I’m frequently asked by people about how they can start an “open data business.” Let me first say that I love that the question gets asked. I love that people are interested in Open Data. I love that people want to learn more, they want to play, they want to think of ways of creating a company. These are, in part, signs of how far the open data discussion has come – people see it as a resource that they would like to leverage.
It is, also, the wrong question.
This is not to say there are not businesses that use open data. Indeed, a vast number of companies use open data (anyone company using census data for even a tiny part of their business qualifies). Nor am I denying there aren’t businesses built primarily with open data – the Open Data 500 list demonstrates there are. Plus I get introduced almost daily to businesses that are: both Ajah and OpenCorporates come to mind (I have donated advice, but have no financial connection, with either).
But from a founder (or, I suppose, investor) perspective there are dangers to thinking about “open data” as a unique business space.
The danger is in failing to understand there is virtually nothing that distinguishes an open data business from any other business. Any business needs to solve a real (or sadly, at times imagined) problem, it needs to find clients (e.g. people willing to pay for that solution), and it needs to execute on a number of other things at least competently (HR, marketing, management, cashflow, etc…).
The danger with putting the words “open data” before the word “business” is that it risks making people think Open Data businesses are somehow unique. They are not. If there is a gapping chasm between the question of “what can I do with software” and “how can I create a viable software company” there is an equally large gap between “what can I do with open data” and “how can I create a viable company using open data.” And the questions you need to ask yourself to figure out that latter question (many of which are nicely laid out in this book) are independent of whether it is a software, hardware, crafts or open data business.
Indeed open source software space gives us a nice analogy. I suspect few people decide to create an open source software company – they decide to create a company and the software license is a reflection of their strategic options. I think it is the same with open data. You don’t start a company saying “let’s use open data.” You start a company to solve a problem, of which using or publishing open data may be the only, or the most strategic, way of doing this.
Some readers may be surprised to see me write this. I am, and continue to be an advocate of open data. But open data is not some magic pixie dust that causes normal business logic to disappear. And it is not that I think people are saying that per se, it is just that I want them to understand that the 99% of the problems that needs to be solved in an “open data business” lie in the third word of that string, and that while the first two do confer some unique advantages and disadvantages, these are relatively trivial.
The real opportunity of open data lies not in the way it creates a new unique type of business, but that it offers a new set of cheap building blocks by which to try to solve problems. In other words it increases the diversity and, lowers the cost of, inputs.
Here again the world of software is instructive. The Economist’s recent survey on Tech Start Ups talks of a Cambrian Explosion because of the availability of “Cheap and ubiquitous building blocks for digital products…” many of which are (and many of which are not) open source. The cheap availability of these building blocks is allowing for a range of experimentation that was previously not possible, or at least, prohibitively expensive.
Open data – whether as an input for software products and services, for analysis (journalistic or corporate) or for scientific research – is cheap (in theory free) and increasingly plentiful. It has the possibility of thus being the equivalent of the cheap code that is powering a great deal of experimentation in the world of software. As an open data advocate the possibility of this increased experimentation has me excited.
So if you are thinking about starting an open data business – that is great! I’m excited to hear that and I am keen to help and be supportive. But focus on that third word – business. That’s the one that really matters.
From a data perspective you should be asking yourself – what real tangible pain does doing something with this data set help me solve that was previously only possible with a more expensive input (e.g. proprietary data) or not possible at all. The second is to think of the impact of using open data on your strategy. Where does it leave you more vulnerable (too copy cats or the whims of the data publisher) and where does it leave you stronger (if the data is commoditized then the axis of competition will lie in other parts of the business).
I hope this is a helpful nuance to the issue of open data businesses, and some helpful input for those looking at open data and thinking about to find business opportunities in it.