MBA Mondays: Revenue Models - Data
The Internet is a data generating machine. According to Eric Schmidt, every two days now we create as much information as we did from the dawn of civilization up until 2003. It's also incredibly good at presenting that data, both to humans and machines.
So it makes sense that collecting and publishing data is one of the primary business models on the Internet. Here are some of the examples that you all created on the revenue model hackpad:
I like to think of data businesses in two categories; businesses that aggregate and then publish data and businesses that generate their own proprietary data by virtue of the service they provide on the internet.
Most of the companies listed in the data section of the revenue model hackpad are businesses that aggregate data from others and sell it. These can be good businesses but they are rarely great businesses.
Google is an example of a business that generates its own proprietary data by virtue of the service they provide. Google doesn't monetize with a data revenue model, they monetize with advertising that is targeted based on the data they generate. But in many ways, Google is a data business. Data is the secret sauce of their business and they have invested heavily in data science to maximize the value of their data.
Facebook and Twitter are rapidly becoming data businesses like Google. They collect a ton of data about users and what they think about and care about by virtue of providing a free and valuable service on the Internet. And that allows them to improve their services, make them smarter, and to target advertising to their users.
Going back to the aggregation model, if you are going to pursue this approach, try to figure out how to make your data as proprietary as possible. Anyone can aggregate so you run the risk of commodification in the aggregation game. If you can create some sort of proprietary advantage, either through exclusive access to the data or through some sort of refinement of the data using your own insights and analytics, that leads to a better aggregation type business.
Most data businesses are subscription based, but data can also be sold on a transactional basis. Transactional models are easy to sell when you are just getting going, but subscription models work better over the long run.
Many data businesses use APIs to make it easy for their customers to get data into their own systems. This is a good idea because it makes it harder for customers to leave if your data is part of their systems. If you can make your data part of a broad ecosystem, that is a good thing.
Selling data is a good way to build a business on the Internet but if you can figure out how to leverage proprietary data produced by your service to make your service even better, that often turns out to be an even better "data business".
You need huge scale though to make your data meaningful and scale is a bigger and bigger hurdle.Question: Facebook. Twitter. Google as you mention got scale, got data and are all media/ad companies. They don’t sell anything or connect anyone through a transaction.We all use the data we collect from our markets to be smarter but examples of companies that sell the data not use it to drive media sales that are less than say, five years old?
Many data companies do both. Examples that fit your criteria, that I can think of: AddThis and ShareThis. They collect massive amounts of data by virtue of being on publisher pages – to provide the sharing buttons – and they use that data themselves as well as sell it to others in the ad eco-system.
Thanks Jim (w/0 an avatar it seems!).I’m a data sleuth big time I’ve just never built a business whether I sold it.
I was unknowingly logged into another account I have when I made that first comment.
This is one of the cloudier parts of the Adtech eco sysytem….What happens to data that is collected via trackers installed on a site? How many other members of the ecosystem get a look at that data? The daisy chain of trackers is absolutely astounding and can be quite revelatory to anyone who makes a living off web data. Evidon Encompass powered by Ghostery.com lets people see this Invisible Web of trackers. Very powerful stuff, if I say so myself.
Agreed. Ghostery is cool stuff.
We would be one example – but an exception that validates. We sell data back to the generator with added value and no added advertising in a B2B context. We notice that we dont “fit” a lot of standard business models.
Thanks James.If there is anything public that deals with the economics and the pricing of data as a product, do link me to it.
R.R. Bowker might be a good example…they are purely in the data business (publishing data)…Amazon and Bn.com were both started by lic. the data from Bowker…Amazon built a transactional system that added a ton of value around the initial data and has never looked back…Bowker remains profitable (I think) but small and niche (and still struggling)…
Thnx, new2me.I know its a generalization but you have to wonder if the data they sell spawns monsters companies but they themselves can’t grow as the provider.
