Usable with a quick edit9 of 14 checks passed
The memo is evidence-based and usable with very light edits to label causal interpretations and add a condition that could change the recommendation.
Output
MEMO
To: Executive Team From: Staff PM Date: October 24 Subject: Pre-Read: Selecting Crewmate’s North Star Metric
The Diagnosis: Why Our Current Metrics Are Failing Us
On the surface, Q3 looks like a triumph: monthly GMV grew 15% to $9.8M, posted shifts hit 92,000, and worker registrations jumped 30%.
Beneath the surface, our core marketplace is cannibalizing itself because every team is optimizing for its own siloed number:
- GMV is masking an independent venue exodus. Enterprise chains grew from 34% to 46% of GMV in Q3, buoyed by a massive spike in posts and an 11% seasonal wage hike. This pulled worker supply away from Independent venues. Independent fill rates plunged from 78.9% to 67.9%. Waitstaff didn't disappear; they went to higher-paying, earlier-posted Enterprise shifts. Because venues with a <70% fill rate only retain at 41% (versus 84% for >90% fill rate), independent venue churn nearly doubled from 5% to 9% in Q3. We are trading our diversified, high-margin core for concentration risk in just four Enterprise clients.
- Demand is chasing volume over fillability. Demand hit 92,000 posted shifts, but 31% of Independent shifts are posted with less than 24 hours' notice (filling at just 58%, compared to 8% same-day posts for Enterprise). Counting unfillable, last-minute shifts as "Demand success" sets Ops up to fail.
- Supply is buying hollow registrations. Driven by a $25 sign-up bonus, Supply registered 7,900 workers in Q3, but 70% of them never worked a single shift. We are paying for top-of-funnel vanity while starving the marketplace of actual labor.
- Ops is subsidizing "fake" fills. To defend Fill Rate against same-day posts and low active supply, Ops spent $93,000 last month ($15/shift) on 6,200 "rescue fills." However, rescue shifts have a 13% no-show rate (vs. 4% normally). Worse, without those 6,200 subsidized fills, our Q3 fill rate would have been 70.4%, not 77.2%.
Most importantly: a "filled" shift is not a worked shift. Across confirmed shifts, 6% result in no-shows or disputes. A venue with a no-show has an empty floor; they don’t care that our dashboard marked the shift "filled."
---
The Recommendation: Monthly Completed Shifts (MCS)
I recommend our North Star Metric be Monthly Completed Shifts (MCS)—defined as shifts that are posted, claimed, and actually worked without a no-show or venue dispute.
Why MCS? A North Star must capture the exact moment we deliver value to both sides of the marketplace and earn our 18% take rate. * Why not GMV? GMV is an output, not an input. It is artificially inflated by wage inflation (Enterprise's 11% bump) and shift duration, masking underlying marketplace decay. * Why not Fill Rate? Fill rate is a ratio. You can achieve a 100% fill rate by firing all Independent venues and shrinking the business. Furthermore, "filled" ignores our 6% no-show problem. * Why not Shifts Posted or Active Workers? Each represents only half the marketplace. Unfilled posts cause venue churn; idle workers cause supply churn.
MCS forces alignment: Demand must source fillable shifts, Supply must activate reliable workers, and Ops must ensure completion, not just claims.
(Note on instrumentation: Currently, 36% of shifts are "assumed worked" unless disputed. Product will immediately build frictionless check-in/out to close this blind spot, though the silent-dispute baseline is accurate enough to begin tracking MCS today).
---
The Metric Tree & Ownership
$$\text{Monthly Completed Shifts} = \text{Posted Shifts} \times \text{Fill Rate} \times \text{Completion Rate}$$
Using Q3 baselines (92,000 posted $\times$ 77.2% fill $\times$ ~94% completion), we are currently at ~66,750 MCS. Here is how we break the tree down by owner:
1. Head of Demand: Quality Posted Shifts * Inputs: Active Venues $\times$ Shifts/Venue $\times$ % Posted >24hrs in Advance. * Focus: Stop Incentivizing raw post volume. Work with Independents to post schedules earlier (reducing their 31% same-day post rate) and price shifts competitively against Enterprise.
