Marketing audit in seven days: checklist and worked example

1. Day one: agree on the decision

Choose one product, market and sales journey. An online store and a B2B service need different evidence. State the decision the audit should support: whether to raise paid acquisition spend, why lead quality declined or what prevents inquiries from becoming customers. A broad request to review marketing easily turns into a presentation covering everything and deciding little.

Collect revenue, contribution margin, acquisition spend, new customers and the rules used to count them. Mark unavailable data explicitly. Do not substitute an industry average without labeling the assumption. Assign a sales contact and an owner for account access. If essential exports will arrive late, narrow the promised scope rather than infer precise conclusions from a few dashboard screenshots.

2. Day two: reconcile the funnel

Describe the stages from visit or inquiry to payment, including the unit and period of each count. Website events, unique people and CRM deals are often mixed. Confirm that stages refer to a comparable cohort before calculating conversion rates. An open funnel and a closed funnel answer different questions, so save the chosen definitions with the result.

Consider an illustrative US service example: 1,000 visits, 50 inquiries, 20 qualified leads and four sales on $4,000 USD in ad spend. Cost per inquiry is $80, per qualified lead $200 and per sale $1,000. These are invented teaching figures, not benchmarks. Without margin and other acquisition costs, the table cannot establish profitability. It identifies the next transition worth investigating.

3. Days three and four: review channels and promises

Separate channels by their intended job and the reliability of their measurement. Last-click reporting may understate content that introduces the product. That does not make an unsupported awareness claim sufficient evidence of value. For each channel, name the expected user action and the method used to observe it. Preserve uncertainty where the reporting cannot answer the question.

Then compare the ad, email or post with its destination. Do the product, terms, audience and next action agree? Review expensive and productive examples, not only the team's favorite creative. Capture the exact wording or screenshot behind a finding. Weak design is hard to act on; the mobile page hides the paid consultation requirement is a specific problem someone can verify.

Completed audit check: cheap leads with a costly qualification gap

This original worked check uses synthetic campaign exports for October 1–7, 2026. All money is illustrative USD. Campaign A spent $600 and delivered 30 unique inquiries; campaign B spent $400 and delivered 10. Sales reviewed the same inquiry cohort and marked six in each campaign as qualified using one agreed rule. This is not a result from Daniil’s client work.

Campaign A has CPL $20 and cost per qualified lead $100. Campaign B has CPL $40 and cost per qualified lead $66.67. Combined spend is $1,000, total inquiries 40 and qualified inquiries 12: blended CPL $25 and blended cost per qualified lead $83.33. The cheaper raw lead source is more expensive at the next stage.

Evidence packet: download the synthetic campaign spend export and the 40-row inquiry and sales-review cohort. The rows reproduce the inputs in this demonstration and contain no real customer records. In your own audit, preserve secured raw exports, use non-personal inquiry IDs and document every filter before drawing a conclusion.

Completed audit row: finding = low CPL conceals lower qualification; evidence = the matched cohort above; confidence = descriptive of these 40 inquiries only; owner = paid acquisition lead plus sales reviewer; decision = inspect Campaign A’s promise and qualification reasons before expanding it. The next experiment can test a clearer message under a stated spend cap. No statistical superiority or future return is established by this small example.

Acceptance check: recalculate both metrics, inspect rejected inquiry reasons and verify that both campaigns have had equal time for review. If Campaign B’s leads mature earlier or the spend exports use different attribution windows, the comparison is incomplete. Put that uncertainty in the audit row instead of ranking channels prematurely.

Download the synthetic audit source exports · Check your funnel assumptions · Reconcile conflicting lead counts · Discuss your audit findings

4. Days five and six: follow the inquiry

Complete the main journey on a phone and a desktop: arrival, offer, form, confirmation and receipt by the responsible team. A successful server response is only one part of that experience. The visitor needs a clear confirmation, and sales needs a usable record with a source and next action. Include external booking or checkout transitions in the scenario.

Compare intended events with actual submissions and CRM records. Sample response delays, loss reasons and incomplete statuses. The funnel review tool on this site analyzes the stages and figures supplied by the user. It does not connect to your CRM or certify an export. Remove personal information and use its suggestions as questions to investigate, then confirm them against the underlying records.

5. Day seven: select three actions

For each candidate, record potential business impact, confidence and effort. A numerical priority score can organize a discussion, but it cannot turn a guess into proof. A reproducibly broken form generally deserves attention before a button-color experiment. Where tasks depend on each other, complete the measurement or functionality prerequisite first.

A useful action card contains the observation, evidence, proposed change, owner and acceptance criterion. For example, a mobile form fails after a specific field selection; the developer owns the repair; acceptance requires a successful inquiry and CRM record on two devices. Evaluate commercial outcomes separately after deployment. Restoring the form is a technical result, not proof of a particular revenue increase.

6. Avoid an audit that ends in a folder

Common mistakes include examining every business line at once, comparing incompatible periods, treating preferences as evidence and presenting hypothetical revenue as a measured loss. If you cannot know how many visitors would have purchased after a repair, say so. You can model a scenario, but retain its assumptions and do not promote it into historical fact.

Deliver the three immediate actions plus a separate list of questions requiring more investigation. Give each a review date and an accountable owner. A month later, reopen those original cards instead of producing a new presentation from scratch. The practical value of the audit lies in a clearer decision and a checked change. The downloadable board keeps the evidence and follow-up attached to the work.