Yes – I think it’s a great example of missing the big opportunity because you are blinded by (shorter term) profit…I suspect this is a realization that a number of recent ‘data’ companies are also starting to wake up to and it will be interesting to see if/how they evolve (ie. bit.ly, chartbeat, klout, etc.)
hmm – why huge scale? it seems to me that 30k frequent users of a cloud based product would generate meaningful data
Guess….but we aren’t talking about meaningful data. We are talking about gathering data and making a meaningful business with it.Got to be somewhat difficult to gather, understand and parse into useful streams to sell for it to be productized and paid for.Where can you gather data that can create a growing biz model from a limited sample that is broad enough and competitive enough to sell broadly?That was my thinking.
ownership of data.did the D of J go after Aaron Swartz to protect the legal right of private corporate ownership of the data we all generate?
TBD WTF The DoJ was thinking.
If you don’t want to give a corporation your data, don’t use it’s services.You’ll have to delete your Disqus account to achieve that.
i try not to, but it’s not easy. i don’t use FB for this very reason. i’m suspicious of GOOG (i won’t use its email service), and i think eventually i’ll ditch Chrome too.the idea of shared ownership with permissions appeals to me, where i get some choice over how my data is mined and who else gets to see it. i think a lot of people would go for that, and perhaps even pay for it.
App.net is a great experiment in that. I personally like being tracked and appreciate the benefits that free web services give me. Being able to surf the web anonymously (e.g. “what is X cancer?”) is just a click away to a different browser.
Bloomberg, Thomson Reuters, and many others that cater to Wall Street are aggregators of data. They do “value-add”, but the raw data comes from stock exchanges and financial institutions. Some of these market participants have done better than others over time. 😉
That’s me ^Not an impersonator. Logged into another Disqus account.
“make your data part of a broad ecosystem”- that’s the money quote. No matter whether it was aggregated, collected or derived, that data must be valuable for someone else.Selling data is like selling a product. Its benefits must be explained. It has target segments, and it must be sold because it’s largely a B2B thing.Take away: Give a service for free so you can collect a ton of data, but use that data to sell another service.
William >>Give a service for free so you can collect a ton of data, but use that data to sell another service.This is exactly our plan when our subscription services see competition.
Exactly. I thought about your service when I said that. Same with Google, searching is free, but being part of every search is not always free.
Being part of every search is not free–?How does Google benefit directly from SEO other than driving more and more consciousness around search and it’s offshoot SEM?
I was thinking about “purchase intent”. That’s the rub with Google. Their whole SEM mantra is based on purchase intent. They are delivering the goods to the advertisers,- otherwise they wouldn’t be collecting billions of dollars from it.
SEOs create content which is monetizable through AdSense, in addition to AdWords (since paid results need organic results to be shown around…for now at least)
Adsense–true.Forgot about it cause I’ve never liked it very much.On your other point….SEO is a platform for PPC. Smart.
It is crazy how many possibilities this creates.
How does it have targeted segments? The problem is a lot of data being sold is of the advertising or the finacial type. There is no reason why you couldn’t aggregate electrical meter data and sell that. Hell, I can even imagine medical data that has been anonymized being bought and sold (for research purposes)
“…can even imagine…”It’s happening 🙂
Most data that is generated is functional – in other words there are fixed algorithms that produce outputs based on incoming data.This has some profound business and non-business implicationsProprietary algorthms with known inputs and outputs can be reverse engineered.- means they will be.- so ‘super-normal profits’ of any data processing service (in a big market) will tend to zero – early wins are not sustainable unless secured by barriers to entry- services need network effects to protect market share or are ultimately eroded.- to support up sales / cross sales (based on information not data) on top of the aspects of businesses that will tend to zero value, exploits the commoditization effects as accessible market grows (this is a longer term winner)The winners of games are thus people who generate actionable information with actions that support revenue models.So while we produce twice us much data every two days than since the dawn of time… Since we are almost no better informed (we are simply informed about different things and make the same decisions based on different inputs served by different media)We can deduce that we are actually filtering data more effectively.So data becomes ‘information’ simply by virtue of passiing BS filters.Since gaming BS filters can be automated, and costs filtration and adds no value (negative sum game), the best channels to exploit are those that serve the source of the data, as they have disincentive to game the data.This suggests in the very long model advertising models do not belong with information services unless as part of an informed cross sell or upsell strategy(I think) ! 🙂
Data is a prime example of Quantity vs Quality.The exponential growth of the data-pool does not necessarily mean an ever-richer pool of data to mine. Often a relatively tiny but qualified and focused sample-size is infinitely more valuable…
Well said.In my wine project my sample size is small but I can tell you over the last six months from a consumer trend perspective exactly what is being tasted (marketed ) to this segment of the public–spirits/sparkling/countries/varieties/ types/organic/natural whatever.What is more interesting is that by making this information publicly available i’m channeling the trends. ‘Festival’s are organically going to happen because if a leading shop is pouring Jura, I’m seeing shops in other neighborhoods who are fed by the same distributors doing the same on the same day.The distributors will learn to love this project.Very cool. Like a market petri dish!
