Last Updated: June 2024

Why We Do This

Making our spreadsheets tidy in a consistent way serves three core purposes:

• It makes it easier to spot when we’ve made a mistake
• It makes them easier to understand for other people who are reviewing your work - including yourself, many weeks or months down the line when you might have forgotten why you originally did something
• It means that work looks professional, and if we had to do, we could share externally

Just as it is hard to write a long document without typos on the first try, it is next to impossible to create a complicated spreadsheet without any errors.  Many studies  have consistently found that almost all spreadsheets contain mistakes.

Unlike a typo, a mistake in a spreadsheet can have a significant impact on the final estimate, changing a key number by an order of magnitude.

For our core models that are repeated multiple times (eg Google, Uber) we are working towards having a version of the model both in code and spreadsheet form. Creating a model in code rather than a spreadsheet is no guarantee that you have not made a mistake - but it is quite hard to make the same mistake in both coding and spreadsheet form.  If the numbers in both match, we can be fairly confident that there are no simple errors.

For pure spreadsheets, we should follow our internal style guide to try and reduce the prevalence of errors. While there is no silver bullet, these practices can help reduce the prevalence of the worst of errors. Your spreadsheet is not ready to go to a client - or your line manager - until it is fully tidied. This will take time to do properly, and you should allow for it. Often coming up with the original estimate is only half the work.

The Types of Errors We’re Trying to Avoid

Among the most common type of errors are:

• Broken or mistaken cell references. (By far the most common.)