I have the opportunity to see many companies doing fascinating things with data at my day job. One truism that needs to be recognized is that generating interesting data is not nearly enough, but rather entrepreneurs need to be laser focused on productizing data. I’ve seen many entrepreneurs miss this point, often times creating an interesting mousetrap for generating, collecting or aggregating data, but stopping there instead of going the distance by productizing the data.
So what does it mean to productize data?
Most generally, there are two types of data-related products:
- Data-driven product. A standalone product that is informed and enhanced by data (ex. Amazon’s recommendation engine)
- Data product. The product is data itself (ex. AP newsfeed)
Let’s look into each of these types of products more closely.
The concept of productizing data is quite intuitive with respect to data-driven products. In these cases a feedback loop is created where some standalone product improves by collecting and effectively utilizing data. Zynga exemplifies the power of data-driven productization by capturing, storing and analyzing almost every interaction users have with their games, and then uses this data to improve game play and optimize monetization. As described in a recent WSJ piece:
To understand why Zynga Inc. is among the tech industry’s hottest companies, consider how it gets people to buy a bunch of things that don’t exist. Last year, Zynga product managers for a videogame called “FishVille” discovered something intriguing while sifting data that Zynga collects when people play its online games. Players bought a translucent anglerfish at six times the rate of other sea creatures, using an imaginary currency people get by playing the game. The “FishVille” managers had artists whip up a set of similar imaginary sea creatures with translucent fins and other distinctive features, says Roger Dickey, a former Zynga general manager who left the San Francisco company recently. This time, they charged real money for the virtual fish, and players snapped them up at $3 to $4 each, says Mr. Dickey.
’We’re an analytics company masquerading as a games company,’ said Ken Rudin, a Zynga vice president in charge of its data-analysis team.
Companies that are able to leverage data to improve products benefit from what we at IA describe as Data Economies of Scale. You can think of Data Economies of Scale as a virtuous spiral in which strong products attract users who generate invaluable data through their usage. Insight gleaned from user generated data is then fed back into product development which helps evolve and improve the product, thus enabling it to attract more users who contribute more data which then feeds into further product improvements, more users and more data, etc. etc. etc.
Data Economies of Scale is an extremely powerful competitive barrier enabling a product to move further and further ahead of new competitive entrants who have yet to achieve enough scale to benefit from a data-driven product feedback loop.
In contrast to data-driven products, it is somewhat less intuitive to understand the concept of productizing data for a product that is essentially, in and of itself, pure data.
Before diving into productization strategies, let’s spend a moment first clarifying the concept of a company whose product is essentially data. Using an example to demonstrate the concept, let’s look at the New York Stock Exchange (NYSE). As the exchange through which stock transactions occur, the NYSE generates an extremely valuable proprietary data asset - market tick data. NYSE then sells this market tick data to banks, hedge funds and media outlets that use it in a variety of productive and profitable ways. Thought of in this light, NYSE is most fundamentally a mousetrap for generating and capturing data and its core salable product is data itself.
Contrary to what you might think at first blush, selling raw data is rarely a ‘product’. The problem with raw data is that it is difficult for end users to consume and it requires users to start from scratch trying to figure out what to do with it. (That said, sometimes users - particularly those who are highly technical and quantitative - want the exploratory flexibility that comes with having the unfettered raw data; but this is the exception, not the rule.)
Instead, data almost always require some sort of ‘packaging’ to become a product. Though I am sure there are more, here are three types of effective packaging for productizing data:
- Structuring data for a particular use cases and ease of consumption, often in the form of an API. A great example of an API product company is The Echo Nest whose entire product suite is a set of highly structured developer APIs that provide powerful access to music data. Another good example of a structured data product is Yipit Data from our portfolio company Yipit. Yipit Data aggregates and structures all relevant daily deal data in an easily understandable and consumable manner specifically tailored for their primary target market of financial analysts.
- Visualizing data in the form of a highly interactive, dynamic and insightful dashboard. Great analytic dashboards allow users to both explore data and be pushed actionable insight. Our portfolio company Next Big Sound does a killer job of taking raw social, event and transactional music data and turning it into easily understandable graphs, charts and visualizations enabling sophisticated exploration of the data and highlighting important actionable insight.
- Creating an experience around data. Sometimes packaging is as straightforward as creating a nice interface and simple interaction tools. For example, Twitter’s product is most fundamentally data (tweets) wrapped in a clean interactive interface that allows users to create, consume and engage with the data. (Interestingly, Twitter initially allowed third parties to create competitive packaging but later strongly discouraged and eventually prevented third parties from creating their own experiences. This strategic move is very much in line with the spirit of the framework I have laid out here – Twitter recognized that a fundamental part of the Twitter ‘product’ is the packaged experience around the data, and as such, they felt it imperative not to outsource this critical element of productization.)
I had a chance to meet with an impressive entrepreneur last week who exemplifies the theme of productizing data. The entrepreneur’s fundamental asset is TV related social engagement data, but his products include 1) multiple well defined structured APIs, 2) an interactive analytics dashboard, and 3) stand-alone applications build on top of the data. It was awesome to see an entrepreneur with such strong product sensibility exemplifying the powerful offerings that emerge when you productize your data.