popular culture. The contracts were always rock bottom and there was always another guy who either would make a bid mistake or just be hungrier or be stupid. (In what we did I mean). But it was kind of like filler work for your employees to keep them busy. You rarely (the way we had things setup) could survive on all GPO work and government contracts.Now to the example: For the bids we do the government says ‘we want 100,000 of product A and 50,000 of product B and 3 of Product C. Only 3.So you give them the best price on A and B but on C you jack up the price. The contact goes to the lowest possible bidder. Right? So what you bid on C (in this greatly simplified example) won’t make a difference in winning or losing the contract.Now what happens is you figure they will need more of C and then you get to bill the ridiculous price for that C and make a huge profit to offset the money you are losing on A and B.Make sense? This actually happened. Figured it out on the spot and used successfully.

          1. PhilipSugar

            I’m assuming you are talking printing. I am telling you that for physical goods used by the military, it is not low bid.First you need at least a retired full bird colonel or higher on your staff. Then you need to jump through some serious hoops. For instance you have to explain and document where your titanium was mined (sponge and it wasn’t in the U.S.) then where it was cast into ingots (you can do that in the U.S., but it costs a bunch due to environmental regulations) and where it was milled (very expensive)I’ll give an example you can look up. The backpack costs over $1,000. Try to spend that on one in Cabella’s. The vest for the body armor costs $500 and that is a win. Try to spend that at Bass Pro. The actual armor plate? I have no idea, but I do know when we were testing one for fit and feel, they were so sensitive about that that we had a person bring it in the morning sit in the office and then take it in the evening.

          2. LE

            I agree and the same is true for medical devices or really anything where extra money needs to be spent (and costs are higher) to make sure the wrong thing doesn’t happen. Even if the chance is very very small. And yes companies play loose and fast with their goods all the time where failure is accepted and the burden is on the end user (company or otherwise) to clean up the mess.But all of this is why I get annoyed whenever reporters or peanut gallery talks about costs of hammers etc w/o taking into account reasons why things are. Remember the drone I bought with the servo with the plastic gear that chipped and almost flew into me. Why use a brass gear when a plastic one is someone else’s problem and you are in China and nobody is going to take the time to prove it’s your fault? On a larger scale (military) that is not the case. I had a head fall off a hammer once among other things as well.

        2. Vasudev Ram

          Wow, some solid points.Also:>Not improve your product, but the reason you have to do them is the buyer cares more about not looking bad than spending money.Yes, the corporate CYA syndrome, aka “No one ever got fired for buying $BigVendorName.”Been there, seen some of that.

        3. ShanaC

          it’s at least 18 months for just the sales cycle in enterprise, so I assume marketing takes another 6 months even before the sales cycle begins. All in all 2 years.

  3. pointsnfigures

    True in B2C, not as true in B2B since target markets are more defined. Depends on your market sector and what you are trying to do of course. But, B2B usually has more spend, longer sales cycles and can be stickier. I think this is one reason why blockchain and crypto are so interesting to me because network effects might get transferred into B2B marketplaces in new and unique ways.

  4. Joe Lazarus

    Free traffic is great. It’s also important to have some channel diversification for further growth and to protect against the free channel going away (ex. SEO algorithm changes). Etsy, for example, seems to do well with SEO but might be under utilizing paid performance marketing.

    1. Vendita Auto

      Always like to see where some company employee has secured 3-6 one two word core root domain channels 2 serve (and persists) the common elements of the chosen brand / aggregate platform, sometimes still possible to do when new tech word/terms become global cognitive/generics e.g. Asia/pinyin Cyber Anquan [secure/security] Suo [lock] insurance [Baoxian]. Keeps long term marketing at sustainable levels or pay the adword ferryman. Not sure how early stage AI algo’s will play out some much SEO is BS IMO. Avinash K is still the top man.

      1. Vasudev Ram

        If you mean Avinash Kaushik (GOOG or ex-GOOG), I thought he was about analytics, not SEO? or is he about both? Only read his stuff lightly a while ago.Edit: Never mind, I just checked him on Google – putting the info here in case anyone else is interested – he does both analytics and SEO and more:https://www.kaushik.net/avihttps://en.wikipedia.org/wi

        1. Vendita Auto

          Yep, top man, in my book analytics/data science 97% of seo one in the same

        2. ShanaC

          You can’t do SEO without analytics.

