A useful campaign report shows how interaction became a valid enquiry, how an enquiry became a sales-accepted opportunity, and what remains unknown. Clicks and form costs can help diagnose a campaign, but they are not enough to authorize more spend. Equally, an agency should not be blamed for a sale that nobody followed up or recorded.
Before commissioning a new report, agree the stage definitions and the source of each number. The missing work may be status recording rather than a more attractive dashboard. Paid Campaign Strategy and Management should connect optimization decisions to business objectives while keeping measurement limits and delayed outcomes visible.
Stop calling every stage a lead
Create a short vocabulary approved by marketing and sales. An event is a measured interaction. An enquiry is a record that can be reviewed. A valid enquiry may still be unsuitable for the offer. A sales-accepted opportunity requires a documented decision. A completed sale requires your chosen commercial evidence, not merely a quotation sent.
| Stage | Proposed working definition | Evidence owner |
|---|---|---|
| Form event | The specified measurement signal | Measurement owner |
| Unique valid enquiry | Non-duplicate record with usable information | Record-quality owner |
| Accepted opportunity | Suitable need accepted under the agreed rule | Sales owner |
| Commercial outcome | Documented won, lost or held status | Opportunity owner |
These are proposed definitions, not universal industry standards. Google Ads describes qualification as occurring in the CRM or internal lead system, with conversion based on a business-defined step. Platform configuration does not replace commercial assessment. Qualified and converted leads.
Reconcile the numbers before drawing the charts
A report should explain how many records were duplicates, tests or unusable messages; how many have not been reviewed; and how many opportunities remain open. Different systems can legitimately show different totals. An unexplained difference is what prevents a reliable conclusion.
Marketing Automation and Lifecycle Journeys can include a conversion-tracking audit and approved repairs without requiring a complete automation programme. Start with the connections needed for the decision. Where systems are separate, define permitted access, the minimum required fields and the owner of delayed or failed transfers.
Do not include customer names in an executive report merely because they are available. Record identifiers, stages, sources and necessary dates may be sufficient. Keep personal details with authorized owners and define a permitted sample-audit route when investigation is needed.
Worked hypothetical reconciliation
These are teaching figures, not an actual campaign or a market benchmark. An illustrative month records 80 form events. Review identifies 11 additional events associated with duplicate requests. Eight of the remaining unique records are tests or unusable messages. That leaves 61 unique valid enquiries: 80 − 11 − 8 = 61.
Sales reviews 43 of them, accepts 19 opportunities and rejects 24 as unsuitable. Eighteen remain unassessed. Acceptance among reviewed enquiries is therefore 19 ÷ 43, approximately 44.2%. Accepted opportunities as a share of all valid enquiries is 19 ÷ 61, approximately 31.1%. Both calculations are correct, but they answer different questions. The first must not imply that every enquiry has been assessed.
Within the 19 opportunities, suppose four are won, three lost and 12 still open. Open opportunities are not losses. Nor should the report add 19 accepted opportunities and four wins to claim 23 independent outcomes: the wins sit inside the accepted group.
Record the date each stage was reached. Otherwise better follow-up this month can look like better quality from this month’s advertising, even when the team has simply cleared an older backlog.
Separate enquiry cohorts from closing activity
“What arrived in October?” and “What closed in October?” are different views. A sale closed in October may originate from a September enquiry. Show a cohort view based on enquiry arrival and an activity view based on the closing date. Do not divide revenue from the second group by enquiry volume from the first without explaining the mismatch.
A sophisticated forecasting model is not required at the start. A relevant follow-up window, an extraction cutoff and visible immature outcomes are. An Internal Operations Portal or Dashboard can use the same definitions and filters if an operational system is justified, but custom software is not a prerequisite for a credible report.
What attribution tells you
Attribution distributes credit across touchpoints under a particular model. Google Analytics documents alternative models and settings. State the model, window and reporting source, and avoid summing attributed totals from different advertising platforms as though they were separate sales. Google Analytics attribution.
An ordinary descriptive report does not, by itself, establish how many additional outcomes a channel caused. Some journeys cannot be connected or occur outside the available measurement. Keep “unknown source” visible rather than assigning it to the most convenient channel. Also separate enquiry fit from response speed and proposal quality; these require different corrective actions.
Acceptance criteria for a decision-ready report
- Every metric has a definition, source, period and extraction date.
- Totals can be reconciled with a permitted sample of underlying records.
- Exclusions, duplicates and unknown states are explained.
- Arrival cohorts are not confused with closing-month activity.
- Nested stages are not double-counted.
- Each recommendation identifies the observation, action owner and agreed review point.
Conversion Rate Optimisation is appropriate when evidence identifies a journey problem that can be tested. A falling percentage is not automatically a redesign brief. It may reflect a different audience mix or a backlog of unreviewed enquiries.
A useful monthly discussion should end with a small number of decisions: repair tracking, resolve unassigned records, narrow the offer, test a specific journey change or gather more mature outcomes. A collection of favorable percentages without these choices is a presentation rather than an operating tool.
Frequently asked questions
Is the advertising platform report enough?
It can answer questions about that platform’s activity. It does not automatically know what sales accepted or why a request was rejected. Those answers require separate definitions and records.
How do we assess the agency if sales does not update statuses?
Separate campaign and measurement responsibilities from business assessment. Show review coverage instead of replacing missing outcomes with undisclosed estimates.
Does an unknown source mean zero contribution?
No. It describes missing information, not proof that no channel contributed. Investigate the gap without inventing attribution.
When should we increase the budget?
When the evidence is sufficiently reliable for the decision, follow-up capacity exists and the findings justify a bounded expansion. There is no universal acceptance rate that makes every campaign scalable.
Request one sample report using test data and written definitions before commissioning recurring reporting. You will then know whether you are buying screenshots of numbers or a useful decision process.
Sources
- Google Ads: qualified and converted lead stages.
- Google Analytics: attribution and models.
- Sources reviewed 26 October 2026. All example quantities are hypothetical.
