The pivot you delayed by a year
The signal shows up in spring. The pivot happens the next winter, after months of burn and the best people leaving. Most of the delay is identity, not data.
Most pivots land six to twelve months late. Not from cowardice. From identity. At one company I watched the signal arrive in spring and the decision arrive the next winter, after $4M of burn and three of the best people gone.
The founder there knew in March. Weekly active accounts, the metric he'd built the year around, had flatlined in February and stayed flat through the quarter. He told me over coffee that month that he was seeing the same data the team was seeing. He wanted to run three experiments first. Each would take six to eight weeks.
The experiments failed. He proposed three more. Those failed too. In November he told the board the company needed to pivot. The internal announcement came the following February. The intervening stretch cost the burn, the resignations, and a public story he had to manage for two more press cycles.
The delay wasn't about evidence. By spring the evidence was in. It was about what the strategy had become to him. He'd raised against it, hired against it, told the press against it, told his parents against it. A pivot wasn't a tactical change. It was a public admission that the bet he'd been celebrated for making was over. That admission is the cost, and the brain doesn't pay it early.
The team saw it first, the way teams almost always do. The engineers shipping features know which ones land and which bounce. The CSMs on customer calls know which use cases stick and which are pretend. By the time the founder was staring at the March dashboard, the team had been reading it since January. They were waiting for him to say out loud what they'd been saying to each other for weeks. That gap, between what the team knows and what the founder will admit, is the most expensive part. It teaches the team that the leader is working from different information than they are, or pretending to.
Early pivots are rare because they look bad. The founder who pivots in May, six weeks after the data turns, looks impulsive. The founder who pivots the next February, after eleven months of experiments, looks thorough. The second version makes for calmer board updates and less awkward press. It also burns the extra millions. The optics reward the slow version. The math punishes it. This is the same trap as how you frame a bet that stopped working: the story you tell yourself decides how long you keep paying for it.
There's one honest test, and you run it privately the moment the data first turns. If I were joining this company today and saw this data, would I bet the next year on the current strategy? If the answer is no, the pivot is already needed. The only open question is what it costs to admit it. Founders who can hold the failure without flinching answer it faster, and they keep more of the team and the cash.
Pivot when the data turns. Not when the experiments fail. Not when the board asks. Not when the press cycle is convenient. The data is the data. The story is what it costs you.