Unclear marketing performance
What do your reports actually let you decide?
The monthly report shows visits, clicks and enquiries. The question you have is different: whether to keep paying for something, why the good enquiries dried up, or where to look next. The report doesn’t say. The same fog settles on a business with three dashboards and one with none. This page is about working out what you can know, what you’re missing and what the evidence will bear. Not perfect attribution. A clearer basis for the next decision.
What decision does the reporting need to support?
Start with the decision, not the data. Three come up more than any other.
Should we keep doing this?
Why has the number of suitable enquiries changed?
Where should we look next?
Each needs different evidence. Whether to keep paying needs what was delivered, what it cost, what enquiries followed and whether they fitted. A drop in suitable enquiries needs like periods compared, source by source, with the meaning of “suitable” held still. Where to look next needs less than people expect: one finding with a question attached.
A campaign report tells you spend and activity. It doesn’t tell you whether the enquiries that followed were any good. A sales report tells you revenue. It doesn’t tell you how each customer first found you. Neither report is wrong. Each answers a question you may not have asked. One dashboard won’t answer all three. The smallest useful set of measures for the decision in front of you will.
Activity, enquiries and sales tell you different things.
Think of each record as a witness. It saw one part of what happened, and none of the rest.
The work log and the channel spend
Saw what was delivered and where the money went. They never saw whether any of it caused a sale.
The visibility and traffic figures
Saw that people encountered the content and used the site. They never saw whether those people were the customers you wanted.
The validated enquiry count
Saw that a contact action was completed and actually arrived. It never saw whether the enquiry fitted, or what happened next.
The qualification and opportunity records
Saw what your team judged and how each enquiry developed. They never saw every marketing touch that led there.
The revenue records
Saw what became a recorded sale. They never saw how much of it one channel caused, as opposed to accompanied.
You don’t need every witness to have seen everything, and nobody has to track every person across every touchpoint. You need to know which record to ask for which question, and to leave the gaps in view.
Where did the enquiry go missing?
Take one good enquiry. There are three ways it goes missing.
01
It was never recorded properly.
The form fired twice, the phone call was counted as a page view, or the booking never registered at all.
A collection problem.
The work is measurement setup, then a check that the action recorded is the action that happened.
02
It was recorded, then lost in transit.
The enquiry exists in the website’s records. Its source, whether it was any good and what became of it live somewhere else, or nowhere.
A continuity problem.
The work is agreed record fields, a connection between systems, or a human habit of updating the status.
03
It was there all along.
The data exists. Nobody agreed what “lead” means, the comparison period keeps changing, and any rise gets credited to whatever ran last week.
An interpretation problem.
The work is a clearer question and stable definitions, not a new tool.
Whichever it is, somebody has to own the definitions, the source records and the later updates. Automation can carry a status from one system to another. It can’t reconstruct a judgement nobody wrote down.
Where did the enquiry go missing?
Take one good enquiry. There are three ways it goes missing.
01
It was never recorded properly.
The form fired twice, the phone call was counted as a page view, or the booking never registered at all.
A collection problem.
The work is measurement setup, then a check that the action recorded is the action that happened.
02
It was recorded, then lost in transit.
The enquiry exists in the website’s records. Its source, whether it was any good and what became of it live somewhere else, or nowhere.
A continuity problem.
The work is agreed record fields, a connection between systems, or a human habit of updating the status.
03
It was there all along.
The data exists. Nobody agreed what “lead” means, the comparison period keeps changing, and any rise gets credited to whatever ran last week.
An interpretation problem.
The work is a clearer question and stable definitions, not a new tool.
Whichever it is, somebody has to own the definitions, the source records and the later updates. Automation can carry a status from one system to another. It can’t reconstruct a judgement nobody wrote down.
Enough evidence to choose the next investigation.
Made up to show the format. Not a client, and not a result.
Observed
Campaign B produces enquiries at roughly half the cost of Campaign A. Neither campaign’s enquiries carry a consistent record of whether they fitted or what happened to them.
Uncertain
Whether B’s enquiries are as good as A’s. Whether either campaign has produced a sale. The last six weeks of B’s enquiries haven’t had time to become anything yet.
Next question
Ask the responding team to mark the last three months of enquiries from both campaigns as suitable or not, and note any that became work. Compare like periods. Then decide where the budget goes.
Nothing in that extract needs new software. It needs an agreed meaning of “suitable”, ten minutes from the person who answers the phone, and the patience to wait one decision cycle before calling anything a failure.
Reporting sized to the decision.
A proportionate arrangement has four parts: the question it exists to answer, the measures agreed to answer it, the sources those measures come from, and the person responsible for the review. Where a gap needs implementation, that’s scoped as its own piece of work.
What you end up with depends on the gap. It might be one measurement definition everyone uses. A validated enquiry action, so a count means what it says. A connection between the website’s record and the CRM’s. Or a short review that names what changed, what’s still uncertain and the next decision. These are examples shaped to the need, not a bundle.
Two things we won’t promise.
Full revenue attribution, because the evidence rarely supports it.
And a single source of truth, because your customers leave traces in systems you don’t control.
What we will do is separate a channel’s observed contribution from proof that it caused an incremental result. They’re different claims, and a report that blurs them isn’t helping you decide.
If the real problem is what happens after the enquiry arrives, that’s a different page. If the constraint is broader and still unclear, a Clarity Audit can establish it. A known measurement fault can simply be scoped and fixed.
What can’t your current reports tell you?
Bring the decision you’re struggling to make and the reports you use to make it. The first conversation is free. A broader review, or fixing a gap, is scoped separately.