Data marketplaces, we hardly knew ye

I’m on a panel at the ODI lunchtime lecture this week, where I’m hoping to help answer the question of “what does a good data market look like?“.

As many of you know I was previously the product manager/lead for a data marketplace called Kasabi. That meant that I spent quite a bit of time exploring options for building both free and commercial services around data, business models for data supply, etc. At the time data marketplaces seemed to be “a thing”. See also this piece from 2011. There were suddenly a number of data marketplaces springing up from a variety of organisations.

The idea of data marketplaces, perhaps as an evolution of current data portals is one that seems to be resurfacing. I’ve already written about why I think “data marketplace” isn’t the right framing for encouraging more collaboration around data, particularly in cities.

I’m not going to rehash that here, but, as preparation for Friday, I thought I’d take a look at how the various data marketplaces are fairing. Here’s a quick run down.

If you think I’ve misrepresented anything then leave a comment and I’ll correct the post.

  • Data Market was originally focused on delivering data to businesses, offered sophisticated charting and APIs. Drew largely on national and international statistics. Great platform and a really nice team (disclaimer: have previously done some freelance work with them). They were acquired by Qlik. My understanding is that this rounded out their product offering by having an off-the-shelf platform for visualising on-demand data. This is no longer what I’d consider a marketplace, more a curated set of data feeds.
  • Azure Data Marketplace is still around but seems to be largely offering only Microsoft’s own data and APIs. Seems to be in the middle of a revamp and refocus on cloud apps and more general APIs rather than a marketplace. In its early stages Microsoft explored iterating this into an enterprise data portal as well as deeper integration with some of their products like SQL Server.
  • Kasabi. Shutdown. Sob.
  • BuzzData. Shutdown.
  • FreeBase. Acquired by Google, continued as a free services for a while and shutdown in 2015. The data is now part of wikidata.
  • Infochimps. Originally a data marketplace, the team spent a lot of timing building out a data processing pipeline using Big Data technologies. They were acquired for this technology.
  • Timetric started out as a data platform focusing on statistical and time series data, now seems to have evolved in a slightly different direction.
  • Factual continue to focus on location data. I was always intrigued by their approach which (at least originally) included businesses pooling their data together to create a richer resource, which was then used to drive additional revenue and sales. While there were suggestions they may expand into other sectors, that hasn’t happened.
  • Gnip and Datasift are still around, both still focusing on services and data analysis around social media data

There are others that could be included in the list. There’s one interesting new contender that shares a lot of similarity with some things that we were building in Kasabi, but they’re currently in stealth mode so I won’t share more.

I also don’t include Amazon Public Datasets or Google Public Data as they’re not really marketplaces. They’re collections of large datasets that Amazon or Google are providing as an enabler or encouragement to use some of their cloud services. Difficult to demonstrate big data analysis unless there’s a nice collection of demo datasets.

So, really only the Microsoft offering is still around in its originaly form of a data marketplace and its clear that the emphasis is is shifting elsewhere. The other services that are still around are all focused on a specific vertical or business sector rather than offering a general purpose (“horizontal”) platform for the supply and selling of data.

This matches what we can see elsewhere, there are lots of businesses that have been selling data for some time. While the original emphasis was on the data, the move now is to sell services on top of it. But they’re all focused on a specific sector or vertical. I think cities are neither.


2 thoughts on “Data marketplaces, we hardly knew ye

    1. Its a cloud based data hosting platform, its speciality is RDF/Linked Data but there are plenty of others that focus on other technologies. MLab for example for MongoDB to pull an example out of the air.

      That’s not a data marketplace though. For me the key characteristic of a marketplace is that it provides a platform for people to list their datasets and wrap accessing them in a commercial model. The marketplace may provide hosting or just act as a gateway.

      Under that definition many of those that are still around, e.g. Factual etc are not marketplaces either: they’re a curated set of datasets offered under a commercial model from a single organisation.

      I think there’s a reason for that though. For a paid service, users expect a more consistent experience when using data. Similar APIs, standardised data model, aggregated data from multiple sources. It can be hard for a data marketplace to provide that across a broad range of sectors. So I think this has lead towards specialists in particular types of data and a single provider model rather than a marketplace.

Comments are closed.