AdSlot 1 (Leaderboard)

Decisions – Data day grind

There was a time when the general ledger was king. You coded your ledger to enable groupings and totals that were of interest to your business to be assembled at some ongoing print run. Then you poured yourself a cup of coffee, thumped the printout onto the table and cogitated, flipped pages and cogitated some more. A view of your business and its health and profitability emerged – if you knew what you were doing.

As time went on, the ledger got crusted over with functionally-specific, whizzy new applications. Functions such as sales order processing, distribution and manufacturing grew into monsters in their own right which sat on top of, and fed information down to, a now thoroughly subterranean general ledger.

With so much specific information being captured, printouts became mammoth. This created fertile ground for a new breed of application, business intelligence (BI), now a billion-dollar industry.

But Paul Yarwood, managing director of The GL Company, says it’s time companies got back to the general ledger as their first port of call for business information. “Go back to basics and get everything you need from the general ledger,” he urges.

Yarwood has grown his client base from 21 companies 15 months ago to 84 today. What exactly are these companies buying into? The answer, it seems, is real-time reporting on critical issues, particularly at hectic times such as month-end close.

The problem with current BI, Yarwood argues, is that it functions by extracting information out of large enterprise systems into an OLAP database.

In other words, you are always working once removed from the real-time data. “These are always staged databases that show a view at a particular point in time, but that view is not necessarily in sync with what is happening in the business,” he says. Sometimes it doesn’t matter. Other times, such as month-end close, it matters hugely.

Yarwood claims his Windows-based software can interrogate the general ledger in real-time in environments where the company might be posting, say, three million transactions a month. The ability to handle large volumes is key to his approach.

Business intelligence vendors make a great play out of having web-based BI portals, where users with particular roles can log on and access key information essential to their role in the business. However, Yarwood argues that HTML and web design is inherently unable to deal with reports culled on the fly from huge volumes of real-time transaction throughput.

Instead, you are referenced back to a staged database with no exact knowledge of how far your staged database has diverged.

“The main reason why people don’t trust all the gauges and dials associated with business intelligence systems (colour-coded key performance indicators, for example) is that they don’t know how in sync with the real world these dials are. If something is showing up as red, for example, is it red because you are half-way through your month-end close and there is still a huge volume of data to go through the system? With our system you can see the totals mounting in real-time.”

Yarwood is not saying don’t use BI. He’s saying hang your BI systems off his system, rather than off an OLAP database. So far, The GL Company has focused on PeopleSoft and JD Edwards enterprise systems users and is working on SAP and Oracle products for the near future.

Julian Caine, European CPM marketing manager for business intelligence vendor Cognos, says there are real constraints to reporting direct off a general ledger. “If it were that easy, then Oracle and companies like it would never have gone to the trouble of building staged databases,” he says.

There are a lot of pressures on the general ledger system at peak periods and adding universal reporting on top of those may not make sense. “The key fact with a staged database is how frequently the staged database is updated. Outside of a real-time dealing environment, few people need ‘next second’ updating. A half-hourly or hourly refresh of the staged database is fine for most users,” Caine says. He also believes user organisations like the security of a staged database since it means no one is getting access to the raw data. “Having views of the data assigned to roles via a staged database is just good IT governance,” he says.

Related reading