Business Intelligence today

“Users constantly complain of information overload. That is too much information that lacks real insight, and not enough time to make sense of it all. More reports faster, isn’t the answer.”

So how does traditional Business Intelligence fare in this age of intelligent business? Almost every major business or organization the world over uses some form of Business Intelligence (BI) to run their business. BI enables us to plan and budget, control costs, figure out how to acquire new customers, and understand how to retain our existing ones; how to comply with regulators and ultimately to report results to shareholders. In short, BI is the eyes and ears checking the performance of our businesses and has grown into a $12bn a year business in its own right.

There is no doubt that BI has helped to publish information, held in static databases, to many middle and senior business managers effectively. It has enabled us to move from consolidating financials monthly, to daily.

We can now view data onscreen in published reports, nicely formatted, in place of the continuous computer paper printouts inches thick delivered on a trolley.

Insight not Information

But simply because you can now automate the distribution of data to users doesn’t mean that in every case you should. Clearly you can generate an ROI that shows how much money you’d save compared with doing it manually, but users constantly complain of information overload.

Distributing reports out to the field doesn’t change people’s behavior.

If the goal is to get them to manage their budgets more effectively, then training might be in order. Just sending a report, where the only indication that they’re heading for an overspend is one number buried on page 26, is not effective at driving change.

The frustration by business users is most often stated as ‘information arrives just too late to be really useful’. While at first blush this appears to be a timing issue, it’s clear that information needs to be acted upon in order to be useful. Business people universally agree that they don’t need more reports. What’s lacking is real insight.

And this insight is needed at the front line in business operations where it can be used to improve performance. There is too much information that lacks real insight and not enough time to make sense of it all.

More reports faster, isn’t the answer. But it’s the answer that today’s BI comes up with most frequently.

The time to insight

Without getting too technical, the real problem we’re talking about is ‘Latency’. Latency is the time taken from something happening or changing to the moment when we can do something about it. It is the root cause of the problem in the architecture used today for BI.

Latency, of course, is not a simple concept, but it is critical to BI. Analyst Richard Hackathorn produced a study of BI latency and concluded that there are three types: Data latency; Analysis latency; and Decision or Action latency.

 

 

Much of the recent work in BI has been to do with reducing the first category: data latency; feeding data faster into the data warehouse. The apocryphal example is the supermarket chain where the data is available for analysis in the data warehouse before the customer has left the car park.

Of course, if the customer happens to be driving away during the analyst’s lunch break, while the data may be updated several times per day, the actual analysis is not going to be real time. In reality, no analyst is going to review the unique shopping habits of Mr. and Mrs. Smith and any impact on the business.

If the analyst were looking at this low level data, he might have noticed is that the Smiths have just defrauded the retailer by returning goods purchased on a cloned credit card, and the packet of cornflakes they bought was the last on the shelf so the shop is now losing sales until someone notices that the shelf is empty. 

So the major problem with current BI is only partly the data latency; just as critical is the manual analysis process, and the time taken for real decisions to be made which can positively affect the business.

A retailer might survive these long delays. But in today’s real time world, many businesses cannot afford to wait and are looking to BI to provide, not only the information, but the insight, the decision and in many cases the automated action.

So how does BI fair? To use an old scorecard favorite: must do better!