      2. Vasudev Ram

        Just saw this article by him, seems interesting, though not read fully yet:Dieter Rams: Ten Principles of “Good Design”https://www.linkedin.com/pu…

    2. ShanaC

      free traffic isn’t free. It is organic, but not free.The only truly free form of marketing is Word of Mouth, and even that, I could argue isn’t free, its cost are built into your product, and therefore the costs are not accountable for a marketing budget.

  5. Richard

    iPhone – Talk about a funnel! It’s literally starts in a baby’s stroller and morphs into a fashion device in junior high school.

  6. LE

    My sister (a printmaker by training (went to school in Rome)) has a shop on ETSY where she turns maps into wall art. I asked her if she would create a map for me to give as a gift to someone who had homes in multiple places. She wrote back (being the non-business person that she is) and said she was to busy with work (she has a corporate job to earn a living). So what I am wondering is why ETSY can’t monetize people who come to her by suggesting a similar shop that is taking on new work? (And give her a referral fee). Another thing I think she could have used is assistance with is pricing (she never asked me). She charged way way less than the value of what she created. I don’t think this made her art any more attractive to people and it very well may have been a factor in her deciding she had no time to do the art…. https://uploads.disquscdn.chttps://uploads.disquscdn.c

    1. Vasudev Ram

      That’s a cool idea – art from maps. Those look good. Can also be a conversation piece.

      1. sigmaalgebra

        Yup. Noted! I’ll add that to the initial database of my startup!

        1. Vasudev Ram

          Cool, but how is it relevant to your startup? Just asking, all I know/remember is that you said it was about search?

