Real Time Reporting

Earlier this week, I represented our global BI group in a review of a whitepaper on enterprise portals being written by another division in our company. The paper covered real-time reporting of operational performance, which of course is very similar to some of our BI work. My review was needed that to ensure that the ideas presented in the paper and the terms used did not conflict with our views of BI.

One of the points raised in the draft whitepaper was about the lag between business-affecting events, their detection by ‘real-time’ reporting processes and the length of time that a business has to respond to them. In general terms, the number of options that a company has to deal with an event diminishes as the time from the event increases. Nothing really new in this, we all know that that companies that ‘keep their eye on the ball’ probably do better than others that put decisions off until there is only one (or even, no) course of action left. But it does make a clear statement that by making the gap between event and reporting as small as possible the more options are open to respond.

The concept of lag between an event and its reporting is interesting in BI terms. Not too long ago warehouses were looking at data from the previous month, later we had a latency of around a week, many DWs now work on the previous day figures. Advances in technology make it possible to add data into a data warehouse ‘as it happens’. But it is debatable whether real-time feeds are always appropriate for a business; Retail BI probably does not need real-time reporting (we have one hospitality customer that moves staff between outlets based on real time reports though), but for utilities or finance users real time BI could be a great benefit.