CPO Days 61 to 90: The Kill List, the Bets, and the Readout

By day 60 of a CPO's first quarter you have banked credibility: the audits were real, the instruments are live, the visible decision landed. Days 61 to 90 are where you spend it. And here is where most new CPOs reach for additions. A new initiative, a new team, a new strategy narrative. Additions feel like leadership. They photograph well.

The higher-return move is subtraction. Every product org carries weight that survives on inertia, your audits told you exactly where it is, and killing two of those things publicly buys more durable authority than launching five new ones. Everyone in the building knows which call is harder.

The short version

CPO days 61 to 90, the closing month of The CPO 30/60/90, has three deliverables. The kill list, built by running the inertia test over everything the audits flagged: does this change a decision? Two public kills, each with a written kill memo. And two or three first bets, reversibility-weighted and evidence-linked, with an explicit anti-bet paragraph. All of it lands in the day-90 readout, an SCQA memo in the Barbara Minto tradition, written to be read rather than presented, delivered first to the exec team and then in a tightened version to the board. The instruments that make every claim in the readout checkable were built in days 31 to 60.

Building the kill list: the inertia test

Pull everything the audits flagged: the initiative with no owner since the sponsor left, the weekly ceremony nobody can explain, the feature with four users and a maintenance rota, the report that gets generated and never opened. For each item, one question.

Does this change a decision?

Not "does someone like it." Not "did it matter once." If the item vanished tomorrow and no decision anywhere in the company would be made differently, it survives on inertia. Onto the list it goes, with two numbers attached: what it costs (people, compute, attention, calendar) and what killing it frees up. The cost-per-outcome instrument from month two gives you the compute number for free, which is one of the quiet reasons you built it.

The list will be longer than you can act on. Good. Rank it by cost-freed divided by political-price, kill the top two by day 90, and put the rest in the readout as dated commitments. A kill list that lives only in your head is a wish. A kill list in the readout is a plan the org can hold you to. The Anti-Backlog is the standing version of this practice once the first 90 days are over.

The kill memo: anatomy

A quiet kill teaches the org nothing and invites the thing to un-die. Every kill gets a memo, published where the whole product org can read it. Five paragraphs, one page, this structure:

1. What we are stopping. Named plainly. No euphemisms like "sunsetting our investment posture." You are killing it. Say so.
2. What the evidence was. Two or three sentences citing the actual numbers: usage, cost, the eval trend, the audit finding. This paragraph is why you waited until you had instruments.
3. What it cost us. The engineer-months, the compute, the calendar hours. People defend things until they see the price tag.
4. Where the freed capacity goes. Name the destination. A kill without a destination reads as cutting. A kill with one reads as investing.
5. What would reverse this. The falsifiable condition under which you would admit the call was wrong. This paragraph does the most work, because it converts the kill from a decree into a testable decision, and it tells every team how decisions now work here.

I have watched orgs where the first well-written kill memo changed the culture faster than any all-hands. The craft of executing the kill itself, the customer migration, the comms sequencing, is in The Deprecation Playbook, and a sanitized end-to-end run is in the killed-feature playbook.

Sizing the first bets

Two or three bets. Not five. Five bets is a roadmap wearing a strategy costume, and it tells the org you could not choose.

Each bet has to clear two bars. Evidence-linked: anyone reading the readout can trace the bet back to a specific audit finding or a resolved contradiction from the listening ledger. "We are betting on workflow X because the money audit showed it is our only workflow with improving cost per outcome and the eval trend is positive" is a bet. "We believe AI will transform our category" is a horoscope.

Reversibility-weighted: for each bet, write down whether you can walk it back, at what cost, and by when. Reversible bets can run on medium confidence, and you should let them. Irreversible bets, the pricing-model change, the platform migration, the public deprecation with contractual teeth, need the highest evidence bar you have, because being wrong on those compounds. I wrote the longer argument in the cost of being wrong. A reasonable portfolio at day 90 is two reversible bets and at most one irreversible one, and if you have zero irreversible bets, that is fine. You have only been there 90 days.