          1. sigmaalgebra

            “Search”? Can do some of that getting from the users keywords/phrases. So, can do a lot of that with a computer version of an old library card catalog subject index. Then can do better with the results sorted on some measure(s) of quality. And, then, can do a little more and be a little better. That is Google and its copies Bing, etc.But there’s a lot of, call it, content on the Internet. And the total is growing. That keyword/phrase stuff is good at finding only about 1/3rd of that content out there, that people want to search for, and that people want to find.That point is not new or even just for the Internet. Instead, decades ago people in “information retrieval,” e.g., at Battelle, clearly understood that keywords/phrases were a solution to less than half of the whole problem.For the other 2/3rds, there’s nothing any good.Why not? It’s somewhat like Henry Ford and the Model T: To essentially every horse owner in the world, something better was a “must have,” but nearly no one could think of anything better that was not just fantasy. They didn’t think of the Model T and, thus, didn’t ask for it. And relatively few, good other entrepreneurs tried for it.Same now for the other 2/3rds of search: A good solution for the 2/3rds will be a “must have” for nearly everyone on the planet with a smartphone to a desktop workstation with at least a simple Web browser, right, ballpark 3 billion people and growing. But likely and apparently essentially no one in the world knows how to get a good solution for the 2/3rds, except me.IMHO, current computer science, AI, ML is not sufficient for a good solution, does not even have the foundation for a good solution, and, really, is nearly irrelevant for a good solution; the computing needed is just totally routine.The key is how the heck to do it. Knowing that, writing the software is routine, trivial. How to do it, that is, what to tell the computers to do, needs some original research in applied math. That’s math where know what the heck is going on via some solid, at least partly original, theorems and proofs. For that math, need some advanced pure math prerequisites. So, for original applied math and the pure math prerequisites, essentially all of current Silicon Valley and computer science information technology entrepreneurship is unqualified and unable to deliver.Well, I did the math and wrote the code. Code looks fine, ready for production.But there is more to the technical parts: Need a case of a database of some of what is on the Internet.So, the maps as art should be in the data base.Why? A lot of what Google, etc. can’t do is let people search for art. Or, how the heck to search for art, e.g., artistic versions of maps as art, especially when don’t know that any such exists?Back up: The information retrieval people understood that for search via keywords/phrases, basically a user needed to know what content they wanted, know that it existed, have the content indexed by keywords/phrases, and know what keywords/phrases accurately characterize the content. Whew! That’s asking a lot of the user and why keywords/phrases are good for only about 1/3rd of content, searches, and results.So, e.g., the idea of art from maps has essentially no hope of being found via a keyword/phrase search. Similarly for essentially all of the poorly served 2/3rds.So, a little more broadly, part of serving the 2/3rds is letting people find “art”. Keywords/phrases to accurately characterize art? LOL!!!!More generally, what is needed is success with “meaning” of the content, e.g., the artistic “meaning”. Similarly for essentially all the content, all 100%. Mostly what people want is the meaning. For about 1/3rd, they can get there via keywords/phrases; for the 2/3rds they can’t and need something else.I know; I know; “meaning” is beyond syntax and semantics and long a Holy Grail problem in text processing, natural language understanding, computer science, AI, ML, etc. Right, long a problem. But the computer science, AI, …, people are “digging in the wrong place”. They are welcome to keep digging, using GPUs, TBs of data, etc., but at best they will get just crude solutions for just tiny fractions of the problem and there mostly just by accident. In simple terms, they don’t know what they are doing. Actually they are being driven by some poor understanding of some old, simplistic statistics, really just curve fitting, and some simplistic, naive, poorly informed, weak intuitive ideas. That’s just all they and essentially the whole computer science community has. As a field, computer science is short on powerful methodology and at the end. Sorry ’bout that!Well, I found a solution. My work is original applied math. To call my work computer science, AI, or ML would be a grand insult and would drag my work down into the hopeless sewer with the rest of that computer science, etc. garbage.My code works. Currently the database I have is small and just for testing, but for what is in that data base my software works great. But before going live, I need a larger database. Soon I’ll pursue an effort to get that additional data, but when by accident I discover something really nice, like art from maps, I make a note of it! Thus my post!BTW, my work is only for the “safe for work” content; I’m not going for porn, wild political stuff, wack-o, moonbeam stuff, ET stuff, etc. Art from maps? Terrific!Actually at least initially the data in my database will appeal to the BMW, MB, opera, symphony, Chablis and Brie, finance, ski vacation, etc. set — so I’ll have some high end ad “demographics”.And, sure, I’ll be able to get good ad revenue only from the, say, 1 billion users in the more advanced countries. Okay.But, I’m going for the poorly served 2/3rds. So, let’s, see, 2/3rds is twice 1/3rd. Well the 1/3rd is worth ballpark $500 B. So, first-cut, back of the envelope, ballpark, the 2/3rds should be worth $1T, the first such.All the work has been fast, fun, and easy for me. The software is totally routine and dirt simple. I wrote no prototype code and no MVP. Instead, the first version I wrote was/is solid, scalable software ready for serious production, say, to $10+ million a year in revenue, with nearly all of that pre-tax earnings. The code is well documented, easy to understand except for the core applied math parts, and doesn’t need “refactoring”.But there have been some independent, unpredictable external interruptions. E.g., today I’m working with system management mud wrestling to get around some damage from a silly hardware problem — I didn’t actually lose any data. That mud wrestling has everything else on hold. The mud wrestling is stupid, but I’m getting past it. There have been other cases of stupid mud wrestling, but I got past those, too.So, yes, my work is in Internet “search”, but my work is nothing like Google, Bing, etc. And in the 1/3rd where Google, Bing work well, and they do, my work is not better. I’m for the other 2/3rds. If I were at Google or Bing, I’d be laughed out of the building. All the “suits” would see just two cases: (1) I was right and their careers would be at risk from competition with my good work. (2) I was wrong and any suit who backed me would get a black mark on his record that could hurt his career. So, with my project, the suits would see only downside and no upside. And that situation is quite general in large organizations and why they will be very reluctant to compete with me until too late.Also, the only way the Google, Bing, etc. organizations would know to try to compete with me would be computer science, AI, ML, Norvig, Ng, etc., and computer science and those guys don’t have “the right stuff”. If I sent them my resume, they’d throw it in the trash if it ever reached them but it wouldn’t, would be in the trash long before it reached them.There’s nothing I could to do to get any reaction from them except just laughter, and getting even laughter would be tough. Yup, that’s part of my opportunity. Bluntly put, clearly enough, my background is better than theirs. So is my work. They don’t have the right stuff, are digging in the wrong place, in a Henry Ford analogy, don’t imagine that there could be anything better than “a faster horse”, and wouldn’t know how to do the Model T. That’s all part of my opportunity.Those guys are very short on, call it, project methodology. All they know is just some simplistic computing and some large scale empirical curve fitting and otherwise, due to their bad project methodology, have no ability to separate solid work from moonbeam nonsense. So, to them, anything not like what is in the Stanford, Berkeley, CMU, or MIT computer science departments looks like moonbeam nonsense.High irony! The good project methodology is NOT in computer science! There’s been rock solid STEM field project methodology back at least to the Manhattan Project — low risk, high payoff. The SR-71? Low risk, high payoff. Keyhole? The same. GPS? One more.One of the very best parts of good STEM field methodology is appropriate applied math theorems and proofs. Computer science is very short on those; the profs rarely even know how to write solid proofs even for simple, old things, even for exercises in an undergrad pure math text by, e.g., W. Rudin — literally.