check out getfoodgenius.com. doing big data with food. you might see parallels to the wine business because of similarity in industries.
.Interestingly enough what you note is the “real” solution to gun regulation in the US.Don’t catalog gun owners, make a list of folks whose behaviors (mental health, criminal records, substance abuse, prescriptions) suggest they should never be allowed to touch a gun.There is a 100% correlation between gun crime and the folks who should be on this list.JLM.
Makes sense. There’s so much data out there but very little truly intelligent correlation – look at the shit the markets throws-up with the ubiquity of algobots etc. Plus, HP/Autonomy, lol. The irony of that one is so funny it’s tragic :-/
under that course of action they will simply categorize anyone who criticizes the government as having a mental disorder. kooks are easy targets here.
Of course they will but being on the list will only get you special treatment.JLM.
So…is “Data” really a revenue model? I think it’s a business model or product. But the revenue model is subscription or pay-for-consumption. Not sure I see the distinction here versus say, mobile phone companies selling minutes.
The distinction is quite thin but ultimately, I see “revenue models” as a list of ways to monetize whatever public facing service you offer. Viewed in that lens, data fits nicely beside subscriptions or licenses.
or both. And disrupting data markets might be a matter of disrupting subscriptions of licenses
I’m glad I read this comment before commenting myself because I had a similar observation. If your product is data then you still have to decide how to generate revenue from that data, right? (subscription, advertising, etc.?) I guess the same could be said for other revenue models like [email protected]:disqus ‘s response below helps and also a comment made earlier in the series that it seemed that revenue model and business model could be used interchangeably.
So I think we’re on the same page.
data is ‘land’, but you have to build something on top of it. that something is the revenue generator.
Can I get some collateralized mortgage debt obligations on that “land”?
they couldn’t give it away
a mobile phone is like a Vegas slot machine, only you can never walk away
Strictly speaking ‘no’ but ‘Data’ or more typically these days ‘Big Data’ is used as a shorthand by many for business models where the IP is either aggregated data or insights derived from datasets.
the dangers are clear, and Fred describes some of them. Many times data businesses become consulting plays, not scalable business models. Still might be a good business-but not an investable one.
the reason for that seems to be largely driven by the following problem: Once I have data, what do I do with it. The consultive play answers that question, the data play just is another toolset
For data businesses to become consulting plays would (IMHO) be an unlikely outcome. Of course you can have a consulting firm helping people implement the analytics platform, training staff in ‘R’ or undertaking business consulting to focus the mind on low hanging fruit, but this isn’t a real data business. Data businesses are product businesses that either package data or insights for the consumption of others or crunch that data to better differentiate their own own virtual offerings.
Wrt ‘big data as a service”http://thinkbiganalytics.com/
Outside of the internet, the grocery business has some interesting parallels. Grocery stores generate immense quantities of data. They don’t share it outside their walls (like to the CPGs) unless there are large $$ exchanged. You can look at companies like dunnhumby that offer their services to sort, sift, and analyze this data. Great company? Not so sure.Catalina, by virtue of their near ubiquitous post-transaction coupon printers, is perhaps the only company that has access to item-level receipts across store chains. But they are likely constrained by contracts about how they can use it. Great company? Could be but seems stuck in its cash-cow biz model.Other companies are doing digital couponing or receipt scanning – getting consumers to give them data. Currently not large enough to be meaningful, but might have a crack at monetizing the grocery purchase data. Great company possibilities? Right now no one has figured out how to scale and/or deal with privacy concerns.And so inefficient grocery pricing/supply/marketing continues until someone figures how to get this data actionable by all partners. Great opportunities, many barriers. Classic case study on data revenue models!
