Your Roadmap Intake Has 340 Feature Requests and Engineering Shipped Twelve in Q3

It is Wednesday, 11:04 AM. Your Head of Product opens Productboard with the second coffee. The intake queue reads 340 requests logged since January. The Q3 shipped column reads 12 features. The "waiting on PM triage" tag reads 187 items, oldest logged March 12th by an AE who left the company in June.

She opens Slack and searches #product-requests for the trailing 30 days. Forty-seven new requests. Twelve forwarded from execs with a one-line "customer wants this" and no account attached. Eighteen tagged by CSMs against a specific renewal book. Four flagged as a churn risk by an AE with the deal review calendar link. The Gong tag "customer requested" ships 34 more per week straight into a bucket the PM opens Friday afternoon if the sprint review runs short.

The founders in this seat keep opening a Senior PM req at $195K and wondering why engineering ships from a plan the CRO argued in the board call last quarter. Roadmap intake is not a headcount. Roadmap intake is a six-stage queue that starts with a customer signal captured in the wild and ends with a shipped feature closed back to the account that asked. Five teams touch the queue and none of them own it end to end.

The 94-day path from customer signal to shipped feature

Sort the roadmap ledger by signal-received date. Median days from a customer signal captured to a shipped feature landing in a release note reads 94 against a Series B target of 42. Day one to day fourteen belongs to the capture window, where the signal lives in a Gong tag, a Slack thread, a Salesforce opportunity note, or a Notion page the CSM wrote after a QBR, and none of the four surfaces sync back to Productboard.

Day fourteen to day thirty-five belongs to the triage backlog, where the PM opens the intake queue Friday, sorts the 340 items by newest, reads the top 40, tags 12 with "duplicate," six with "need account context," four with "escalate," and closes the tab. Day thirty-five to day sixty belongs to the RICE cycle, where the PM rebuilds the prioritization sheet against six factors nobody agreed on, argues the top ten with engineering on Monday, and loses two slots to a CRO ask that landed Sunday night.

Day sixty through day ninety-four belongs to the build-and-ship window, where engineering scopes the top three, cuts the fourth for a platform migration, and ships the release note to a template the CSM finds Wednesday morning. The account that filed the original request in April never hears the feature landed. The renewal book carrying 60-plus balances opens the QBR asking whether the ask made the cut and the CSM checks the release notes on the call.

Stage one: the intake channels that never converge

Open the request ledger. 340 items in Productboard. 187 tagged "waiting on PM triage" since March. Cross-reference the CRM and Gong. The AE team logged 214 opportunity-linked feature notes in the trailing 90 days, and 84 of them never made it into Productboard because the workflow reads "manual PM copy on Friday." The CSM team logged 147 QBR action items with a product ask, and 62 sit in Gainsight custom fields the intake queue never reads.

The Series B version of this function reads every intake surface on the day the signal lands. Pulls the Gong tag, the Salesforce field, the Gainsight action, the Slack thread, and the exec forward into a single intake object with the account, the ARR, the AE, the CSM, the renewal date, the churn risk, and the verbatim customer quote attached. Deduplicates against the existing 340 items in Productboard by embedding similarity, not by keyword match on a title the AE typed in a hurry.

Flags any signal filed against an account carrying more than $80K ARR with a renewal inside 90 days for PM read the same day, not the Friday six weeks out. The 340-item queue collapses to 187 unique items inside a cycle, and the PM stops sorting duplicates on Friday afternoon. Every signal lands with the account context attached before the PM opens the tab.

Stage two: the RICE sheet the PM rebuilds every quarter

Every quarter the PM opens a fresh sheet with six columns: reach, impact, confidence, effort, ARR at stake, exec heat. Fills 40 rows by hand. Argues the top ten with engineering on Monday. Cuts two for a CRO ask that landed Sunday night on Slack. Loses another to a customer the CEO met at SaaStr. The sheet dies Friday and the next quarter opens a fresh copy against a different 40 rows.

The stuck version scores the sheet against the PM's memory of last week's deals. The Series B version scores every intake object against the live pipeline, the renewal book, the churn signal, and the last four quarters of shipped-feature attach rates. Pulls reach from the account book with the feature request tagged. Pulls impact from the ARR at stake and the win-rate lift the sales team saw on the last three deals that referenced the ask. Pulls effort from engineering's actual scope estimate against the last five features of the same shape, not the PM's guess Monday morning.

