The H1 postmortem: three questions

The half-year review has a way of becoming a ceremony: a deck, a retelling of the number, a slide of learnings nobody acts on. The version worth doing fits on one page and consists of three questions, answered in writing, by the people who own the pipeline. Each question exists to force a specific kind of honesty, and the third one produces the only output that matters, a decision about July.

What did we believe in January that turned out to be wrong

Every plan written in January contained beliefs dressed as facts: the segment that would convert, the persona that would answer, the message that would land, the ramp that would take a month. Six months of dials have ruled on all of them. Name the beliefs that lost. Not to assign blame, the January plan was the best available guess, but because a wrong belief left standing keeps collecting budget. The commonest survivors are a target segment kept on the list out of optimism, the pattern we flagged in your ICP is a hypothesis, and a January value proposition still being said in June even though live conversations rewrote it months ago. This question has a due date, because memory is already editing the story; we made that argument for month one in grading January honestly, and it holds double at the half.

What worked that we ignored

Somewhere in H1, something outperformed and got no follow-up. A vertical that answered at twice the expected rate but stayed a sliver of the list. An objection response one rep improvised that never got written down. A referral source that produced two of your best meetings and zero process. These things hide because nobody planned them, so no line item reports on them, and the review agenda walks straight past. Finding them is mostly archaeology in the notes and the win column: read the accounts that closed fast and ask what quietly helped. The answers here are the cheapest growth available, because they are already proven at small scale inside your own motion. Scaling something that worked costs less than fixing something that failed, and it is the more neglected of the two moves.

What one bet does H2 get

The first two questions produce a list of candidate changes. The third question is a discipline: pick one. A half is only twenty-six weeks, and a bet needs most of that to prove itself through a full cycle of list, message, meetings, and closes. Teams that leave the postmortem with five initiatives run five underpowered experiments and reach December knowing nothing. The one bet should be stated in a sentence with a number and a check-in date, and it should trace directly to an answer from question one or two: double down on the segment that overperformed, or kill the belief that failed and redirect its budget. If the bet needs new capacity to run, the planning for that starts now, not in September, the lead-time argument we made in draft the H2 plan while H1 is alive.

We run this exact exercise with clients at the half: the funnel provides the evidence, the postmortem provides the reading, and the one bet gets written into how the rep's list and message change in July. If your H1 taught you things nobody has written down yet, book a strategy call this week. Bring your January plan, and we will run the three questions against what actually happened.