Define the outcome
Distinguish an interaction from a qualified enquiry, booking or sale.
Campaign measurement
Beyond the numbers.
Find the meaning.
A report should help you make a decision. Connect advertising activity with meaningful outcomes, and understand the context behind every result.
Discuss your brief
The perspective
Measure what matters.
Know what it means.
Impressions, clicks and enquiries describe different parts of a journey. The useful question is how each relates to the business objective, and what the available evidence can genuinely tell you.
Define the action before choosing the metric. Agree which records are reliable, how outcomes connect and where the gaps sit. A clear measurement brief creates a stronger foundation for review and the next decision.
Check the evidence
Understand definitions, duplicate events and the limits of attribution.
Inform the decision
Use the findings to identify what needs attention and what to explore next.
Start at the business endName the action
Name the action
that matters.
For an appointment business, for example, a completed booking may be more useful than a page visit. It still needs a clear definition: confirmed, attended or paid?
ACTIVITYMEANING
Always name the denominator
A rate only makes sense when you know what it divides by. Clicks per impression, enquiries per visit and sales per qualified enquiry answer different questions. Compare like definitions over comparable periods.
02 / Read the signal
Four levels.
Different questions.
Moving up the steps requires more context, not just more activity. There is no automatic path from a display to a sale.
From signal to meaning
Choose a signal to unpack it.
Read what it describes, what it cannot prove and which context makes it useful. Each signal tells part of the story; context makes it useful.
Reach & impressions — Was there an opportunity to see it?
What it describes
Impressions count recorded displays; reach describes an estimate of distinct people or accounts reached. The same person may account for several impressions.
What it cannot prove
Neither proves that someone noticed, understood or remembered the message. Reach can be estimated and is not automatically comparable across platforms.
Context to ask for
Ask how the platform defines a display and a reached person, which period is included, and whether frequency or placement changes affect interpretation.
Clicks — Did someone take the next step?
What it describes
A click records an interaction as defined by the platform. Click-through rate relates counted clicks to counted impressions, using a clearly specified click type.
What it cannot prove
A click is not necessarily a loaded page, a useful visit or an enquiry. Different platforms may count different interactions as clicks.
Context to ask for
Compare the chosen click definition with destination visits, check the page works and keep dates, placements and campaign scope consistent.
Enquiries — Did interest become a conversation?
What it describes
An enquiry is an agreed contact action, such as a completed enquiry form or a relevant call. A conversion rate needs a named denominator, such as eligible visits.
What it cannot prove
An event count alone cannot establish that an enquiry is genuine, unique, relevant or commercially useful. A form opening is different from a completed submission.
Context to ask for
Agree what qualifies, how repeat contacts are handled, and who checks quality. Establish the source of truth before interpreting the total.
Business outcomes — Did the action matter to the business?
What it describes
A qualified opportunity, completed booking or sale can connect advertising questions to commercial activity. Define the outcome and the period in which it is recognised.
What it cannot prove
A sale following an advert does not show that advertising alone caused it. Revenue is not profit; refunds, margin and other costs change the interpretation.
Context to ask for
Ask how enquiry records connect to outcomes, how long decisions take and which costs are included. Record where evidence is incomplete rather than assuming a return.
03 / Read the small print
The number depends
on the conditions.
Agree how information will be collected and interpreted. A neat report cannot repair unclear definitions or missing evidence.
Establish what can be collected.
Ask what information is needed, what permissions apply and who is responsible for collection and account access. Some activity may be unavailable or unobserved; do not treat an incomplete record as a complete view of everyone’s behaviour.
Check the action behind the event.
Define the exact action, where it is recorded and how it would be checked. A test submission, a broken destination or a changed form can alter a count. Record changes so that a rise or fall is not automatically interpreted as a change in demand.
One person can leave several traces.
A reload, repeated submission or more than one collection method can count the same action twice. Several contacts can also belong to the same opportunity. Agree whether the report counts actions, unique enquiries or customers and who resolves duplicates.
Credit is a convention, not certainty.
Someone may see an advert, search later and then contact the business directly. Different attribution rules and time windows can assign credit differently. Platform totals may overlap, so adding them together can overstate outcomes. Ask which rule is used and what it leaves out.
04 / Before a report is agreed
Bring the question.
Agree the meaning.
Explain the outcome you want to understand, the records currently available and any known gaps. Use that starting point to discuss the reporting scope and the decisions it needs to inform.
Discuss measurement- Definitions
Which actions and denominators would be used? What counts as a useful enquiry or completed outcome?
- Ownership & access
Who owns each account, supplies business records and checks data quality? What access would be necessary?
- Cadence & responsibility
Who prepares and interprets the report, how often, and who acts on unresolved questions?
- Limits & decisions
Which gaps, attribution rules and changes should accompany the numbers? What decision could the evidence reasonably support?