So you just found an error in one of your data models.
Let’s say the sum-function did not sum the whole range – only the top 55 rows, the last 6 were unaccounted for.
“Congratulations, I found an error” is probably not what you think. But you should.
Your thoughts are more like “What is this?”, “Since when is this error in this model?” and most important “What kind of data did I show at the board-meeting last month?”
Let me be very clear here: errors in your data-model are very common. More common than you might think. We are all humans.
Popular stats say “over 80% of all spreadsheets have errors” – but this is other people…not you, right?
Don’t be frustrated when you find an error, rather be elated because you know, your data model has now one error less. And finding one error often leads to investigating whether there are more errors.
The cost of finding an error is small. Your frustration, your time and your embarrassment when you need to admit it to others.
The cost of not(!) finding that error could be very substantial.
You might win a tender based on a wrong calculation and the actual cost of delivering the job is $300K higher than your false estimation. Ouch!
You might invest in a marketing campaign based on wrong interpretation of your sales data. So instead of backing your best product, you are promoting a dud. Money down the drain!
You might present a healthy surplus to the board only to come back with the real figures a month later – showing the actual situation with way higher costs. That misleading information could cost you your job! (pun intended)
And arguing that nearly all spreadsheets are faulty, to get you out of jail free, does not work.
Here’s what you actually need to do to minimise the risk in your data model:
- Testing, Testing, Testing
- Get another pair of eyes to see and test your model
- Explain and show your model in detail to another person
- Get your model audited
- Engage a modelling expert
Maybe something to consider for 366 more relaxed days in 2020.