The settled bet

Spring 2020. Ecommerce volumes hit levels the pre-COVID trendline assigned to a future decade. Lutke's public framing at the time: 2030 pulled forward to 2020. Nobody, anywhere, had data on whether the shift was permanent or elastic, because the event producing it had no precedent. Shopify's merchants were drowning in demand; the infrastructure question wouldn't wait for the answer.

Treat the surge as temporary and let service degrade while the fog cleared. Or hire and build as if the new level was the new baseline, and be wrong in public if it wasn't.

Bet permanent: aggressive hiring, logistics network buildout, capacity everywhere. And, the part that makes this file canonical: he stated the bet in writing, as a falsifiable claim, while placing it. Shopify was betting that the share of dollars flowing through ecommerce would permanently leap ahead by five or ten years. A specific, checkable claim, with his name on it, made at the moment of maximum uncertainty.

The July 26, 2022 memo, "Changes to Shopify's Team," published on Shopify's own site, in Lutke's own name. The structure, in order:

The prior, restated: the bet as originally written, five or ten years pulled forward, permanently.

The evidence: it's now clear the bet didn't pay off. The 2022 numbers had reverted toward the pre-COVID trendline, and he showed the chart.

The ownership: *"Ultimately, placing this bet was my call to make and I got this wrong."* First person. No pandemic to blame, no macro, no team.

The action, in the same motion: 1,000 people, 10% of the company. In May 2023, another 20% alongside the sale of the logistics arm to Flexport.

This is the Call Memo and the tripwire, demonstrated at maximum scale. State your bet as a falsifiable claim while making it, in writing. When the evidence lands against you, publish the update in the first person and act in one motion. The written claim is what made the clean settlement possible: an unwritten bet is infinitely renegotiable by the person who made it, and the 2022 version of an unwritten 2020 bet gets argued with for another year while the payroll burns.

Contrast is instructive: the record's failure cases (Katzenberg at Quibi, Johnson at JCPenney) share one structure: the bet was never stated falsifiably, so no evidence could ever officially defeat it. Lutke's memo is the opposite machine. The claim had a shape, the shape met the data, and the data won in public.

That's what this system's memo field EVIDENCE and the tripwire discipline are for. You will not out-predict the world. You can out-settle almost everyone.