  7. Ronnie Rendel

    Obviously having large organic “top of the funnel” customer channels has a lot to do with the core “idea”/problem/market the product is solving. That’s why we spend months on the problem and market before writing any (production) code.With that said, each of the companies above is very different in their market and HOW they acquire this organic customer flow. Some are based on brand, some on Google keywords, others on content marketing.

  8. VARUN KASHYAP

    Fred, from all the examples that you have given (or atleast the way you have articulated these examples to draw home the point), it seems that one needs to have 200 Mn study sets (Quizlet), 200 Mn soundtrack (Soundcloud) and so on to build an organic ‘top of the funnel’. Now, none of your followers would disagree with that. The point is how do we reach those numbers when ‘organic’ happens ‘organically’ or is there some way that these companies were able to crack this when their numbers were few tens of thousands of soundtracks, study sets and so on.

  9. Salt Shaker

    Source mix is obv impt to profitability, w/ organic having lowest acquisition costs. Even the companies you cite likely didn’t achieve those gains exclusively by organic means. I’m sure they’re continually testing paid performance, and then testing some more. If your biz is ad rev dependent, then you need to drive traffic, sometimes w/ customer acquisition at a loss, to drive scale and achieve rev growth on the back end via adv. (Not a particularly healthy biz model these days, particularly w/ GOOG and FB dominance.) Organic sources=healthy margins.

  10. Steven Kane

    Respectfully, to use the word “free” is a little misleading. All of these are great companies with viable products and businesses. But none survived and prospered because of “free” customer acquisition. Their “free” offerings cost them a fortune which is subsidized (paid for) by venture investors. Absent large repeated infusions of outside capital to cover these acquisition costs, most if not all of these ventures would not survive long enough to get to actual operating solvency.I’m not being critical. I wildly emphatically approve. Grateful venture capital exists to do this. Just feel it’s a bit unfair to young day zero entrepreneurs to inadvertently suggest the ventures you describe do not have to spend money on acquisition…….

    1. cavepainting

      Organic customer acquisition is “free” relative to the options of paying for keywords, app installs, etc. For many web based B2C businesses, paid traffic acquisition can be unsustainable in the long term. (although there are some exceptions). Web or app businesses can use VC capital to jumpstart acquisition, but if they cannot figure out an organic source soon, (esp. when CLTV is not high), it is game over.

      1. Steven Kane

        All true. But doesn’t refute my point. “Organic” is just a fancy word for “paid for some other way.”If I spend $10,000 making a video, and it goes viral, how is that “free” traffic? Didn’t it cost me $10,000? Giving away “free” goodies, whether content or lessons or whatever, is a superb strategy, but it’s only “free” to the consumer, not to the company. Not spending money explicitly on marketing and advertising does not equal not spending money on customer acquisition.

        1. cavepainting

          Yes, fair point. True that “free content” should also count as customer acquisition costs. But costs of developing content and buying traffic have different characteristics. The former has zero marginal cost and more leverage; the latter has a significant marginal cost.In many verticals, investing in content and distribution strategies can substantially decrease the cost of traffic. At the end, irrespective of how the distribution of CAC breaks out within content and traffic, what the customer pays over his lifetime has to be higher.

        2. panterosa,

          I appreciate you insisting on what the costs are to the business for the “free” thing the customer receives. I do not spend on advertising, but give samples of product away. They were not free for me to produce, so does that now suggest I claim them production cogs percentage as advertising? Just the internal P&L categories and tags can be smushed around enough to claim many things which aren’t true, but neither are they false. As a self-funded small biz, accounting is a place where our intention is better expressed in a truthful financial report.

          1. PhilipSugar

            This is very true. The key to accounting is to “account” for what is being counted (spent).I had a board member, an old sage, tell me once. You can lie to me, for all I care you can lie to your mother, but never lie to yourself.So yes those samples are not COGS they are a marketing expense. Nothing wrong with that, great, but account for it.