.The loyalty programs — all of which have now devolved to using phone numbers for identity — have tremendous amounts of information and they use it.They now have begun to gravitate those programs to the Internet by forming clubs and using the info to populate their infobase. The web based coupons are better than the cash register receipt coupons.I buy an odd type of very, very thick cut bacon. Hard to get.Until I started getting web based coupons from Randall’s — grocery store where I usually buy it. Same thing with Squirt soda which is virtually impossible to find but now Randall’s always has at least 3 cases of it when I come in.This big data set and actionable info is very, very interesting.JLM.
I’d be really interested in diving deeper into this, reading several case studies of biz’s that have leveraged their data for revenue (but maybe have done so in a way not so obvious to the casual observer).Anyone have any recommendations there? Book or blog post or…?
Grab a copy of Creative Entrepreneurship from http://kbsp.vc/book.html (bonus that one of the essays is from a little known VC in these parts that goes by the name of Fred Wilson)Anyway – one of the initial essays in the book is this one -> oreilly.com/web2/archive/wh… which talks about a lot of great stuff (some related to this)…Oh and the book is *FREE* thanks to @dherman76 and the kbs+ team! 😉
Awesome! Thank you 🙂
Wow! Thanks! A whole new world. 😉
Even better…about 16% of that world currently uses gawk.it to power their blog search ;-D
We can all learn from you, Kevin! Respect. #Marketing.
heh thanks…I’ll feel more accomplished when that ratio is flipped though and it’s 84% that have us installed. 🙂
The problem with most of the folks on the data aggregation side of things is the low quality of data they are selling. We built a tool to help people see the information that bluekai/exelate are selling about them – it’s really surprising to see bad it actually is (http://enliken.com/discover).I think the obvious solution to the quality problem is to involve the source of the data, the individual, by getting people to transact using their data. The trick is going to be figuring out how to incentivize individuals to start transacting (eg – what do companies need to offer you for access to your data?).
i’m very bullish on the human component of data at this time. meaning i think human analysis and re-packaging of data is a growing center of value creation and profit extraction.
Right, because what is data without analysis?
It’s data. ;)As Kid says, the value is in information.
Funny. But interesting to think about how it goes from data to information. I mentioned analysis above but really, I think it is even more than that, it is interpreting the data that makes it meaningful information.
Mind you I was only joking a little. There is actually a formal definition which states (paraphrasing) that information is interpreted data.
http://www.nytimes.com/2013…”I confess I enter this in a skeptical frame of mind, believing that we tend to get carried away in our desire to reduce everything to the quantifiable.”
Zillow, where I used to work, is most certainly a data business. Like Google, they monetize the traffic viewing that data rather than sell it. They could have easily sold that data to businesses (mainly financial institutions) for additional revenue — we got asked all the time while I was there — but it was a matter of focus. I agree with the exec team there, that the opportunity around being the “real estate data” company in the eyes of the consumer is a bigger long term opportunity than a b2b data sales business.
“we got asked all the time while I was there — but it was a matter of focus.”There is a company that provides data to the NY Times and it’s a going joke with me about how literally every week they get mentioned in a real estate story.The company is http://www.millersamuel.com/ and it’s President is always mentioned. “According to Jonathan Miller of Miller Samuel…”I had a conversation recently with a writer at the Times (that used to cover real estate) who is a customer and I asked her why they only use that company in their stories. She said that they don’t have anyone else to get the market data from or they would.Perhaps Zillow is only looking at this from one angle “how much we can charge”.But If they were the source of data for the Times (and you can see who else MS gets their data to) there is no question it would be a benefit to them marketing wise by offering the data for free and getting mentions. It looks like they are providing some info from a quick look at zillow.com but could be much more of a media resource in this area.Note to startups: Journalists are always looking for info, trends, data to write about. They don’t have time and if you can provide something of value to their readers they will write about you and use you as a resource. Then you can use those clippings to legitimize yourself with other media. I’ve done this it’s not that hard to do.
