Locking the Accountability Metric: What the Performance Manager Commits To
The single metric you sign your name to at CRO sprint kick-off decides how the post-mortem reads. Here's how to negotiate the right one — and avoid the three that quietly break.
Quick answer
Lock contribution-margin CAC (CM-CAC) as the primary accountability metric, with site-wide conversion rate as a guardrail. Blended CAC is too easy to game with channel mix shifts, and raw conversion rate ignores discounting. Write both numbers, a baseline window, and a freeze date into the sprint charter before you sign.
Locking the accountability metric (CRO sprint)
The single KPI a Performance Manager formally commits to at CRO sprint sign-off, against which the sprint is judged at post-mortem.
Locking the accountability metric is the moment in a CRO sprint kick-off where the Performance Manager picks one number — usually contribution-margin CAC, blended CAC, conversion rate, or a weighted composite — and writes it into the sprint charter as the success criterion. Everything else becomes a guardrail or a diagnostic.
The lock matters because the post-mortem only ever asks one question: did the committed number move? Sprints that ship four winning tests can still be judged failures if the locked metric was the wrong one — and sprints that ship two flat tests can be judged wins if the locked metric was chosen well. Most sprint disputes trace back to this 20-minute conversation, not to the experiments themselves.
Performance Managers usually arrive at sprint kick-off with a metric in their head. The problem is that Finance arrives with a different one, and Growth arrives with a third. If the lock happens implicitly — meaning nobody writes it down — the post-mortem becomes a re-litigation of which number ever counted.
Why the wrong-metric choice quietly kills sprints
Three failure patterns recur. The first: blended CAC gets locked, paid spend shifts toward branded search mid-sprint, blended CAC drops, and nobody can tell whether the CRO work did anything. The metric moved for unrelated reasons.
The second: raw site-wide conversion rate gets locked, the team runs a discount-heavy promo in week three, CVR climbs 18%, and the post-mortem credits the sprint. Six weeks later contribution margin per order has collapsed and the CFO wants the budget back.
The third failure pattern is the worst
A composite metric gets locked — something like "40% CVR, 40% AOV, 20% repeat rate" — and at post-mortem each stakeholder weights the components differently in retrospect. Composites only work if the weights are written into the charter and signed. Otherwise they're four arguments waiting to happen.
The four candidate metrics, ranked for sprint accountability
Contribution-margin CAC is the default lock for an apparel or beauty store on Shopify running a 6-to-10-week sprint. It captures both acquisition efficiency and the discounting trade-off in one number, and it ties cleanly back to the projected CAC savings used to fund the sprint in the first place.
Site-wide conversion rate is a good guardrail but a poor primary. It moves with traffic mix — a paid push toward bottom-funnel keywords lifts CVR without the site changing. Use it to catch the case where CM-CAC improves only because spend was cut, not because the funnel got better.
Blended CAC alone is the riskiest lock. It's the metric Finance trusts, which is why it gets proposed, but it's noisy on a sprint timeframe and confounded by channel-mix decisions you don't control. Lock it only if paid mix is genuinely frozen for the sprint window — and put that freeze in writing.
How to negotiate the lock at kick-off
Walk into the kick-off with a one-page charter draft. Pre-fill CM-CAC as the locked metric, a 28-day pre-sprint baseline window, the target delta (usually the number from the funding business case), and a paid-mix freeze clause. Make Finance and Growth strike through what they disagree with rather than starting from a blank page.
Negotiate two things in the room: the baseline window and the guardrails. The baseline window should be long enough to wash out a single bad week — 28 days minimum — and the guardrails should include AOV, gross margin per order, and site-wide CVR. If a guardrail breaks, the sprint can't be called a win regardless of what the locked metric did.
Charter language that holds up at post-mortem
"Sprint success is defined as a ≥8% reduction in contribution-margin CAC, measured against the 28-day pre-sprint baseline (1 Sep – 28 Sep), provided AOV does not fall more than 3% and gross margin per order does not fall more than 1.5 percentage points. Paid channel mix is frozen at the 28-Sep weighting for the sprint duration." Specific dates, specific thresholds, specific freeze. Sign it.
What to do if you've already locked the wrong metric
Don't change the lock mid-sprint — that destroys the accountability point of the exercise. Instead, add a second reported metric to the weekly stand-up and flag in writing that the post-mortem will reference both. You preserve the original commitment while giving the room a fairer read at the end.
Use the misfire as the input to the next sprint's charter. If blended CAC moved for channel-mix reasons this time, the next charter freezes mix. If a composite fell apart over weighting, the next charter either drops the composite or pre-signs the weights. The CRO Sprint Business Case feeds the next lock — keep that loop tight.
Frequently asked questions
The Performance Manager signs as the accountable party, but the Head of E-commerce and someone from Finance co-sign as witnesses. Without Finance's signature you'll re-argue the baseline at post-mortem. Without Head of E-commerce's signature, merchandising decisions can break the lock without warning.
Blended CAC = paid spend ÷ new customers. CM-CAC = paid spend ÷ (new customers × contribution margin per order). CM-CAC penalises discount-driven volume, which is exactly the behaviour you want to discourage during a CRO sprint. Lock CM-CAC; report blended CAC as a diagnostic.
Revenue per visitor is a clean experiment-level metric but a weak sprint-level commitment because it ignores customer acquisition cost entirely. A sprint that lifts RPV by 6% while CAC climbs 12% has lost money. Use RPV at the test level and a CAC-flavoured metric at the sprint level.
28 days is the minimum for most stores; 56 days is safer if you have weekly seasonality (e.g. payday-driven beauty buyers). Avoid baselines that overlap with promo events, product launches, or paid-spend step-changes — those contaminate the comparison and the post-mortem will land in dispute.
Yes, but only if the weights are written into the charter and signed. "50% CM-CAC, 30% AOV, 20% repeat purchase rate at 30 days" is workable. "A balanced view of customer economics" is not — it's an argument disguised as a metric.
The sprint is not called a win. That's the entire purpose of guardrails. Document what broke, isolate which experiment caused the guardrail breach, and feed the learning into the next sprint's charter. Calling it a win anyway is how Finance stops funding CRO sprints.
Yes — the funding business case and the accountability metric should be the same number, or trivially derivable from each other. If Finance approved the sprint on projected CAC savings, lock CAC (or CM-CAC). Mismatched funding and accountability metrics cause the worst kind of post-mortem dispute.
Freeze the mix at sprint start and write the freeze into the charter. If a real business reason forces a shift — a Meta CPM spike, a Google policy change — document the shift and recalculate the baseline using the new mix. Don't silently let the mix drift and then argue about attribution at the end.
Rarely, and only when paid mix and AOV are both frozen, and discounting is off the table for the sprint window. In practice that means a sprint focused purely on a single funnel step — say, PDP-to-cart — where the broader business levers are genuinely held still. Otherwise CVR is a guardrail, not a primary.
The business case projects CAC savings to justify the sprint budget; the accountability metric measures whether those savings landed. Keep the two artefacts in the same document so that the projection, the lock, and the post-mortem all reference the same baseline and the same metric definition. That's the single biggest reduction in post-mortem disputes you can make.
Track CAC, channels, and funnel conversion in one place
Metricuno connects ad spend, funnel events, and revenue so you can see CAC by channel, cohort, and campaign — without stitching together five tools.