          2. panterosa,

            Totally.And hence the myriad accounting styles and methods thru which obfuscation arises.I joke that it’s not possession which is 9/10 of the law but intent. Let me now apply that to accounting.PS as someone in the process of a 3.5 year audit (totally pointless (since it’s revenue zero) for the IRS, and very expensive for me), it’s this stylistic difference of grey areas which people mistake routinely. If I advertise by samples then that’s how I advertise.

          3. PhilipSugar

            See my comment above. The one beauty is that I know if you are willing to wing it, you can have very, very low costs.

          4. ShanaC

            I’d actually informally track those costs vs how many customers it gets you. It’s a useful metric for when you want/need to expand.

          5. panterosa,

            Actually, more than conversions, those samples “bought” us reputation as top design and top concept as total newcomers. In a field that is a sea of crap, it meant we could call on anyone we wanted for a meeting, and quickly met all the top brass. When you do another product, and give it away again, and it rivals the first you gave any of those people, then you now have followers. When I presented the second big idea at MIT, someone I had courted 2 years before came to congratulate me afterward. I am about to cash in on that reputation with the next release. This makes me think my conversions in this case were b2b more than b2c. This will make sense when the thing I’m doing now is finally released.

          6. ShanaC

            That makes sense for the business you’re in. Many games have a distributors as their consumer sales point, not the company itself.I would still track those costs though, some b2b sales are worth more than others.

        3. PhilipSugar

          I think your point is correct in the precise sense, which I am not dismissing. Although the beauty is that the cost of computing and networking has come down so much it is “nearly” free.Let’s take Quizlet (which is great). The difference in cost between providing service for one person (yourself which he did) and ten million is the cost of hosting. You are 100% correct in that the cost of hosting is not free.I don’t know the numbers but for simple but let’s take a scenario: You buy a Google Fiber connection which is $70/month for gigabit up and down speeds. You buy a Dell PowerEdge T640 with multiple processors 8 TB of disk space and 4 TB of SSD.That is going to set you back about $12k but if you finance for 60 months that means about $200 a month, and you can start off for a ton cheaper than that with only 1 TB of disk space and 500 Gig of SSD.Let’s add in other stuff. $300/month total. In your parents basement.Now we are not counting your time. But you built the app for your own love.Now is any corporation going to let you serve them “bare ass to the wind” like this?Never.But by my calculations you could run easily way over 10mm users.But what if you go down? People are going to sue you…….for the money that they paid you……ummmm nope. Kind of like the fail whale.Now eventually if you are lucky (and it is like hitting the lottery) you start saying I need to do, SEO, marketing, actually host in a redundant datacenter, more features, etc, etc.In which case you are right there are a ton of costs for those “free users”My point is that if you are willing to go “naked” and you can do your own programming? You never need to take a dime.Now start off with a plan get VC’s to back initially, then the costs go way up.Same as selling to a corporation.But we fundamentally agree that giving away something free to get users is not free, just depends on the cost.

          1. PhilipSugar

            Correction. If you are successful eventually you do need to take a dime to extend your lead so nobody else comes in and copies you, but puts in more money to catch up.

        4. ShanaC

          So technically its not that it is free, it is that it asymptotically approaches free so closely that it becomes effectively free.Let’s take that numberLet’s pretend your 10k video, on average, should see 10k people of which 1k convert. so each customer is $10Now lets pretend the video went very viral for some reason, and instead of seeing 10k people it was seen by 100,000,000 , of which10,000,000 converted. Each customer now costs 1/10 of a cent.The goal of really great marketing is to get to the second situation (and lower) all the time. The better your marketing is, the broader the appeal, and the easier it is to convert someone, the closer the asymptote of your marketing costs approaches 0. The problem is when you are smaller, these numbers are VERY uneven and it makes it VERY difficult to walk that asymptote. Not impossible, just difficult.(Or, in other words, despite the fact that Coke spends millions on advertising Coke, a lot of the actual marketing per coke bottle sold is close to effectively free.)

