Your PMM Shipped Four Launches in Q3 and Sales Mentions Two on Calls

It is Friday, 3:47 PM. Your Head of PMM opens the launch tracker with the last coffee. Four features shipped in Q3 against the roadmap commit. Two of the four appear on the current sales talk track PDF. One of the four made it into the demo script the AEs read on discovery calls. Zero made it into the pricing page. The Q3 launch budget reads $184K across creative, video, and paid amplification.

She opens Gong and searches the last 40 AE calls for the feature name. Two mentions across 40 calls, both from the same rep who joined the beta. The CRO opens the QBR Monday with a slide titled "Q3 velocity" and the four features do not appear in the pipeline attribution column.

The founders in this seat keep hiring another PMM body and wondering why launches do not convert. Launches are not a body. Launches are a 21-day queue that starts with a PM's Notion doc and ends with an AE citing the feature by name on a discovery call. Five teams touch the queue and none of them own it end to end.

The 21-day path from launch decision to first AE mention

Sort the last four launches by feature-flag-on date. The median lag from launch date to first cited mention on an AE call runs 47 days. The healthy target sits under 14. Day one to day seven belongs to the launch brief, which the PMM drafts against the PM's PRD and the design mocks, then bounces between four reviewers, none of whom own the sign-off.

Day seven to day fourteen belongs to the enablement package, which the PMM builds in Google Slides against a template last updated at Series A. Day fourteen to day twenty-one belongs to the launch-day content, which ships the blog post, the changelog entry, the email, the social copy, and the demo video across five surfaces with mismatched positioning.

Day twenty-one through day forty-seven belongs to the silence. The launch shipped, the announcement went out, and the AE talk track never updated. The battle cards never updated. The pricing page never updated. The onboarding drip never updated. The battle card queue already flagged the pattern in April. Four features shipped, one shows up in demo, one shows up in the last month of AE calls, and the pipeline attribution column reads zero.

Stage one: the launch brief that lives in six drafts

Search the shared drive for the last launch brief. Six files come back. Two are named v3, drafted by the PMM and the PM against different positioning. One is named "final for review" and predates the pricing change that moved the feature from add-on to bundled. One lives in the PM's private folder against a segment the CRO deprecated in June.

The Series B version of this function ships one canonical launch brief with a version tag, a change log, and a single approval line. Reads the PRD, the design mocks, the beta usage data from the product database, and the last three competitor releases in the same category before the PMM opens the draft. Writes the positioning against the two segments the CRO named in the last QBR, not the five the PMM inherited from the last cycle.

The stuck version ships six briefs and lets the reviewer pick. The Series B version ships one brief and locks the others. Same PMM, same headcount. The design team opens a request that already names the two hero images, the demo GIF spec, and the video length. Four rounds of "which brief is current" collapse to zero.

Stage two: the enablement package the AEs never opened

Every launch ships a slide deck, a one-pager, a Gong call snippet, an email template, and a Slack announcement. The stuck version drops all five into the enablement folder on launch day. Opens rate on the folder link in the trailing 30 days: 8 percent of the AE team. Two AEs opened, one bookmarked, none cited.

The stage-two function reads the Salesforce opportunity feed for the two ICP segments, the trailing 90 days of Gong calls, and the discovery-call template every AE runs. Rewrites the talk track injection against the actual language buyers used in the beta interviews. Ships the update into Chorus or Gong as a call snippet the AE hears in the coaching loop, not a slide the AE never opens.

Drops the demo-script update into the sales-enablement wiki with the two-line delta highlighted against the prior version. Pushes the pricing-page update into the CMS as a draft the CMO reviews Monday, not a Jira ticket that sits three weeks. The 47-day lag to first cited mention collapses to 9. The AE opens the discovery call Tuesday and the feature name lands in the third question, not month two.

Stage three: the launch retro nobody ran

Every launch ends with a Slack thread that reads "great launch team." The stuck version ships the launch and moves to the next PRD. The investor update references the feature by name and the board asks about pipeline impact. The PMM opens Gong, counts two mentions, and closes the tab.

The stage-three function reads the launch-day analytics against the four surfaces on day 7, day 14, and day 30. Counts AE mentions on Gong calls by feature name, by segment, and by call stage. Counts CRM opportunities with the feature tagged as a decision driver in the notes. Counts pricing-page traffic to the feature block and conversion from the block to the demo request.

Ships a two-page retro into the exec channel on day 30 with the four numbers, the segment breakdown, and the three enablement edits that would move the numbers next quarter. Loops the edits back into stage one on the next launch. The launch queue starts learning against its own output, and the fifth launch ships against the pattern the first four exposed.

The unit economics of a PMM hire against an agent stack

Run the two paths against the Senior PMM req sitting in the drafts folder at $165K base plus 15 percent bonus. Path A closes the req in October, ramps the hire through Q1, and adds a second PMM to the team owning launches, competitive, and pricing. Loaded year-one cost lands $240K to $290K. The 21-day path to launch collapses to 17, the 47-day lag to first mention collapses to 34, and the pipeline attribution column reads two features out of four.

Path B ships a three-sprint fractional AI product marketing function at $72K to $108K in build across the first 60 days and $5K to $8K a month to run. Sprint one lands the canonical brief lock and the beta-usage feed. Sprint two lands the enablement injector against Gong and the CMS. Sprint three lands the launch retro against the four-surface analytics.

The fractional AI department runs the launch queue on the cadence engineering ships against, not the cadence a PMM writes decks against. Loaded year one lands $132K to $204K plus a fractional PMM operator at $6K a month who owns positioning judgment and the retro read. Total $204K to $276K. Four launches ship with four cited AE mentions inside 14 days. The 14-day sprint that stands up the enablement injector costs less than the search retainer for the Senior PMM.

The four numbers a CMO runs before the next PMM req

The founders reading this are two weeks from posting a Senior PMM req on the strength of a "launch velocity" narrative. Before the offer letter goes out, run four numbers against the launch function, not the headcount. Score the queue, not the hire.

Launch-to-first-mention days. Measure the median days from feature-flag-on to the first AE citation of the feature by name on a Gong or Chorus call. A healthy function runs under 14. A stuck function runs 40 to 60, and the QBR slide reads zero pipeline attribution.

Talk-track update lag. Measure the days between launch date and the pricing page, battle card, and discovery-call script showing the new feature block. A healthy function runs under 7. A stuck function runs 30 to 90, and three surfaces still show the pre-launch positioning next quarter.

Enablement folder open rate. Divide the AEs who opened the launch enablement folder in the trailing 30 days by total AE headcount. A healthy function runs above 60 percent. A stuck function runs under 15, and the CRO stops treating the launch package as a source in the deal review.

Launches with a shipped retro. Count the launches in the trailing 90 days that shipped a written 30-day retro with cited-mention counts, opportunity attribution, and three enablement edits queued for the next launch. A healthy function ships one per launch. A stuck function ships zero, and the fifth launch repeats the pattern the first four exposed. Any two numbers in the stuck zone means the queue is the problem, and you can scope the launch function in a 30-minute call this week.