Chris Kanarcus pointed me to an embarrassing story about how an error (cockup?) in a spreadsheet has led to Tibco investors losing $100 million when and if the planned acquisition by private equity
bandits company Vista Equity Partners goes ahead. According to regulatory filings:
"In its Sept. 27, 2014, presentation to the Board of Directors, Goldman Sachs, using the spreadsheet, used a share count that overstated [Tibco's] fully diluted common stock by the number of shares of restricted stock...”
This is being pitched as existing shareholders getting a 'slightly lower' payout to the tune of 61 cents a share but come on people - $100 million? That's a tasty number in anyone's book. It's like me saying to Bill Gates: "Heh Bill, how about a check for 0.1% of your wealth?" That would still add up to $55 million but who gives a damn? He sure wouldn't miss it. Would he?
More seriously, I've been saying it forever - like since at least 1997 - but it's worth repeating:
Spreadsheets are general purpose tools that can do many powerful things but they are a programming environment and should be treated as such. That means testing and documenting according to good programming standards.
It is hardly surprising to learn that although beloved of many a CFO, spreadsheets are treated with an astonishing degree of abuse. There are conferences on spreadsheet management, Professor Ray Panko has been documenting the incidences of spreadsheet error for many years. Now and again I hear of particular cases, often involving huge sums of money. Ones that immediately spring to mind include:
Spreadsheet hell, an inconvenient truth
- Heinekin overstating interest payments by more than $1.5 million per month in one financial reporting statement. Cause? Undocumented calculations.
- Mortgage book overpayment on handover: $200 million. Cause? Logical error in calculating years outstanding.
- PG&E reject NAPP bid but it included a single cell error that miscalculated demand generation requirements. Note - PG&E is no stranger to spreadsheet error. There are basic issues like failure to apply correct addition formula in documents going back to 2008.
- Those with a taste for the morbid will enjoy this collection of clangers that include Kodak, Fannie Mae, Nevada City and many more.
Back to TIBCO. Reuters puts it well:
As in the spreadsheets of widely cited economist Thomas Piketty, however, even minor infelicities undermine credibility. With Goldman in line for a fee of nearly $50 million, Tibco surely expected exemplary attention to detail. Then again, shareholders might have thought Chief Executive Vivek Ranadivé and finance lieutenants would have the enterprise’s dollar value to hand in real time, and to have noticed when Goldman’s number was off in the fairly obvious second significant figure.
Tibco touts the ability of its products “to capture the right information at the right time.” The lesson of Goldman’s snafu for companies and financial advisers alike is to double-check, in spreadsheets and any other software, that’s what is really happening.
Maybe they should have used Tibco's Spotfire?
- Many vendors have gone the path of least resistance, assuming that you can't wean finance types off the ubiquitous spreadsheet. I call that lazy. Customers deserve better. Ring fencing the spreadsheet is a half way house solution.
- The incidence of publicly reported failure remains low but studies consistently warn that upwards of 90% of spreadsheets likely contain material errors. The economic impact, while hidden, must be horrendous. Why do we put up with this kind of thing?
- Dedicated analytical and reporting tools are far more accessible today then ever before. There really is no excuse for not considering modern solutions where the algorithms and calculation engines have been built with rigor in mind.
- Almost every case study I have seen on the acquisition and use of analytical tools over close to 20 years cites the need to move off spreadsheets. Why do finance people cling onto them?