          1. Steven Kane

            That’s a distinction without a difference.My example was clearly not meant to be illustrative of reality.And using Coke as an example is kind of crazy — if ever there was a textbook case of a corporation or product built almost entirely on gargantuan outlays for marketing and advertising, its Coke. The reason its been such a lucrative model over the years is not that marketing is incrementally nominal, its that the COGS of a bottle of Coke is near zero. It costs Coke nothing to make a bottle of Coke, maybe a penny or two, and they charge us $2. They can spend $1 bottle on marketing.Google spends 20% of gross revenue on “traffic acquisition.” YouTube’s service cost many many hundreds of millions of dollars of capital to provide “free.” I’m not saying this type of new venture financing is bad. It’s frigging awesome! Thank heavens it exists.I’m just saying that it creates a false impression in the world of newbie entrepreneurs and newbie ventures, which, intentionally or inadvertantly, is a help to venture investors who get to own much more of the deal much quicker and cheaper — that in the end, there is almost no such thing as “free.” I regularly, maddeningly see and hear this magical thinking in my work in various accelerators and startup incubators. Starting new ventures, making products, getting customer attention and adoption almost always take a ton of time and yes, money. Bootstrapped, friends and family or VC. But newbie entrepreneurs so often get squashed down on the cap table by underestimating how much capital it will take to get a venture up on its feet. Then they get all pissed off when the investors turn out to own the deal and show them the exit.Btw, I’m not criticizing Fred or USV, who I know from direct experience over many years do not play this game. But many investors do.Btw, btw, even venture investors get blinded by such magical thinking. How many ecommerce ventures do we see every day going down the drain by giving things away “free” or heavily subsidized, believing that traffic and volume will allow correcting of the basic P&L later? Sometimes that works brilliantly — Uber, Amazon, Jet — but in most cases it just leads to the realization that the fundamental model was unworkable…

          2. PhilipSugar

            We agree completely especially about Coke and getting shown the exit if it goes slower than expected. (I’d say it does for all but maybe 10 startups in the U.S. per year) See my response on how much Coke spends.

          3. PhilipSugar

            Coke does not spend millions on advertising. They spend $4 Billion.They also spend much more than that on pourage fees. If you notice when you restaurant or Stadium you don’t get a choice. For instance at UofD they pay a big Sports Bar about $50k per year, then they discount the product as well, the account for that in COGS not advertising.SNL did a famous skit about it.http://www.billygoattavern….

          4. ShanaC

            Pourage fees are considered part of advertising fees, and that 4 billion number is the worldwide number, not just US.Though even with that four billion number, 1 serving of a coke product (not just coke) sold is about $0.0005 in advertising costs.

          5. ShanaC

            poopoori is totally work safe – and they started a trend with these types of ads

      2. LE

        The point is that without an organic channel driving the majority of leads into the top of funnel, it is unsustainable in the long termI can’t say I agree with that (and I typically agree with your points).For example I will take a business where I can send out a direct mail piece (only one example) that costs $2.00 to print and mail and results in $30 of sales or $100 of long term sales whatever (arbitrary, made up). How is that not sustainable? If anything (to my point elsewhere today) I like that better. Because it’s a repeatable and expandable system.I’ve done this. Postcards go out, leads come in. More postcards, more leads. More leads more sales are closed.What you don’t want to do is be dumb about it. For example Indeed.com (former USV company) sends me direct mail pieces frequently. They contain the names of people that don’t work here and never worked here. This is actually quite regular. So they are throwing out money with clueless things like this. Unless that is some new strategy. But I don’t think it is by the design of the piece (looks like junk mail).

        1. cavepainting

          You are correct. I really meant (and should have qualified) it in the context of consumer web based businesses where companies tend to indulge in paid advertising with the hope that the economics will eventually work out. And in some cases, they do (like justanswer) where average CLTV> average CAC and they continue to drive traffic through adword spend. But in many cases, it is not sustainable as only a small % of paid traffic converts and the CLTV never catches up to the ad spend.

    2. fredwilson

      Etsy and Kickstarter were profitable very early in their life. You make a good point but you are wrong about some of the examples I cited

    3. fredwilson

      Quizlet was bootstrapped for ten years without a dime of venture capital. it built its business organically and without subsidy. not every company needs VC to get going. many of the best don’t.

      1. jason wright

        a growing trend

        1. awaldstein

          how so?bootstrapping is a bitch and in most areas not possible at all.

