On 14 March I wrote, in public and with a date attached, that the mid-market workflow category would consolidate from eleven credible vendors to five within two quarters. It went the other way: the count is fourteen today and two of them raised at higher valuations than the leaders did in 2025.

The forecast, unedited "Consolidation is a function of distribution, and none of these companies has any. Expect two acquisitions and three shutdowns before September."

I am keeping the sentence exactly as written, because a prediction you can quietly revise is not a prediction.

The 41 days that broke it Between March and April the three largest players in the category shipped free tiers within 41 days of each other. That single move inverted the assumption underneath my forecast: distribution stopped being the scarce thing, and price became the weapon. Cheap distribution expanded the category rather than consolidating it.

I did not see it because I was watching balance sheets instead of product pages.

What I will change next time First, I will write the falsifying condition next to the forecast — the specific, observable event that would prove me wrong, not a feeling. Mine would have been "a leader ships a free tier," and I would have caught it in week two.

Second, I will grade on time, publicly, even when the answer is embarrassing. The deadline is what makes the record worth anything.

The forecast was wrong. The logging was right, and that is the only reason I can tell you why.