I view data sales and data for the media as two very different things, as does Zillow. They are very much in the data for the media business. I assure you, if the Times is not using Zillow data it’s not due to lack of trying on Z’s part. Data for NYC is a bit harder (I actually spent several months researching the issue in 2006) than it is across most of the country, as NYC is truly its own beast with a number of differences in terms of how real estate data is collected. My gut is the data you are referring to is NYC-specific data — which is Jonathan Miller’s specialty.
Data is relatively easy to collect but selling it is hard work. Carving out a market, making sure you aren’t commoditized, keeping an eye out for new competitors – it isn’t easy!My parent company is doing a fairly hard pivot from selling news to selling data subscriptions – the two overlap though so I doubt it will ever be a complete switch. It is a big challenge to build up a knowledgeable sales force as it takes people who are a bit geeky to get the data and also good at sales – not something you run into every day.For businesses looking at monetizing their data I’d highly recommend a blog that Fred mentioned a few months ago, http://www.asalesguy.com. Keenan is the epitome of the thinking salesperson.
At Dash (www.dash.by) we are building a product to lay the groundwork of an automotive data platform. Given the sheer amount of data exhaust from vehicles and their sensors today, as well as driver/passenger smartphone sensors, social profiles and ambient information to do with car travel (weather, traffic, playlists, local biz), the ‘auto graph’, as we call it, is vastly untapped. In terms of business models, our approach is to build a utility for our users, which is free, and leverage that data for medium-term transactional models (unlock app features or affiliate sell-through to insurers, OEMs, resellers). While the app’s utility (performance monitoring, real time diagnostics, incentivizing safe driving) must be compelling for users to install, the longer term benefit is to use the scaled data to make the roads smarter, safer, cleaner and more affordable. That’s the real ROI for us – using data to improve our lives.
In addition, we’ve just nailed the groundwork for our API, which we hope to expose soon to third party developers.
I can understand mining data to build business and I’m a hybrid who spent many years in sales with Fortune 500 companies, then later I learned to write code so I wear both hats and try to see both sides. Right now the data side is out of whack with value and that has happened over the years and we have way too much value based on selling algorithms. Sometimes it’s like “hey dude lets mine some data, query it and sell some analytics”…people to market this and write reports are plentiful, but does it have real value and we come back to “it’s all about context”.There’s garbage out there with a lot of flawed data with combining credible with non credible and people who to take it to the bank as the gospel. I call it Algo Duping and its running a much and mixing with the good stuff and how in the world do you tell it apart?We didn’t have a bank crisis without some fictional based formulas and the software developed to make it work did we? It’s all about about code an formulas running the world and how much folks will suck up at times, and again not meaning any disgrace here to anyone, but it is what it is. I keep saying we should license and tax the data sellers so we know who they are as billions upon billions are made selling data with banks, companies, social networks, etc. Consumer has no clue on this and the data is used against them. Here’s a link on this topic and read the FICO scam with using credit reporting to score and determine if someone will take their prescriptions? huh you say, yes they are selling that garbage to pharma and insurers and we get scored on this crap. Again I come back to “context” here.Some revenue models keep inequality growing in the US sadly, those with unintended circumstances or you could argue that some are indeed intentional.http://ducknetweb.blogspot….
That last paragraph is exactly what I was looking for.Question for Fred, or anyone interested in answering:Would you consider targeted content production inspired by proprietary data (or even data gathered from social communities & engagement) a good way make a data business better; regardless of the industry?My reasoning is that if original content production has the ability to keep users or consumers engaged and loyal to a brand / company, it may lead to more sales of products / services.An example of this would be Nike+ Technology and Data. What are you thoughts?
Interesting relative to this discussion disqus “recommended content”. On the list attached 2 of the links were pretty good as far as things that I would find interesting and might read.
That’s exactly right. Disqus are using their aggregated data to recommend content and they derive revenue from those recommendations. It’s a ‘data’ business.
