Monday 21 April 2014

How the DfE could turbocharge innovation in edtech

I was at the hugely inspiring Emerge Education event last Thursday night. A big theme of the night was how to remove the barriers to innovation in UK edtech. Listening to the diverse pitches, it struck me that most of the products in the first Emerge cohort need some form of data interoperability to make them work. That’s not easy in the UK right now.
Here’s the problem. If you dream up a great edtech idea with the goal of improving student outcomes (like the promising, Emerge-backed Reward System, for example) the chances are you want your product to interact with pupil information. That data is stored in the School Information System (SIS) - eg SIMS, CMIS etc. Right now it is hard to get to that data, for the following reasons:
  1. It's complex and costly to connect to all vendors. If you want to sell to all schools in England, you need to interact with quite a few school information systems. SIMS may have 83% of the market but RM and Advanced Learning also have over 1,000 schools each, and increasingly, recent entrants like Arbor and ScholarPack are gaining ground (my previous blog post covers the market in more detail). So if you want to interact with all of them, you need a way of maintaining multiple interfaces.
  2. It’s complex and costly just to connect to SIMS. OK, you might say, so I’ll just sort out an interface with SIMS, since that covers the majority of schools. Well, Capita will let you do this, but they'll charge you to be a partner. Their APIs also don’t give you access to all data in all ways, and you need Visual C skills to use them.
  3. Vendors have different data models. In an ideal world, all vendors would coalesce around a common data interoperability standard to make it easier for people to interact with different SIS. Something like SIF, for example. The problem is, in the UK we’re a long way away from an ideal world. Opinion is divided on whether the SIF specification adequately covers everyone’s needs, and anyway, adoption is patchy. In the US, initiatives like Ed-Fi offer an open-source data model (a bit like SIF) which states and districts are adopting. That leads vendors to standardise their data models around that framework. However, one of the reasons this approach is gaining traction there is that the American SIS market is much more fractured. In a fractured market, adoption of a common standard benefits everyone. That is not the case when there’s a dominant player, since there’s no incentive for the market leader to accept the trade-offs that come with standardisation. In fact, there's a powerful disincentive, since openness encourages greater competition.
  4. The UK doesn’t yet have its own “Clever. Looking to the US once more, Clever is solving the problem laid out in point 3 above by offering an easy-to-access, best practice (“REST”) API layer with start-up friendly business models. The closest thing to that in the UK is Groupcall Xporter, which lets you access data from multiple MIS. It’s helpful, but my understanding is that it is a scheduled extraction tool, rather than a REST API, which would offer full read/write access to an application. Groupcall also comes at a cost which may prove be prohibitive for startups.
So how do we solve the problem of data interoperability in the UK? Well, one great place to start would be for the government to make data interoperability an eligibility criteria for the next IMLS framework. This framework is the main procurement route by which schools, LAs and academy groups choose their SIS. In other words, it's important, and software vendors care about being on it.

Specifically, the DfE could specify that products need to have free, open source, fully documented,RESTful APIs to be IMLS-compliant. That would make it much easier and cheaper for edtech companies to interact with the SIS, and it would also lead to the rapid emergence of cost-effective, Clever-style products. It would also nudge vendors from a closed-product mentality towards a platform mindset, where others are encouraged to interact with or build on top of the system. (As I’ve said before, I like platforms.)
This may all sound rather niche, but the more I look at the area, the more I think that the DfE should take a stand on data interoperability. It could be the small step that leads to big innovation in edtech.