Measurement
Part of How marketing agencies measure campaign performance
How a marketing agency should choose client reporting metrics
Which client reporting metrics are worth a marketing agency's time: how to pick four, baseline them honestly and report them without overclaiming.
What to take away
Key metrics for a marketing agency are the small set of numbers that show whether client work produced the outcome the client paid for.
- Choose measures that map to a client objective, such as qualified enquiries or revenue per campaign.
- Baseline each one before you change anything, so the after view has something honest to sit against.
- Write down the definitionswhat counts as a lead, when a sale is credited, which system is the source.
- Report a range and name the source, and say plainly what you cannot yet prove.
- Review the set each quarter, because a dashboard nobody acts on teaches clients to ignore reporting.
What makes a metric worth reporting
Reporting reach alone is easy and rarely settles the argument. When a client asks what the retainer bought, impressions do not answer.
Does the metric change a decision?
Does it change a decision?
keep it, report it
drop it, it is noise
A metric earns its place when it changes a decision. Qualified enquiries, cost per accepted lead, revenue attributed to a named channel and retention by cohort all pass that test.
Pick two to four. A longer list divides attention and usually hides the one number that would have prompted action.
What the four metrics mean in practice
- Qualified enquiriesenquiries that meet the client's agreed criteria for a sales-ready lead, counted once the client or sales team has accepted them. Put that wording in writing, or two people will count the same month differently.
- Cost per accepted leadthe total campaign spend over a period divided by the number of leads the client accepted in the same period. Both figures must cover the same dates.
- Revenue attributed to a named channelsales credited to one channel under one stated attribution model, so paid search and paid social cannot both claim the same sale.
- Retention by cohortthe share of clients who started in the same month, quarter or year and are still active after a set number of months.
Keep these definitions in the report itself, in the client's own language, so the numbers can be checked rather than argued about.
How to baseline before you change anything
A baseline is the value of a metric over a stated period, pulled from a named system on a known date. Without those three details it is an anecdote.
Baseline needs three details
- Value of the metric
- Stated period
- Named source system
- Known pull date
- Export last twelve months
- Note gaps and tracking changes
Export the last twelve months from the platforms and the client's own systems where you have access. Note gaps, tracking changes and campaign pauses in the same file.
Source systems are usually a small set: Google Analytics 4 for on-site behaviour, the advertising platforms such as Google Ads and Meta Ads for spend and clicks, and the client's CRM, such as HubSpot or Salesforce, for accepted leads and closed sales. Record which one supplied each figure, because the same number can differ between them.
Once a baseline exists, it feeds targets and review cycles, which helps when a client wants a number before the work starts.
Sector context helps you separate your performance from the market. The Office for National Statistics on GOV.UK publishes the economic and business indicators that agencies commonly use as that backdrop.
Before and after: a reporting change
Small structural changes do most of the work. The table below compares a common starting position with a tidier one.
Before
- Metric set
- Twelve activity measures
- Baseline
- Not recorded
- Attribution
- Last click
- Format
- Monthly slide deck of screenshots
- Review
- Once a year
After
- Metric set
- Four measures tied to client objectives
- Baseline
- Twelve months, with source system and pull date
- Attribution
- Channel value model, named in the report
- Format
- One page: result, range, next action
- Review
- Each quarter
Keep the monthly report to a page so the result, the range and the next action sit together.
Tying metrics to channel value
Attribution decides which metric gets the credit, and different models disagree. Google's help page on getting full value from web and app channels explains how conversion value is credited across channels. That matters when paid search and paid social both claim the same sale.
Analytical skill limits what you can attribute with confidence. The data and insights courses at CIM cover the measurement ground that agency reporting depends on. That is worth checking before you promise a model the team cannot maintain.
How to report the numbers
Every figure in a client report needs a source and a pull date. For example, a team paying £400 a month for a reporting tool should be able to say which platform each number came from.
Show a range where the data is thin. A labelled estimate is more useful than a false decimal point.
State what you cannot yet prove. Clients accept gaps; they do not accept being surprised later.
Dashboard layout is the next decision after the metrics: choose panels that match the agreed objectives, and set a refresh timing the team can actually keep.
Common questions
How many metrics should an agency report each month?
Two to four for a single client objective, with supporting detail available on request. More than that usually means the team has not agreed what the work is for.
Is a click a key metric?
No. A click measures interest, not commercial outcome. Treat it as diagnostic detail rather than a headline result.
How often should the metric set change?
Review it each quarter and change it when the client objective changes. Changing definitions mid-quarter makes the baseline useless.
What if tracking breaks during a campaign?
Record the outage in the report and mark the affected figures. Hiding a gap damages trust more than the gap itself.
