bug in an Excel spreadsheet influenced the outcome of an economic study. Flawed numbers can lead to bad decisions and/or take the credibility out of your presentation. “If the number on this page is wrong, I might as well have to check the numbers on every page....”

I have built my fair share of models (mostly discounted cash flow valuation models of big companies) and a small bug can make a few billion dollar difference in valuation.

My main strategy against bugs:
  • Simple formulas: plus, minus, times, divided by. 
  • If possible only 2 numbers per formula, intermediate results appear as another line
  • Round up to a unit that leaves you with one digit behind the dot (millions, thousands, etc.)
  • Everything points down, a result always depends on values higher up in the spreadsheet, never the other way around
  • And finally: adhere to the rule that if it looks wrong, it is probably wrong. Averages should be within the minimum and the maximum value, gross margins are usually around 50%, etc. etc.