PM Days 31 to 60: Ship One Artifact That Compounds

Somewhere around day 35 of a new PM job, the advice arrives: find a quick win. Fix a small bug, run a crisp meeting, present a tidy deck. Show momentum.

I think quick wins are mostly stunts. They are small enough to be ignored, cosmetic enough to be resented by the people who were too busy with real problems to do them, and forgotten by the next sprint. What month two should produce is an artifact: a thing that keeps working after you ship it. This is the deep dive on the second month of The PM 30/60/90, and it assumes you spent month one the way the first 30 days post describes: archaeology done, signal map built, opinions held.

The short version

PM days 31 to 60 have one job: ship one compounding artifact and let it earn your seat in decision rooms. Four options: a synthesis of 20 customer calls, an eval set for the next roadmap feature, a live page replacing a manual weekly ritual, or a prototype that ends a circling debate. Choose by team weather, not preference: synthesis for narrative-driven teams, eval set for engineering-led ones, live page for data cultures, prototype for opinion deadlocks. Build it in two to four days on the agent stack from The PM Agent Stack, deliver it with a short memo, then read the reaction data. Artifacts get forwarded, and forwarded artifacts get you invited.

The four options, dissected

The pillar chapter names the four. Here is how to actually choose, with effort estimates.

The synthesis. Twenty customer calls turned into themes, verbatim quotes, and one recommended action per theme. Effort: about two days, most of it human review on top of agent runtime. Pick this when the team is narrative-driven, decides through stories and documents, and the leadership meeting runs on prose. Also pick it when your month-one signal map showed a pile of unprocessed calls, which it almost always does.

The eval set. Take the next feature on the roadmap and write its eval before anyone asks: 30 to 150 labeled examples defining what good output looks like. Effort: two to three days. Pick this when the team is engineering-led and respects rigor over rhetoric. The quiet power move here is that whoever writes the eval defines quality for the feature, and now that is you. The Eval Is the Spec is the method, and the five-row eval template is the fast start.

The live page. One URL with the numbers someone currently assembles by hand every week: adoption, eval scores, cost, open customer themes. Effort: three to four days, most of it fighting data access. Pick this when the team is a data culture, argues from dashboards, and you noticed in month one that a person spends every Thursday afternoon building the same deck. You are not adding a dashboard. You are deleting a ritual. The staff meeting live dashboard shows the pattern.

The prototype. Find the debate that has circled for weeks and end it with two working versions built in an afternoon. Effort: one to two afternoons. Pick this when the team is stuck in opinion deadlock, the same agenda item has appeared in three consecutive meetings, and everyone is arguing about a thing nobody has seen. Prototype in 60 minutes is the proof this is not an exaggeration.

The selection rule: match the culture's native language, not your own strengths. I have watched PMs ship a beautiful dashboard into a narrative culture and get polite silence, because nobody there decides from charts. The artifact has to land in the format the org already uses to decide.

A worked example: the 20-call synthesis

Generic and sanitized, but this is the full pipeline from raw input to delivery.

Day 1, morning: the brief file. Before running anything, write a one-page brief the agent reads with every call: the product context, the three to five bets currently under discussion (you know these from archaeology), and what you want out of the batch. Skip the brief and you get generic synthesis, and generic synthesis is the artifact nobody needed.

Day 1, afternoon: the agent run. Pull the 20 most recent calls that touch your area from the recording library you subscribed to in month one. Run each through the triage pipeline from the 20-minute customer call triage: quote bank, theme rollup, recommended action per call. Batch runtime is under an hour. Then a cross-call pass: which themes recur, at what frequency, with which verbatim quotes as evidence.

Day 2, morning: the human review pass. This is where the artifact becomes load-bearing. Spot-check quotes against transcripts, they must be verbatim. Rename themes in language the team already uses, the agent gets clustering right and naming wrong about a third of the time. Cut anything you cannot defend. A synthesis with one wrong quote is a synthesis nobody trusts.

Day 2, afternoon: the delivery memo. One page on top of the synthesis. Structure: what I did (20 calls, date range, segments), the three themes with one killer quote each, one recommended action per theme, and what I am deliberately not concluding yet. That last section matters more than people expect. It signals you know the difference between 20 calls and certainty.

