Spreadsheets, spreadsheets everywhere

Everybody loves spreadsheets. But pretty much everybody acknowledges that spreadsheets are not the right tool for managing critical aspects of a business.

Earlier this week I had the pleasure of hosting a seminar with the ‘father of data warehousing’ Bill Inmon. He presented his ideas about the architecture of the next generation data warehouse. To our surprise, Bill spent a good amount of time discussing the dangers of spreadsheets. That’s surprising as hardly anybody suspected this topic to appear in that type of presentation. But Bill felt that it was relevant to highlight the dangers. That’s yet another sign that the proliferation of spreadsheets is indeed a huge problem.

At the DW 2.0 workshop

Bill told the story of the CFO of a large oil company. This CFO is apparently very frustrated about the use of Excel within his organisation. He sarcastically states that ‘spreadsheets are the best decision-making tools’: If you want to make a certain decision, you can use a spreadsheet to create and provide the supporting information for that decision. Spreadsheets are ideal to not only manipulate data but to also make serious mistakes with the data acquisition and preparation. And of course, somebody else will be able to produce yet another spreadsheet that provides information that speaks against this decision. Big discussions and data comparison exercises are the result. Trust in data goes down the drain.

True: Spreadsheets are risky. It starts with the data load (where did the data come from, how was it acquired, when was it loaded). And then there is the omnipresent danger of breaking links and formulas. To add insult to injury, spreadsheets are also fragmented – not part of a common data repository.

In summary, spreadsheets are so flexible and fragmented that we can use them to influence our information to a degree that is dangerous. Controlling data integrity is virtually impossible. Bill Inmon clearly stressed the dangers associated with liberal spreadsheet usage. And we should listen to those warnings. The answer: a solid data warehouse along with the proper reporting tools?