The anti-bet paragraph. Directly after the bets, one paragraph listing what you are explicitly not doing and why. "We are not entering segment Y this year, despite field pressure, because the money audit shows we cannot serve it above water at current cost per outcome." This paragraph is the cheapest alignment tool in the building. Every future hallway debate about segment Y now ends with someone pointing at it. Write it with the same care as the bets, because it will be quoted more often.

The day-90 readout, section by section

One memo, SCQA-shaped, following Barbara Minto's pyramid: the structure forces the argument to hold without a presenter in the room.

Situation. What the audits found. Where the product actually is: the cost picture, the eval picture, the decision-audit picture, stated as findings with sources, not impressions. This section earns the right to everything after it. Keep it cold. If a finding is uncomfortable, that is the finding doing its job.

Complication. The two or three constraints that actually bind. Not ten. The discipline of cutting to two or three is the analysis. A constraint is something like "our largest workflow loses money on every run and the loss grows with usage," not "the market is competitive."

Question. One sentence. Given the situation and the constraints, what should this product org do in the next year? Writing the question explicitly looks pedantic and is not. It is the contract for the answer, and boards notice when it is missing.

Answer. The bets, the anti-bets, the kill list with dates, and the scoreboard you want to be judged on, which you pre-negotiated with the CEO at day 50 so this section contains zero surprises for the one reader who matters most.

Write it as a memo, not a deck. A memo cannot hide a weak argument behind a strong slide. Six pages maximum. If you need the visual version for the board, build it from the memo afterward, never the reverse. The skeleton is in the strategy memo template, and the board adaptation is covered in Investor and Board Narrative.

Running the read-through

The exec session. Send the memo 48 hours ahead with one line: "Please read before the session, we will not be summarizing it." Open the session with 25 minutes of silent reading anyway, because half of them will not have read it, and the silent read resets everyone to the same page. Then discussion, steered to the two or three places you expect resistance, which your coalition map already predicted. Do not defend every sentence. Concede the small points visibly; it buys you the big ones.

The board version. Shorter, colder, and reordered: scoreboard first, because that is the contract the board cares about, then bets and kills, then the audit findings as an appendix. Pre-wire it with the one or two board members your CEO flags, in 30-minute calls, before the meeting. A board should never encounter your kill list live for the first time. Surprise in a boardroom converts allies into auditors.

Failure modes

The readout as victory lap. A memo that says everything is on track tells the board you found nothing, which they will not believe, or you are hiding something, which they will. The complication section should sting a little. Naming the hard constraint before someone else does is where day-90 credibility comes from.

Betting on what you knew before you arrived. If your day-90 bets match your day-zero priors exactly, the 90 days changed nothing and the audits were theater. Check your priors file. At least one bet should be something you did not believe on day one, and you should say so in the memo. It is the single most credibility-building sentence available to you.

The kill that silently un-dies. Three months after the memo, the killed initiative reappears wearing a new name, usually attached to a sympathetic exec. This is why the kill memo includes the reversal condition: if the condition has not been met, the resurrection has no legitimate path and you can point at the paragraph instead of refighting the war. Put a calendar check on every kill at 90 days post-memo. Zombies move at predictable speed.

Pick one thing this week

Run the inertia test on your own calendar before you run it on the org. Every recurring meeting you own: does this change a decision? Kill one, with a two-line memo to the attendees explaining why. It is the smallest possible rehearsal of the month-three muscle, and it costs nothing.

The full arc starts with the listening ledger and runs through days 31 to 60. The readout skeleton, the kill memo template, and the bet-sizing worksheet are in the CPO First-90 Kit.

Sources: Barbara Minto, The Pyramid Principle for SCQA and memo structure, Michael Watkins, The First 90 Days for the transition arc this series rebuilds, Marty Cagan / SVPG on product strategy as choice.

Further reading

• The CPO 30/60/90
• The Deprecation Playbook
• The Anti-Backlog