# Marketing measurement checklist

By Tugelbay Konabayev · 4 November 2026

Use this to inspect one acquisition path before spending more on it. Record the answer, source, date and person responsible for each check. An unknown answer is a follow-up task, not a failed business.

1. **Choose one customer problem.** Name the buyer, the situation that makes them act, and the result your offer promises. Write down what the offer excludes.
2. **Agree on the stages.** Distinguish a visitor, a form start, an accepted enquiry, a qualified opportunity, a paid order and a retained customer. Do not call all of them leads.
3. **Inspect the measurement window.** Use complete days and a consistent timezone. Mark tracking migrations, outages and changes to event definitions before comparing periods.
4. **Separate unusual traffic.** Look at source, country, device, session volume, engagement and duration together. Keep raw totals. A suspicious segment needs investigation; it does not prove that an entire country consists of bots.
5. **Check search demand by page and query.** Keep property totals, page totals and visible-query reports separate. Anonymous queries and different aggregations can prevent simple subtraction.
6. **Follow one enquiry.** In a permitted test environment, verify validation, server acceptance, the CRM record, notification acceptance and delivery status. Force one provider to fail and repeat the request. Record what survives.
7. **Reconcile the money.** Match accepted orders with payment receipts, refunds and actual payouts. Keep pending affiliate commissions separate from approved commissions and cash received.
8. **Connect content to a next step.** Choose a useful action that fits the reader's task. Measure whether they saw it, used it and became a suitable enquiry. A click on a free tool is not a sale.
9. **Define activation for a product.** Identify the first useful result and repeated use. A store click does not prove installation; installation does not prove retention.
10. **Separate AI evidence.** Record a page citation, a brand mention, a recommendation and an actual referral as different outcomes. Preserve the question, engine, date and response; do not add vendor samples into one audience total.
11. **Check the claim behind a case study.** Identify the period, baseline, method, personal versus team contribution, permitted evidence and limitations. A tidy dashboard is not proof of customer impact.
12. **Choose the next small improvement.** Specify the observed problem, expected change, effort, owner and a check that can show failure as well as success. Stop changes that do not improve a meaningful outcome.

Keep a compact weekly record: qualified enquiries, proposals, paid work, gross margin, delivery hours, customer outcome, and the next unresolved measurement question.

For a scoped review of your acquisition, analytics and CRM handoff, see the [Revenue Leak Audit](https://konabayev.com/services/ai-marketing-audit/). Scope, availability and terms are confirmed before payment; no ranking or revenue result is guaranteed.