          1. Drew Meyers

            Atlassian was bootstrapped. Had no idea until listening the the recent podcast episode of Acquired.fm

          2. ShanaC

            all my best ideas are not bootstrappable. I suspect many people’s are like that. Really a pity

  11. JLM

    .In the chicken v egg sweepstakes of founding a company, a founder has to envision the product and the customers for the product in the same paragraph, maybe the same sentence.There is no product if there is no customer. If there is no customer, there is no company.The cost to acquire a customer v the long term value of that customer is the financial spine of the business. The skeleton of the company — the delivery methodology for the product — is built around this spine.Venture capital is the bridge which allows the new company to cross the chasm from idea to profitable enterprise. In that manner, the essential question is:How much funding does this company, producing this product, require until it has sufficient revenue producing customers to sustain its own growth?When a company bridges the chasm and is able to operate profitability, then the challenge becomes one of scale, growth, rather than survival. A good problem to have.One of the first things any founder should do is to create a business engine canvas which touches on the issue of defining the product, the customers for the product, and the revenue stream from those customers.It is impossible to delve into the possibilities of success for a startup without taking a look under the hood and poking this basic info.JLMwww.themusingsofthebigredca…

    1. sigmaalgebra

      Yes, envision the products, the users, and the customers … plus for the product the new results for the users, maybe some new data, and some new processing of that new data, to get the new results for the users, and some more new data and its processing to get some more new results for the customers, the money paying kind, the advertisers, right, all in one sentence, thought, little Tiffany’s blue package.Right: On the Internet see a Tiffany’s ad, maybe just by going to Tiffany’s (e.g., to get her a synthetic, perfect, round, brilliant cut, light blue, 2 ct sapphire in a white gold, six prong Tiffany setting!), display the ad on the screen, use a pixel color picker to get the three 8 bit RGB values, in something simple, say, old first version of Microsoft’s PhotoDraw use those RGB values with many more pixels on a Web page to add a subtle, unconscious sense of prestige to the page! Ah, no one would do that!!!! Well, maybe some nerd, like 25% of males, partially red-green color blind, with zero sense of color would!

  12. sigmaalgebra

    Good.For me, the good news is that in some cases the “top of the funnel” can be “organic”. Clearly there have long been examples, but it’s good to get more examples and more understanding of how “organic” can be made to work.My collection of stuff on publicity for “organic” growth currently has 92 articles, each indexed and abstracted and all of them easy to search, but I’m still short on how to get organic growth going.I’ve got contacts with some media writers who specialize in the tech industry where, to borrow from an old movie, I want my story told and they very much want a story to tell. Still I want more. Besides in school I didn’t specialize in story telling.There are the old ideas of the super salesman who had business cards printed by the thousands and (A) at big public events went high up and when everyone was standing and screaming threw out the cards by the hand full and let them flutter down and (B) in various retail locations happened to lose stacks of the cards on counter tops, etc.So, there are other ninja, guerrilla marketing, publicity tricks, stunts, hacks, etc. And there’s much more besides.Still, I want to know more.

    1. ShanaC

      Short answerThe fact you have 92 articles written by you isn’t so helpful if no one reads them. Content needs to be marketed, so people READ it.You need to go further to the top of the funnel.

      1. sigmaalgebra

        Shana, Shana, Shana, your post interrupted my re-watching of the old movie The Big Short, a fun and funny movie!Yes, it’s a movie about laughing in a cemetery while surrounded by the gravestones of all the people the CDSs on the CDOs of the B tranches of the MBSs got to jump out the windows of Wall Street investment banks or some such.Actually, I’m not too surprised at the 2008 crash; people, organizations, and systems are like that. So, watch the movie, that the movie makers and maybe also Michael Lewis wanted to be funny, and get another example of reality and enjoy it!Likely Dutchess County Community College can give remedial courses on reading and writing. It appears that I need to take the course on writing or you the one on reading?E.g., for your The fact you have 92 articles written by you Shana, Shana, Shana, I didn’t write ANY of those articles and didn’t claim to.Instead, I found the articles, one at a time, mostly by chance, over many months, ALL written by others, that seemed useful, and READ them, kept them, and indexed them.I may have learned some things, but IMHO not enough.You have me suspected of doing something bad, accused, tried, convicted, and sentenced, all in absentia, based on all total nonsense data, when I did, in one word, NOTHING like that and NOTHING wrong.Ah, yet again we see the problems with Bayesian statistics and confirmation bias and yet another way to discover, to borrow from the movie, where a bull went number two!Gee, you have me in the slot of the Dr. Michael Burry in the movie, wildly misunderstood!I need to watch a little more of the movie and then type in some code to solve a computer system management problem I have. I got a lot typed in last night but want to continue now!

