Metrics

Built around Meridian Intelligence, the hypothetical company on the Case page. The metrics, definitions, and views below are my own designs for this case and all figures are illustrative. Sources are at the bottom of the page and on the Sources page.

In brief: every program and lifecycle stage has a measure. This page starts from the decisions leaders need to make, records a baseline before anything launches, keeps the headline list to five, and holds the one definition of every measure on the site. The first version runs on shared definitions and tidy inputs, with dashboards later, because Meridian can’t afford custom reporting. Short on time? Read section 1 and section 4.

1. Start from the decision

Three numbers anchor everything, and they match the Onboarding page: time to productivity (how fast a new hire reaches independent, useful work), ramp-quota attainment (whether reps hit quota on the committed ramp schedule), and retention (whether people who complete onboarding stay and perform past year one, and whether customer accounts renew).

Metrics work backward, as Kirkpatrick recommends: the business result first, then the behaviors that produce it, then the learning behind them. They come in three kinds. Readiness indicators ask whether people are ready (certification pass rates, role-play scores, the New Hire Pulse). Leading indicators ask whether they follow the process (calls meeting the discovery standard, handoff briefs completed, coaching sessions held). Lagging indicators ask whether results are arriving (win rate, quota attainment, retention). The first two move first, so they are the early warning.

Meridian runs on a handful of KPIs that enablement should agree with leaders before anything launches. If they hold different views of how enablement will be judged, even good results get discounted. The leadership scorecard on the Collaboration page rates each defining objective green, yellow, or red. Each objective gets one measure that moves first, with a lagging check beside it. If these five improve, the three outcome numbers should follow:

Everything else is diagnostic. It explains why a headline number moved, and it isn’t reported upward unless it changes a decision.

How each Case target is tracked

2. Definitions by lifecycle stage

Every lifecycle measure has one written definition, an owner, and a source. Enablement owns the program measures (certification pass rates, coaching activity, content freshness, messaging consistency). The functions that run each stage own the lifecycle measures. The last column shows how much enablement drives each one, and the attribution caution in section 4 applies to every “contributes.”

Average deal size and sales velocity are useful context. Velocity can look strong because of a few large deals, so report the four inputs beside it. Also watch for warning behaviors in CRM history: deals advanced without meeting exit criteria, deals that move backward more than once, late-stage deals whose amount keeps changing, and late-stage deals created by mistake. They are reviewed in coaching, not used for discipline.

3. Program measures and what people tell us

Each program page has its own measures: Onboarding (milestones, gates, day-one readiness), Training (pass rates, skill gap closure, drop-off), Coaching (sessions held and documented, scorecard results, the player-coach signal), and Collaboration (content use, messaging consistency). Skill gap closure is (new score minus starting score) divided by (target score minus starting score), on the same criteria before and after. A rep who moves from 1.5 to 2.5 against a target of 3 has closed about 67% of the gap.

Three inputs sit beside the numbers:

• Sales confidence survey: a 1 to 10 survey on pitching, competitors, and assets. Quarterly, but not in the last weeks of a quarter, when responses drop.

• Satisfaction with enablement: a rating for each deliverable, so weak ones are visible.

• Win/loss interviews: buyers who chose Meridian or someone else, as deals close. Open-ended comments from these, the Pulse, and churn debriefs are coded with the Interview Coding Workbench, so the “why” behind a number is visible.

Survey results are reported only for groups above a minimum size, and no individual’s answers reach their manager or their performance review.

4. Keep the inputs clean, and read the numbers honestly

Reports are only as good as the records behind them. That means one definition per metric with a named owner, dates recorded every time, structured certification and coaching records, consistent recording and consent, a monthly data check, and a change log for every program change.

Meridian is small. With about six AEs and three CSMs, one deal or one departure moves a percentage a lot. So:

• Report counts next to percentages, and pair every number with what people say.

• Look at the spread, not just the average. With six AEs, show each rep’s result and describe the pattern in words.

• Baseline before launch, and compare each cohort to it.

• Keep one definition for each ratio, so the same win rate can’t look different in two reports.

• A program is a contributor, not the sole cause. Pricing, seasonality, manager quality, and process changes also move results. Staggering a rollout gives a comparison group where possible.

• A fast first sale isn’t proof. Time to second and third deal shows whether it lasts.

• Activity isn’t impact. Sessions held tell you enablement is busy, not that it worked.

5. The dashboard: start small

RevOps has no capacity for custom reporting, there’s no LMS, and the budget covers roughly one new tool. So the first version is shared definitions, then spreadsheets and CRM exports for the first two quarters, then dashboards once definitions are stable. The one-tool budget goes where it removes the most manual work. My working hypothesis is call review and scoring (see the Roadmap).

Design choice: five headline measures, spreadsheets before dashboards. The alternative was a full dashboard from day one. Definitions have to settle before anything is automated. It would change once definitions have held for a few quarters.

Five views serve the audiences: a leadership scorecard (quarterly, with a monthly glance), a manager view (weekly), a cohort tracker (at days 30, 60, and 90), an account health view (weekly), and a program view (monthly). The Collaboration page has the cadence for each.

6. Enablement ROI

ROI = (gain minus cost) divided by cost, times 100. The gain is estimated from measured changes, such as revenue reached sooner from a shorter ramp. The cost is salaries, tools, content, and training time. With a team this small the result is lumpy, so report a range, state the assumptions, and use conservative ones.

7. Where AI helps

Three tools from the Real Work page fit into this design:

• Opportunity Grading Agent (designed): scores opportunity quality from a consented transcript and fills the CRM fields, so reviewed calls feed the metrics without manual entry. It builds on a transcript-to-CRM agent I built at a previous organization.

• Interview Coding Workbench (built and tested): codes churn debriefs, win/loss interviews, and Pulse comments.

• Smart Brevity Summary Generator (built in a past role, rebuilt as a Claude project): turns long reports into leadership summaries.

AI reduces the manual work. Humans keep the judgment: what a metric means and what to do about it.