The delivery itself. Do not book a meeting to present it. Send the memo to your manager and the two or three people whose work it touches, with three sentences: "I synthesized the last 20 customer calls in our area. Three themes recur, the quotes are inside. Happy to walk anyone through it, but it should stand on its own." If someone asks you to present it anyway, the verbatim open: "Nothing in here is my opinion. These are 20 customers in their own words, and I have organized what they said. The recommendations are starting points, not conclusions." That framing is armor. Nobody can fight you on it without fighting the customers.

Reading the reaction data

The artifact is also an instrument. The two weeks after delivery tell you more about the org than month one did.

Who used it. Someone quotes a theme in a planning doc, pastes a quote into a debate, asks for the underlying transcripts. These people decide from evidence. They are your future allies, and often the org's real decision-makers regardless of title.

Who ignored it. Two readings, and you have to figure out which. Either you misread their priorities and the artifact missed their actual problem, or they decide through relationships and instinct, and no document will ever move them. The first is your error to fix. The second is information you will need for every future proposal: those people get pre-wired in person, not convinced by memo.

Who got defensive. Someone owns the official narrative your synthesis quietly contradicted, or owns the report your live page just made redundant. Defensiveness means you exposed a gap, usually by accident. Do not apologize and do not gloat. Book a one-on-one within the week: "I think this overlaps with your work, I would rather build on what you have than around it." Sometimes that sentence turns the most defensive person into the strongest ally. Sometimes it does not, and at least you know.

The mechanism: artifacts get forwarded

Here is why this beats the quick win, mechanically.

A quick win gets mentioned once in standup and dies. An artifact gets forwarded. Someone pastes your synthesis into a strategy doc. The eval set becomes the acceptance bar for the feature. The live page URL ends up in the staff meeting agenda. Each forward carries your name to a room you have never been in, and forwarded work generates a specific question: "who made this?"

That question is the actual quick win. The next planning meeting, the next roadmap review, the next escalation, your name is on the invite, because the room wants the person who makes things like that. You cannot ask for a seat at the table in month two. You can ship something the table needs. The seat follows the artifact, and the same reaction data feeds the coalition logic that The CPO 30/60/90 runs at larger scale.

Failure modes

The artifact nobody asked for and nobody needed. Usually caused by skipping month one. If the synthesis covers calls about a segment the team already understands, or the dashboard tracks metrics nobody decides from, the work was real and the value was zero. The signal map exists to prevent this: build the artifact at the biggest gap between captured and routed, not at the spot that interests you most.

Polishing instead of shipping. The artifact is due in week six or seven, not on the last day of month two. I have watched PMs spend three weeks perfecting a synthesis that was useful at day four, and by the time it shipped the planning cycle it should have informed was over. A compounding artifact at 80% polish beats a perfect one that missed the decision.

Claiming credit too loudly. The artifact already has your name on it. Presenting it in three forums, posting it twice in the big channel, opening with "as I discovered" is how an artifact becomes a stunt retroactively. Ship it quietly and let the forwarding do the talking. The people whose opinion matters notice quiet competence faster than loud competence, and they trust it more.

The stunt dressed as an artifact. A deck is not an artifact, even a good one. A reorg proposal is not an artifact. The test is mechanical: will anyone touch this thing in 60 days without you reminding them? A dashboard people check weekly, yes. An eval set the team runs on every release, yes. A one-time analysis with a memorable chart, no, that is a stunt with production values.

Pick one thing this week

If you are in days 31 to 60: pick your artifact today using the weather rule, and block the two to four build days this week. The deadline is week seven, not week nine.

If your team has a pile of unprocessed calls, and it does: that is your artifact. Start the brief file tonight.

If the artifact already shipped: spend this week on reaction data. Write down who used it, who ignored it, who got defensive, and book the one-on-one with the defensive person. Then read PM Days 61 to 90: Own One Bet and Kill One Commitment(coming Jun 29), because the reaction data is exactly what month three runs on.

Sources: Michael Watkins, The First 90 Days on early wins, where I think the framing needs the artifact correction, Teresa Torres, Continuous Discovery Habits for the synthesis discipline, Marty Cagan, Inspired on prototypes as the cheapest form of truth.

Further reading

• The PM 30/60/90, the full arc
• The Eval Is the Spec, if the eval set is your pick
• Prototype Before You Spec, if the prototype is