The most common reason marketing budgets are cut is not that they failed. It is that nobody could demonstrate what they achieved. Reporting built on reach and impressions describes activity, not outcomes, and a finance director is right to be unconvinced by it.
Set measurement up before spending
Attribution cannot be reconstructed retrospectively. If conversion tracking is not configured and verified before a campaign starts, that campaign’s contribution is permanently unknowable — you will be left comparing traffic charts and guessing. Verification matters as much as configuration: a tracking tag that is present but firing on the wrong event produces confident, wrong numbers.
Measure what the business cares about
- Cost per qualified enquiry, not cost per click.
- Which channel and campaign produced each enquiry.
- How many enquiries the sales team judged worth pursuing.
- The proportion that became customers, where your process allows that to be tracked.
Consent changes what you can collect
Analytics that respect a visitor’s cookie choice will record less than analytics that ignore it. That is the correct trade-off, and it means your reports should be read as a consistent directional signal rather than as a complete census. What matters is that the basis of measurement stays the same between periods so comparisons remain meaningful.
Read the numbers honestly
Small samples produce large swings. A channel that generated three enquiries last month and one this month has not necessarily declined by two thirds — it may have produced no signal at all. Give a campaign enough volume and enough time to say something before acting on it, and be equally willing to stop something that is genuinely not working.