The smart way to use BI

Having intelligence is not as important as knowing when to use it, just as having a hoe is not as important as knowing when to plant.
– Chinese proverb

A couple of years back, Fortune magazine reported on a then-new category of software that was helping consumer goods marketers and retailers understand their consumers better.

The software category was none other than business intelligence (BI), and they had mentioned how Ben & Jerry’s (a popular ice-cream maker in the US) had solved a rather knotty problem of customers complaining about the company’s cherry-flavoured ice-cream not having a sufficient quantity of the key ingredient, cherries.

The BI solution revealed that there was nothing wrong with the ice-cream. It was simply a case of the cartons being mislabelled with pictures of cherry yoghurt—a product that happened to contain more cherries, a fact that was quite clear from the picture. The minute Ben & Jerry’s got the labels right, consumer complaints stopped.

There’s a lesson in tales like this. BI can be a powerful tool for marketers in consumer-facing companies. However, if it’s not done right, BI can end up as an oxymoron like MI (Military Intelligence).

Let me clarify that. If you’re an OEM supplying parts to a handful of large manufacturers, you probably won’t have too big a problem figuring out that one of these big guns is unhappy with a particular product of yours. On the other hand, if you are selling soap to millions of Indians, you will need a tool that will analyse the millions upon millions of sales records and tell you where you’re going wrong—or right, for that matter. So who does BI benefit? Banks that lend to individuals (as opposed to specialist banks that lend primarily to corporate clients), telcos, FMCG companies and organised retail.

So BI’s useful, that’s pretty much a given. Then why do some BI deployments leave companies wondering why they ever invested in the ruddy software? The answer is one with many shades.

First off, no BI solution is going to work unless your data is clean. For that you need ETL (Extract Transform Load) software to take all the data lying in your legacy, ERP and other systems. ETL involves scrubbing your data till it is squeaky clean and accumulating it in data marts or warehouses.

This brings us to our second problem. For BI to be useful you need large data stores. How large? US retail chains have data warehouses that can top a hundred terabytes.

Used properly, BI can be a marvellous tool for CIOs to help transform their company’s business by providing vital inputs to business heads. For this, ideally, BI tools should be tied into a company’s extended ERP systems such as CRM and SCM so that the insights dug up using the BI toolkit can be translated into action. This could take the form of a marketing campaign organised using CRM to prevent customer churn at a telco after a BI tool provides a list of users who are likely to switch loyalties. As the proverb at the top of the page says, having information on hand, intelligence at your beck and call, is all very fine. You just have to know when to use it.

Prashant L Rao
Head of Editorial Operations