Content marketing agency metrics: trace every number

What to take away

In content marketing, the most common mistake is treating a platform figure as a business outcome. A post that reached 40,000 accounts tells you distribution happened. It does not tell you whether anyone changed their mind, enquired or bought. If that distinction is not in your reporting template, your client will eventually notice the gap.

Define the metric before the work

A metric needs three things written down: what it measures, how it is calculated, and who owns it. Without those, two people can report different numbers from the same campaign and both be confident.

Start from the decision the number will inform. If the decision is whether to renew a retainer, a useful metric is cost per qualified enquiry, not total sessions. If the decision is whether to change a headline format, a useful metric is click-through rate on a defined set of pages.

Write the metric specification into the brief before work starts: the decision it supports, calculation, source platform or report, date range, filters and owner. Ask each agency to agree that definition in writing and flag any proposed change before comparing results.

Trace every number to a source

The Office for National Statistics publishes its statistical methodology guidance for exactly this reason: definitions, sampling and revisions change what a figure means. You do not need to run a national survey, but you should copy the habit of documenting how a number was produced.

Use the original platform report or export, not a chart alone. In Google Analytics 4 (GA4), record the report, date range, filters and attribution setting; in Google Search Console, record clicks and impressions separately. If leads are passed into HubSpot, note the lifecycle-stage definition used to count a qualified lead; if Looker Studio presents results, retain the underlying source reports as evidence.

The Data & Marketing Association's industry research library can help locate sector research. Before using a benchmark, record the specific report, publication date, sample and methodology; treat the result as context, not as your own campaign result.

Separate outcome from attribution

HM Treasury's Magenta Book guidance on evaluation makes a point that applies well beyond public policy: you need a credible counterfactual before you claim your activity caused a change. In content marketing, that usually means a comparison period, a control group of pages, or a holdout segment.

If you cannot build one, say so in the report. A labelled estimate is more useful than a confident number that falls apart under questioning.

A worked example with labelled figures

A small B2B team spends £3,000 a month on content production and £1,200 a month on distribution, so £4,200 a month in total. Over a quarter, that is £12,600.

The campaign produces 180 marketing-qualified leads. At that volume, the illustrative cost per qualified lead is £70. Of those, 24 become sales-accepted opportunities, giving an illustrative cost per opportunity of £525.

Now the honest part. The team has no holdout, so it cannot prove the content caused all 180 leads. It can only report that leads rose from a prior-quarter baseline of 120. The report should present the £70 figure as an observed cost, and the causal claim as unproven. That distinction is what keeps a report credible.

A dashboard is a presentation layer: whether it is built in Looker Studio or another tool, label every chart with its source, date range, filters and limitations, and retain the underlying export. If reporting includes personal data, apply the UK GDPR and the Data Protection Act 2018, follow Information Commissioner's Office (ICO) guidance on data protection, restrict access and avoid putting identifiable lead data into a widely shared dashboard.

Report the limits as well as the results

Every report should carry a short limitations note. Name the sample size, the tracking gaps and any platform change that affected the data. Keep it to four or five lines.

This protects the relationship. A client who understands the limits of a number will trust the next one more, not less. It also stops the agency being asked to defend a figure it never claimed.

Common questions

Which metrics should a content agency report every month?

At minimum: qualified leads or enquiries, cost per qualified lead, and progress against one agreed content objective. Add channel-level detail only where it informs a decision.

How do I know if a benchmark applies to my client?

Check the sample, sector and date. A benchmark from a different market or an older period is context, not a target, and should be labelled as such in the report.

What if the client wants a single headline number?

Give one, but define it precisely and state what it excludes. A single number without a definition invites disputes later when the figure is challenged.

How often should the metric set change?

Review it at each planning cycle, not mid-quarter. Changing a metric after results arrive looks like moving the goalposts, even when the change is justified.