How to read a cohort report.

Before the revenue review, open a cohort row and trace an account back to its entry event. Check the date, source, opportunity, and current status. A report that cannot explain a single row needs a data review before it can support a budget decision.

Once that connection is sound, read the report in a fixed order: definitions, completeness, age, outcomes, economics, and action. The sequence keeps an attractive percentage from deciding the meeting before anyone checks what it contains.

If you need the rationale for this view, start with why we measure cohorts. This guide covers the operating review.

Ask RevOps to put the report's rules beside the report. A metric contract is a short agreement about what each field means. It should cover:

• Membership: the account or person being counted, the entry event, entry window, exclusions, and deduplication rule.
• Time: the reporting cutoff, timezone, elapsed-age calculation, and whether each metric is cumulative or a current-state snapshot.
• Outcomes: the definition of an accepted opportunity, closed-won status, closed-lost status, and the amount field used for bookings or revenue.
• Attribution and cost: the source assignment rule, influence criteria, unknown-source treatment, and included costs.

Name the denominator for every rate. Accounts that generated an opportunity divided by acquired accounts is an account-to-opportunity rate. Won opportunities divided by decided opportunities is a decided-opportunity win rate. Labeling both as conversion rate hides a material difference.

Keep the definitions stable across the comparison. Annotate changes that require a new baseline.

Reconcile cohort membership with the underlying entry records. Check that opportunity links, stage histories, and cost imports are complete through the stated cutoff. Confirm that duplicate contacts from one account do not create duplicate wins.

Show the counts and share of records missing each required field. Missing acquisition source should remain unknown. Missing revenue should remain unknown. A zero belongs in the report only when the relevant records were checked and the measured result was zero.

Separate observed history from reconstructed history. If the system only stores current stages, it may be unable to tell you where an account stood at an earlier age. Label that limitation and repair it before claiming an equal-age comparison.

Do not rank channels whose data is materially incomplete. Assign an owner to the missing join or import first.

Compare cohorts at the same age from each member's entry event. A cohort acquired earlier can have a longer lifetime view, but that view cannot fairly stand beside a newer cohort's shorter observation window.

Use only members who have reached the chosen age, and display the eligible count. If part of a monthly cohort is still too young, label the result partial or wait until all members are eligible. Never fill future-age cells with zero.

A fictional worked example

Every number below is fictional and exists only to demonstrate the calculation. Entry means the first qualifying inquiry from a net-new account. The example assigns each account once to its earliest recorded eligible acquisition touch before that inquiry, with complete source records assumed.

Both cohorts contain accounts observed for exactly 90 days after entry. This is an example window, not a recommended sales-cycle target. Each account has at most one accepted opportunity in this example.

• Acquisition cohort: January; Accounts: 40; Accepted opportunities: 8; Closed won: 2; Closed lost: 2; Still open: 4.
• Acquisition cohort: February; Accounts: 60; Accepted opportunities: 9; Closed won: 3; Closed lost: 1; Still open: 5.

All statuses are measured at the same elapsed age. The remaining accounts have not created an accepted opportunity by that point. Their eventual outcomes are unresolved.

January's account-to-opportunity rate is 8 / 40 = 20%. February's is 9 / 60 = 15%. Each cohort's acquired-account-to-won rate is 5%: 2 / 40 and 3 / 60.

January produced a higher opportunity rate at this age. The account-to-won rate is the same. Neither observation tells you what the open opportunities will do.

In the example, January's eight opportunities reconcile to two won, two lost, and four open. February's nine reconcile to three won, one lost, and five open. Open opportunities stay outside the loss count.

The decided-opportunity win rates are 2 / 4 = 50% and 3 / 4 = 75%. Those percentages describe only the decided subset. They leave out unresolved opportunities and are based on small counts, so using them as final cohort win rates would overstate what is known.

For an influenced report, inspect the qualifying interaction and when it happened. A webinar attended during an existing opportunity does not automatically become that opportunity's source. Keep sourced wins separate from influenced wins, and deduplicate opportunities before combining programs.

Recorded association does not establish that a touchpoint caused a sale. If incremental impact is the question, use an appropriate experimental or comparison design with its limitations documented.

Suppose the fictional cohorts have assigned media costs of $12,000 for January and $18,000 for February. Media cost per won account to date is $12,000 / 2 = $6,000 and $18,000 / 3 = $6,000.

These figures exclude production, technology, sales effort, and delivery costs. They are media-cost measures, not fully loaded customer acquisition costs. Further wins or allocated costs can change them.

For the combined account-to-opportunity rate, add the underlying counts: 17 / 100 = 17%. The simple average of 20% and 15% is 17.5%, which gives the smaller cohort equal weight. Use the combined counts whenever the populations and definitions are compatible.

Inspect contract size, margin, and cash timing separately before judging affordability. Pipeline value belongs in its own field. It is not realized revenue.

Read the account histories behind the change before choosing a fix.

• If suitable accounts enter but fail to reach accepted opportunities, review the teaching asset, follow-up, qualification, and sales response together.
• If opportunities progress but approvals remain open, ask sales for the next buyer-owned milestone and evidence that it is active.
• If mature cohorts produce unacceptable economics under complete cost accounting, revisit the audience, offer, delivery model, or channel allocation.

Record the decision, owner, evidence gap, and next review condition. Preserve the definitions and snapshot so the next meeting can test what changed.

For help connecting the report to an education-led acquisition system, bring the cohort rules and the decision currently blocked by the data. Get in touch.