This question has probably come up already. Several of these revenue models could have more than one way to monetize. When do you make that decision? I would have thought pre-Web 2.0 that this decision would be part of the business plan but it seems that businesses spring up without a revenue model or even a monetization scheme and seem to be vital businesses– well, except for lack of revenue, if a business can truly be considered vital without this, regardless of how much funding it has received. Is it just that the monetization scheme is not publicized or do these businesses really start without a clue as to how the business will eventually generate revenue — with the belief that if they are providing a valuable product, a plan for monetization is inevitable?So if I have a product idea, is it even worth pursuing if I don’t have some idea up front how it will generate revenue?
There are usually three big (very big) hurdles when it comes to data1) What do you do with it? Your data has to make a case for itself. (or alternatively join into some sort of other groundbreaking business that needs unusual sources of data) If people can’t find a use for your data, they won’t buy it.2) Regulatory. There are certain very valuable types of data, that if they were treated, would be useful if bought and sold. Ex: aggregate medical data and then selling it would make it easier to do research if there was a market for it the same way there is a market for hela cells. We don’t allow (currently) for people to buy and sell your medical data at scale for ethics reasons There are all sorts of data that falls into this category. We’re going to have to make some ethical choices in the near term about these data types, since the usefulness of the data may outweigh the right to privacy. (One of the reasons I care about regulatory is that I am smack in the middle of that issue, for the sake of full disclosure)3) The data body being weird: One of the humongous problems of data (particularly data on the internet) is that it is an uncanny valley of yourself. If you tried to recreate you out of your data in one central database, it would seem like a person similar to you with lots of facts missing out, enough that it would be eery. Using this data becomes suspect because of the differences of behavior vs mind divide. At some point you reach the limits of data, and selling around that limit is difficult.
Really cool to see Yieldbot on the list. As Fred alluded to with Google while we are a data business (right now we crunch 10TB a day and make over 25M real-time data driven decisions a day) what we sell are clicks. We spent the better part of 2 years building technology to make our data proprietary and it was well worth the investment.Couple things I’ll add from our experience:1. The faster you can make your data accessible/actionable the more value it has (in our case that is real-time).2. The more you can bundle/parlay your data with other assets the more valuable it becomes and the more valuable the underlying asset becomes (in our case that asset is media)3. The most valuable use of data is to fuel a marketplace business (in our case keywords/consumer intent)Wall Street is maybe the original template for the above points but the domination of Search advertising, the recent rise of data driven ad exchanges for display and even app businesses like Uber and Hotel Tonight serve as other examples of this valuable “data-based” dynamic.
Data is the raw material. It can be sold in raw form or refined. When refined it represents insights. Insights can be extremely valuable. I agree with Fred that there are 2 primary models:a) simple aggregation – this is a relatively old business. It has low barriers to entry unless augmented by proprietary data. Lexus and Bloomberg spring to mind. b) insights derived from truly massive aggregated datasets is a relatively new business and an exciting one. It has been made possible by: (i) exploding sources of such data, and (ii) tools that enable the analysis of such huge datasets. These tools include everything from enormously powerful distributed processing systems to advances in data science. This model is enormously disruptive because it can introduce the ‘free’ business model into new segments. This intrusion is very hard to resist because the incumbent is hit not only by the fact that the product is free but by the fact that the product is a function of a lot of technology to derive the insights and this may not be quickly replicated.So we know the free product charge for insights model is potentially hugely disruptive. What is much harder to assess is how to build a true business model that can estimate likely costs – which may be dropping fast but are still not negligible for huge datasets – and likely revenues. In fact it can be very difficult to predict the monetizable value of insights unless you are already in a business built on such insights e.g. ad tech. A company that is built on the assumption that a large network of engaged users delivering up data that can be used to derive insights that will be lucrative needs bold investors with a medium/long term view because for many such companies in this new wave it is unlikely to be luminously clear how the insights will be best monetized. With the best will in the world, this will require a leap of faith.
Scoble has interesting things to say about the use of data to create context – he is writing a book about it.And I recommend reading an article and viewing a slideshow here:http://www.iaventures.com/c…
If you are not paying for it – you are the product.In the cases where the data is generated by the users, the real question is who SHOULD own the data?I know TOS of many services claim the rights to the data during the signup which most people do not read. But ethically who should own the data, and what can the services sell. These are tricky issues.