Publishes the ranked list to the sprint planning doc the engineering lead reads Sunday night, with the second-order dependencies flagged and the two exec asks explicitly scored against the same rubric. The CRO stops arguing the roadmap in the boardroom because the sheet already scored the ask two weeks ago. The PM stops rebuilding the sheet Monday morning and starts writing the three specs the sprint needs.

Stage three: the close-the-loop nobody runs

Every shipped feature ships a release note into a Notion page the CSM finds if the sprint review runs long. The account that filed the original request in April never hears the feature landed in September. The renewal conversation opens with the CSM promising the ask is "on the roadmap" against a feature that shipped three weeks ago and sits behind a flag nobody flipped for the account.

The stage-three function reads every shipped feature against the intake objects it closed. Ships a one-paragraph note to the AE and the CSM inside the hour the release lands, with the accounts that asked, the exact request quote, the flag status, and the enablement doc the CSM sends. Files an outreach task in the CS tool against every account that logged the ask, with a draft message referencing the QBR the CSM ran and the release timing.

Reopens the intake object if the flag rollout misses an account for more than 14 days and pings the PM with the accounts still waiting. The case-study pipeline draws from the same closed loop the CSM never ran, and the launch story lands with real reference accounts named on the page.

The unit economics of a Senior PM hire against an agent stack

Run the two paths against the Senior PM req sitting in the drafts folder at $195K base plus 15 percent bonus and equity refresh. Path A closes the req in October, ramps the hire through Q1, and adds a second PM body owning the intake queue against the same 340-item backlog. Loaded year-one cost lands $268K to $324K. Median signal-to-ship stays at 84 days because the PM still opens the queue Friday and the CRM still fails to sync on the day the AE logs the ask.

Path B ships a three-sprint fractional AI product function at $72K to $108K in build across the first 60 days and $5K to $7K a month to run. Sprint one lands the intake convergence layer against Gong, Salesforce, Gainsight, Slack, and the exec forward inbox. Sprint two lands the scoring engine against the live pipeline, the renewal book, and the historical attach data. Sprint three lands the close-the-loop routing against the CS tool and the AE task queue.

The fractional AI department runs the roadmap intake queue on the cadence the pipeline forecast runs, not the cadence a PM can chase between sprint reviews. Loaded year one lands $132K to $192K plus a fractional Head of Product at $8K a month who owns the roadmap judgment, the exec conversation, and the spec sign-off. Median signal-to-ship collapses from 94 days to 46. The Q3 shipped count moves from 12 to 21 against the same engineering headcount because the sprint plan stops absorbing the CRO's Sunday-night ask. The 14-day sprint that stands up the intake convergence layer costs less than one recruiter contingency fee on the PM search.

The four numbers a Head of Product runs before the next PM req

The founders reading this are three weeks from posting a Senior PM req on the strength of a "we need to scale product" narrative. Before the offer letter goes out, run four numbers against the roadmap intake function, not the headcount. Score the queue, not the hire.

Signal-to-triage days. Measure the median days from a customer signal logged in Gong, Salesforce, or Gainsight to a matched entry in Productboard with an account attached. A healthy function runs under three days with the sync running on the intake object, not the PM's Friday sweep. A stuck function runs 21 to 42, and 40 percent of AE-logged asks never reach the queue at all.

Orphaned-request percentage. Divide the intake items with no account, no ARR, and no owning AE or CSM by total intake items in the trailing 90 days. A healthy function runs under 8 percent. A stuck function runs 35 to 55, and the PM sorts duplicates against a sheet the CRO argues from a different copy of.

Close-the-loop rate. Divide the shipped features that reached every requesting account with a note, an enablement doc, and a flag flipped within 14 days by total features shipped in the trailing 90 days. A healthy function runs above 85 percent. A stuck function runs under 20, and the CSM opens the QBR promising a feature that shipped three weeks ago.

Roadmap-to-forecast alignment. Count the top-ten roadmap items where the reach and impact scores reference the live pipeline and the renewal book, not the PM's memory. A healthy function runs ten out of ten. A stuck function runs three out of ten, and the CRO argues four slots back onto the plan every board meeting. Any two numbers in the stuck zone means the queue is the problem, and you can scope the roadmap function in a 30-minute call this week.