Guesstimate: A spreadsheet for dealing with uncertainty

Spreadsheets are great for concrete numbers. But what if the numbers are a little … squishy? Then there’s Guesstimate

By Matt Carroll <@MattatMIT>

Data is dirty. Always. I was a database reporter long enough to know the problems that can crop up. Put data in a spreadsheet and those black-and-white numbers look so … accurate. But don’t be blinded. There are problems, even in the best data. For instance, think about this: How reliable was the original data collection? Who created the data? Did they have biases?

Let’s face it, data can be screwed up in a thousand ways. Which means reporters need to have a healthy skepticism about the results they pull from the data.

So it’s nice to discover Guesstimate, a new tool that introduces fudge factors into the numbers, in the form of ranges. It’s a “spreadsheet for things that are not certain,” as described by Ozzie Gooen, its creator. (Learn more here.)

This is how Gooen introduces it:

Unfortunately many important things are not known. I don’t yet know if I will succeed as an entrepreneur, when I will die, exactly how bad sugar is for me. No one really knows what the US GDP will be if Donald Trump gets elected, or if the US can ‘win’ if we step up our fight in Syria. But we can make estimates, and we can use tools to become as accurate as possible.

Guesstimate allows you to use “best” and “worst” estimates, or anything in between.

For instance, let’s say we have a web site and we’re trying to figure out how much time people might spend on our next story. Let’s create a (very) simple sheet, called “Time spent reading my news story.”

We have some figures to play with: An average story on our site gets between 2,000 and 5,000 reads, and average read time is 6–8 minutes.

Plug those numbers into a couple of cells, using Excel-like formulas, then pump out an answer: ~410 hours.

Of course, these numbers are only as good as your data — remember that talk about dirty data? So if your numbers are crap, the final answer will be crap, too, and “guesstimating” won’t help you at all, and might give you a sense of false confidence. But if you have a range of numbers, and the numbers are solid, this might be something you can use.

It also breaks out the means for you, along with histograms of the numbers. Guesstimate is as easy to use as Excel, so if you’re OK with that, then you not have a problem with Guesstimate. There’s also a live chat to ask questions, although I didn’t check that out.

Granted, there’s no heavy lifting with my example— we could have done this alleged problem with a calculator — but the tool opens up interesting possibilities. I’ll be curious to see if it’s adopted by journalists, or for that matter, who ends up using it.

It’s a nice tool. I’ll be keeping an eye on it in the future.

Matt Carroll runs the Future of News initiative at the MIT Media Lab and writes the ‘3 to read’ newsletter, a weekly email about the media industry.

Show your support

Clapping shows how much you appreciated Matt Carroll’s story.