ODCamp 7: The ROI of Open Data

How do we show the value of opening up data? That was the question asked by Dan Winchester, who runs a company called Get the Data, in the second round of sessions at Open Data Camp 7.

“I am a data publisher and also a data consumer,” he explained. “I create data sets and put them out in the world and people consume that data. But I have very little information about what is created. Is there value: economic, or social from it? That makes it hard for me to know whether I am putting resources into the right place.

“While, as a data consumer, I am seeing benefits from using data, but do the people who are publishing the data aware of those. Am I doing as much as I could do to help them make the case for allocating resources to the data I am benefiting from?”

There were, he said, lots of ways to address these questions. There might be a register of people producing and using data sets, that could be used to quantify their value – “a Google Analytics for data usage.” The way this would work would be that if Dan used, say, a piece of Ordnance Survey data, he would tag that, which would enable the OS to see that the information had been used and what use had been made of it. After which, an economic value could be attached.

“It struck me that the open data community does this kind of thing all the time,” he added. “We create data and standards and technology, so we should be well placed to do this. The question is whether there is an appetite to do this.”

OCD7 ROI

Yes, but…

Campers thought there would be an appetite for doing this kind of thing. In fact, one suggested that some data set publishers already do it. However, someone else felt that handing over information in return for open data should be optional. Otherwise, the community would be back where it was in the early days of open data, when organisations asked users to register to obtain information: which can act as a barrier.

“One of the issues with this kind of sign-up process is that, now, the people who want this data will not be computer or tech savvy,” another camper agreed. “They want data, but they want it immediately, and if there is a sign-up process in the way, they won’t continue with that.”

The reasons for the publisher asking for this information should also be clear, another participant suggested. And the process of collecting and analysing it would have to be easy, because otherwise publishers might not have the time and resources to do anything with it.

Still, publishers in the room argued that it is useful to have some information to make the case for investing in or maintaining an open data set to politicians or other stakeholders or funders. “Another thing that would be very interesting is to see what data sets are never used. A bit like when Battersea Dogs Home tweets about a dog that hasn’t found a home, to see if it can find one,” said another participant, flipping around the same point.

Returning to the issue of friction, another camper suggested that ‘super users’ of data might put some “skin in the game”, by not only registering for a data set or providing some information about what they were using it for, but working with the publisher on maintaining something critical to their site or business: a quid pro quo that would benefit both sides.

After all, another consumer pointed out: “As a power user, I really want to be able to come back and say: here are the problems with this data, and here’s how to put that right. I don’t want to have to maintain my own set of corrections.”

If there was a conclusion, it was that there is an appetite for making sure that the publishers of open data can see what value is being generated by it, and for the consumers of open data to help them. But camp felt that any mechanism for doing that would need to be as friction-free and two-way as possible. And nobody stepped forward to create a register or other mechanism; although some campers suggested they should continue the conversation.

A Google Analytics for Open Data?

Asked to explain more about his own ’Google Analytics for open data idea, Dan said: “My idea is that you have a website or app and you can measure the number of visitors or users. You add a little pixel to the information, that reports back to the register, and let’s the publisher of the data see what is being done, and perhaps gives them more ideas. Because what you are really interested in is how many end-users you are touching in a hopefully beneficial way.

“There is a bit of a model already, in that Data.Gov used to let you showcase your app if you used their data. And people did do that because they got a link back from Data.Gov which was useful for SEO (search engine optimisation). And the police did a similar thing. So it might not be that everybody would do it, but enough people would do it to get some useful information. Even if the information wasn’t comprehensive, even if you got just 20% of what was happening, that might be very useful.”

Anybody interested in this idea, or the wider subject of how to get consumers to help publishers to show the value of open data can email Dan: dan@gethedata.com

[Session Notes]