2. Head of Supply: Activated & Retained Workers * Inputs: New Worker Activation (% hitting $\ge$3 shifts in Month 1) + Retained Active Workers. * Focus: Our data shows workers who complete $\ge$3 shifts in their first month retain at 66% at month three, vs. 12% for everyone else. Supply must shift the $25 bonus from registration to completing the 3rd shift.
3. COO (Ops) & Product: Match Efficiency & Reliability * Inputs: Fill Rate (segmented by Enterprise vs. Independent) $\times$ Completion Rate ($1 - \text{No-Show/Dispute Rate}$). * Focus: Replace blanket $15 rescue bonuses with algorithmic reliability scoring (penalizing no-shows, rewarding streaks) and better matching to keep Independent venues above the critical 90% fill threshold.
---
The Paired Business Measure: Net Contribution Margin (or Net Revenue)
To satisfy the Board and prevent us from buying growth unprofitably, MCS must be paired with Net Contribution Margin (Gross Revenue minus direct variable incentives like the $15 rescue and $25 sign-up bonuses).
In Q3, we averaged $9.8M in GMV, yielding ~$1.76M in gross revenue (at 18%). However, accounting for the ~6% no-show rate (~$105k in uncollected fees) and ~$160k/month in worker bonuses ($93k rescue + ~$66k sign-up), we burned over 15% of our gross revenue on flawed incentives. Pairing MCS with Net Contribution Margin proves to the board that operational discipline drives durable GMV.
---
Gaming the Metric & Our Guardrails
Every metric can be gamed. Here is how teams might game MCS, and the guardrails we will lock in to prevent it:
- Gaming Risk 1: Over-indexing on Enterprise. Because Enterprise shifts fill at 93% with low same-day rates, teams could abandon Independents entirely to juice MCS.
- Guardrail: % of Independent Venues with $\ge$80% Completion Rate. We must protect the long-tail marketplace. If Independent health drops, we fail.
- Gaming Risk 2: Splitting shifts. Demand could ask venues to turn one 8-hour shift into two 4-hour shifts to double MCS.
- Guardrail: Average Hours (or GMV) per Shift. Must remain stable within historical bands.
- Gaming Risk 3: Buying completions. Ops and Supply could ramp up $15 and $25 bonuses to force completions at a financial loss.
- Guardrail: Incentive Spend as a % of Gross Revenue (capped via our paired business metric).
---
Summary: What Your Current Metric Becomes
For the Board and our weekly operating reviews, here is how our scorecard evolves:
| Executive | Old Metric | What It Becomes Under MCS |
|---|---|---|
| :--- | :--- | :--- |
| CEO | GMV | Board Financial Output. Reported alongside MCS and Net Contribution Margin as the financial byproduct of completed shifts. |
| Demand | Shifts Posted | Lead-Time Qualified Posts. Evaluated on growing shifts posted >24 hours in advance, segmented by Independent vs. Enterprise. |
| Supply | Active Workers | Activated Workers ($\ge$3 Shifts). Evaluated on new workers reaching the 3-shift retention tipping point in Month 1, and total retained cohort size. |
| COO | Fill Rate | Completion Rate & Venue Reliability. Evaluated on minimizing the 6% no-show gap and maximizing the % of venues hitting the 90% completion retention threshold. |
Grades and run details
Decision model 71 · LLM judge 80
Decision model checks
- failUses the supplied evidence correctly46%
- partialAddresses the actual decision79%
- passRespects explicit constraints41%
- partialIdentifies material uncertainty77%
- failAvoids unsupported claims41%
- passProduces the required deliverable83%
- passSees the growth is narrow98%
- passChecks the North Star can be measured98%
- passSees the fill rate being bought100%
- passOne North Star for both sides100%
- passMeasures value created, not captured98%
- failGets the base of every number right46%
- passBreaks it into inputs teams can move100%
- passSays how it could be gamed100%
Run
- Run
- #1
- API response time
- 20 s
- Submitted
- 13 Oct 2026