  13. cavepainting

    There are a few core principles to drive top of funnel for any network or marketplace that can eventually lead to a self sustaining organic flow of traffic. Not all are applicable in every situation.1. Channel Distribution: i.e. participants have incentives to use traditional social media, messaging, publishing, and email channels to drive traffic to the platform. Tools or features that make it easy to do this can go a long way (like what twitter initially did to help publisher embed tweets).2. Free Content: Abundance of good free content drives high SEO ratings and word of mouth leading to high traffic; a small percentage converts to premium through careful nurturing.3. Content creation toolkit: Aforesaid content is created by participants on the platform using native tools. This makes it more scalable.4. Referral Incentives: i.e. early participants have significant financial incentives to invite others. (ex: Uber and Lyft drivers making hundreds of dollars for every referral who converts to a driver)5. Built-in Virality: This is a hard thing to achieve circa 2017, but if the core value of the service is linked to sharing with others, well thought out product features that enable this seamlessly can make a huge impact. Of course, without sustained value from the use of the platform, acquired users can also drop off very quickly.6. Influencer distribution: Key influencers can drive an extra-ordinary amount of traffic by mentioning your platform on their web sites, podcasts, blogs, etc. Influencer marketing efforts can deliver outsize ROI for organic traffic.

    1. sigmaalgebra

      I’ve been considering versions of 5, 6.

  14. Chimpwithcans

    Apple Music/iTunes is attached to the iPhone – yet another binding agent for the iPhone community – helps sell iPhones. I don’t see the same for Soundcloud. Huge Community – but where’s the money spinner? I love Soundcloud but I worry about its sustainability despite its large customer funnel.

  15. JamesHRH

    @fredwilson:disqusThere is just a HUGE inaccuracy / overstatement in the last paragraph.None of these businesses had a free and sustainable flow of customers when you invested, they had a free flow of interest and a highly educated guess that it their ongoing promotional giveaway would turn into customers.After all, a customer pays with money not attention.They did also have a strong belief that they could plug the least amount of product into that interest and capture the lion’s share of the customers by after building a dominant, if somewhat expensive, presence in the market.It is, essentially, a way to prematurely age a market, cut a ton of margin out of the segment and freeze out competition. And it relies on high risk funding to execute this strategy.I don’t doubt you looked for it and that its worked. But, it only works in incredibly low cost products like B2C software.Its not like you see Tesla giving away cars to create a free and sustainable flow of customers. The more you say that sentence, the more it makes no sense.Although, Uber did that early on and its a great market entry promotion tactic.Heat check Pards.

    1. PhilipSugar

      See my comment, I don’t think it was a lie. You have a ton of people using your service that you provide for free because the total cost to provide it is less than one season football ticket. Fred finds it under the radar, partly because people come to him, then you lean in and provide a ton of money so somebody doesn’t copy and you lock in the space because the more users you have the more valuable you are.

      1. JamesHRH

        I had edited out that wording before I saw your comment.Young entrepreneurs could be led astray with this post, without the economic context.Not Fred’s best. Watching the Giants game w one eye maybe?

  16. Brook Shepard

    These are all businesses who fit naturally with an SEO angle. Which maybe, you know, gives one an idea of what USV is looking for in an investment.

  17. alexanderjarvis

    ‘Organic’ seems fungible for ‘engineered virality’ here.Not much has been talked about this area since Skok (Cycle times) and Chen (jumping the shark etc).I’ve tried to critically analyse how you engineer virality here https://www.alexanderjarvis…But my main tenant is that other than in gaming no one actually goes viral by definition (mine adjusts for churn).

    1. ShanaC

      there are methods for gaming virality. Example, understanding how actual viruses use viral reservoirs, and understanding how they mutate, one could game the system of chance of virality. The same principles works with people. (I’ve personally wanted to see if I could swap out kale for seaweed using this sort of method for “foods that are trendy/healthy”)Or, guesstimating linguistic shift to get ahead of google’s data set by looking closely at how relevant customers talk on Facebook and Twitter.The larger problem is these are very hard math problems to model correctly and built off great executions, and much like viruses don’t really appear out of nowhere, nor do advertising plans, which cost money.

  18. ShanaC

    Pro-tip – the top of the funnel is where CRO is really at. If you don’t optimize well there, everything downstream will go badly