The proprietary date discussion in this post reminded me of your data networks effect lecture at the NYETM at the end of last year:In a data network effects model, every new user makes the network more valuable through the contribution of some sort of data. Every new user makes it harder for all users to get off.
Data network effects are indeed extremely valuable. And if the data is accessed by means of a free product that has formed habits then it is very defensible.
I would love to look a bit into the legalities attached with selling customers’ data. But the bottom line is that data on its own cannot really generate any revenue. It has to be complimented by something else (for instance advertisement, improved service, etc).
The legalities are regime specific. Some territories place more emphasis on privacy than others.
“The Internet is a data generating machine”the internet is a data capturing machine, we are the data generating….er, ‘machine'(s). i hope that’s not how GOOG sees us, but when i read Eric schmidt’s comments on various things, like its source code secrecy, and cyber warfare, and China, i begin to wonder where he’s coming from and how he sees.. our.. future.
He is going to publish a book shortly so you are going to be able to find out. 🙂
There many important semantic differences that many people on this comment thread have touched on that illustrate my POV: The path from “data” to “revenue model” is not as short as this post suggests.Data must first be turned into intelligence, i.e. – Google’s machine learning as discussed by Mr. Wegner. This is where I think proprietary intelligence extraction is more important than proprietary data. If you have proprietary technology or superior talent to glean superior intelligence from public data, the opportunities are just as big. Obviously creating a consumer-facing product that catches lightning in a bottle and creates a ton of useful proprietary data is preferred. But if this data is public, an intelligence-based business model is very valuable. I think as the web evolves and consumers are more comfortable with open data, potential gate-keepers can’t justify locking down their data because it opens up huge business opportunities (transparency/shameless plug: I work for Pongr which uses superior and proprietary computer vision capabilities to extract intelligence from the content of social photos).Then you need to ask, “Who cares?” You need to establish a customer base.Once you’ve decided on a set of potential customers for your intelligence, you need to ask “Is my intelligence and data source the most reliable and efficient methodology for these customers?” For example, if your data is going to tell a large CPG the demographics of their consumer based on social data, CPG’s will respond, “We knew that…. anything else?”If your data is unique, accurate, and statistically significant, can you make it actionable? A dashboard is nice, but unless you can facilitate action or at least consult on the action the customer needs to take, you won’t have a big business. There are huge barriers and major players in the benchmarking business. Google can make their intelligence actionable as part of their product. Albert’s point is that this intelligence extraction capability has many more use cases. Make your intelligence actionable for customers and you can prove that your actions impact a customer’s business. Impact a large group of customers’ businesses in a unique, scalable way with unique intelligence and you have a huge business.
In B2B Saas data can provide an asymmetrical business model. Your product offers a service to some industry that they pay for but because of the data flowing through it in aggregate you can offer value to another payer. POS data for manufacturers or healthcare data for insurers. http://www.practicefusion.com/ offers free EMR because they are planning to do a lot more with the aggregate data—like build an entirely different product for medical decision support. Of course, you have to make sure that your customers are okay with this.Would love to see what Salesforce knows from their data. They could probably predict shifts in entire industries.
My impression is that this business model – so much in the news – has elicited a surprisingly small amount of discussion. This forum is normally so lively and yet this topic just died. Why is that?
This is an interesting article – http://www.ibm.com/smarterp…An MIT Sloan School / IBM study of big data, analytics adoption and value realization in companies.Not least is the importance of presentation (visualization) cultural factors and management processes to realize value irrespective of the potential of the insights. This has major significance for organizations selling into such companies.
Interesting to read this post following the news that twitter is acquiring bluefin. While the later group did build a propriety data collection system, they found it hard to sell probably because the market is still not sure what to do with it. Still, it is a good move by twitter.
A recent exit in the data business … http://www.ft.com/cms/s/0/657ded70...
Hey @fredwilson:disqus, does Bloomberg fall under your first example as a data aggregator? If so, what has made them continue to be successful as more financial data is freely available?
I see them as a network because of the underlying communication system