When you organize an event, sometimes a local TV news van shows up. And people ask, “are we going to be on the news?” The answer is almost certainly yes. The station only has so many vans. The decision about whether or not the event is newsworthy was made when they decided to send the van.

That’s what’s bogus about “data-driven” organizations. They’re still making decisions based on their values and on vibes, but the decision-making step is moved, to the step where they decide what data to collect to justify the decision.

Here’s a good recent example. When you cancel Microsoft 365 Business Basic (which is, as far as I can tell, one of the confusing names for what’s basically “MS-Office on the web”) you get a survey about why you decided to cancel.

Here are the options:

From the Microsoft 365 Business Basic cancellation survey

Notice anything? There’s an option for “My company needs an AI solution” but no option for “This service has too much AI”. So back at slop HQ in Redmond, Washington, they’re going to be making “data-driven” decisions based on how many customers left because they wanted more “AI” but they get no data pointing the other way.

ICYMI, a toot from Charlie Stross explaining why he canceled “Microsoft Office 365” (which is possibly the same thing as the above? It’s a platinum sandwich situation with MS-Office and Office-adjacent products and services over there now.)

CoPilot in Office would open me up to accusations of breach of contract—my book contracts warrant that they’re all my own work: CoPilot brings that into question.

This could be a whole niche for an office suite. A lot of people are using office software to make some kind of files (documents or presentations or whatever) that they have already signed a contract about, promising that the files will be their own original work. And a lot of them know Microsoft Office already and would prefer not to have to learn LibreOffice, the maintained successor of OpenOffice. Is there a market opportunity for “Microsoft Office: Authentic Human Content Edition”? Maybe, but the way that the “data-driven” process works, Microsoft will never know.

For what it’s worth, the “data-driven” situation is a big part of why corporate internal prediction markets never took off. Prediction markets have a bad habit of contradicting things that executives just know, on vibes. More: How we get to the end of prediction market winter