See why with SeeWhy

SeeWhy is a start-up (founded in 2002) headed by Charles Nicholls (ex Business Objects amongst others), which describes itself as providing operational business intelligence. Now, this is a term that is becoming increasingly popular, so we need to be clear about what it means or, at least, what it means to SeeWhy.

What SeeWhy means by operational BI is the ability to perform analytics against operational data. And by operational data it means transactions, messages and other real-time event-based data. And it also means real-time analytics as opposed to historic queries. For example, you can be monitoring business processes and perform analysis against that data in real-time – this is what I referred to in a previous article as BAMBI – but which only represents a subset of what SeeWhy is providing.

Perhaps the key enabling technology in SeeWhy is the integration layer which, in effect, listens to, say, a message queue, without interrupting the transactions passing along the queue, but capturing relevant (user-defined, which may include threshold-based) data in passing, which can then be analysed using SeeWhy’s in-memory calculation engine.

The point about SeeWhy is that it allows you to compare what is happening today with what happened in the past. For example, you may know that some of your goods are affected by seasonality, but how exactly? With SeeWhy you can compare today’s figure with last month’s, this day last year, other days with the same temperature (if you have that information) and so on. To put this another way, you can see why things are happening.

You can also make comparisons by person, department, stock unit and so on. In other words, it allows you to put today’s data into context, whether that is for seasonality, shrinkage, loyalty card fraud, supply chain monitoring or whatever. Another potential application would be in Internet betting for identifying patterns of betting that might lead one to suspect a coup, or to close arbitrage opportunities.

On an ongoing basis SeeWhy can build up this context (in its Interpretation Store – repository – which is currently SQL Server based but with Oracle and DB2 planned for the future) but on initial implementation you will not have this; so the company provides facilities for loading historic data which enables you to create a baseline against which to compare.

More complex facilities are also provided. For example, there is a data mining capability (though you could plug in a third party product such as SAS Enterprise Miner or SPSS Clementine if you wanted) and you can use this to build a predictive model from which you can extract metrics to be deployed within the calculation engine. Needless to say, there are various graphics and presentation capabilities, as one would expect from a BI product. The whole environment is J2EE-based.

While there are other companies that support complex event processing (notably Progress [with its recently acquired Apama product] and Streambase), as far as I know SeeWhy is the only company to specifically target its product at support for business intelligence as opposed to processing in general. No doubt others will join the party now that SeeWhy has started to validate the market (it has several blue-chip customers) but at present it enjoys a clear market advantage.