Why Information Models Matter

TOGAF is a great framework for Enterprise Architecture. Unlike some, I am a pragmatic fan and am always happy to pick and choose the elements that matter when they matter.

I think however that TOGAF has missed an opportunity. Where the framework is happy to discuss Technical and Data Architecture, they do not focus sufficiently on Information Architecture.

Data Architecture relies too much on an As Is approach. What data do I have, where is it, who owns it and what value does it hold? All very admirable, but slightly misses the point.

You could draw an analogy between Data Architecture and On Premises Applications, Information Architecture and SAAS Applications. The Data Architecture is about what the customer has, while the Information Architecture is more about what the customer should have (reference architecture) and how they can move towards that.
Designing and agreeing an excellent Information Architecture, preferably based on a Reference Architecture, such as the CAUDIT Architecture for Higher Education, is the springboard for a number of key activities, as can be illustrated below.


Using a Reference Architecture massively reduces the time to impact, keeps the stakeholders engaged and gives confidence to decision-makers (also countering the naysayers who think this is in the ‘too hard’ bracket).
Coordinating the activities above is also much easier with an agreed Information Architecture and dependencies can be managed effectively.

So why aren't more organisations doing this?

I'll give the same answer as a myth-buster will always give. There is little money to be made by big software vendors or systems integrators, as this is ‘making the best of what you have’ and requiers in-house knowledge, more than out-of-house consulting skills (unless you can find people who understand your sector). with this approach, they are less likely to be able to sell you a Master Data Management, Data Governance or SOA system quickly, let alone a standalone application, although these are still important and must be integrated into the Information model.

Working at University College London and before that at Imperial College has given me a glimpse of the possible.
At Imperial, we designed a model which could be used to simplify point to point interfaces, starting with the key entities that the College business understands.

IC Concepts

At UCL, we went one step further and have built an Operational Data Store, initially for the HR Transformation, but ultimately for the wider UCL applications domain. This Store holds a single, trusted record of a person today and can be read easily by other applications, facilitating key information exchange and federated reporting. This is an example of what the Information Model can lead to, so this is not an intellectual exercise.

Cost savings from reduced interface costs, improved data governance, better decisions through federation of information all point to the value of this understated model.