Microsoft targets mother of all professional people master data with LinkedIn buy

So you saw the announcement yesterday: Microsoft to acquire LinkedIn for over $26bn. Our team were contemplating the meaning of the deal. One thread came out: this is about the data.

This deal is on the heels of several other notable acquisitions as organisations ramp up the serious battle for ownership of the digital economy's fuel: data. October last year saw IBM acquire the digital assets of the Weather Channel (not the TV component). There are other deals like this – another one that is ongoing concerns Beyer and Monsanto; though this is touted as 'all about the seed' there is a hefty data part here too with Monsanto's data business.

The truth is that "analytics" and insight and decision making are all made up of two key parts. Most folks equate decision making with analytics, so the analytic itself, or shall we say the machine learning, is one big part. Another big part is the data that feeds those machine learning engines or algorithms.

And the latest set of engines (think IBM Watson, Microsoft Cortana etc) feed on as much data as you can give them. They need "big data" or even "biggest data" so that they can triangulate on the real hidden insight. Thus the battle is for the algorithm as well as the data.

Microsoft's move is about the data – but more specifically – the master data that sits at the centre of the world's professional network.  So when we say the battle for data is on, in truth all data roads lead to master data. If the core, underlying data objects are not defined well enough, then the engine has to cope with this by doing extra work. And sometimes that work is too much – for humans as well as for machines.

Microsoft is securing a nice chunk of master data – and so other companies may need to licence the data (a revenue stream to Microsoft) or replicate the data (a barrier to entry) or cope with substituted data or bad data (again, costly alternatives).

But acquiring data, and owning access to the best machine learning capability is only half the battle – though it is the battle that is visible and being written about in the press. Decision making in this new digital economy, digital decisioning if you will, needs more than data and machine learning. There are two more elements needed to power the reimaging of digital decisioning. The complete stack looks like this:

1. Data, to fuel:
2. Analytic or machine learning algorithm, in context to:
3. Business process - outcome; delivered in the form of:
4. Application

The business process concept is needed in order to put the decision in context.

In other words, what does the future state outcome and effort to get there look like?  To answer this question you need to understand the business process and outcome.

Lastly, you need to actually deploy the result in some fashion – thus an application of some kind is needed to forge the other three parts. And beyond this, you need "feedback" on the changed outcome in order to determine what to improve in the 4-part model.

Article by Gartner research analyst, Andrew White