Measuring campaign reach honestly means separating three things that reports routinely blend: how many people had the opportunity to see something, how many actually attended to it, and how many did anything differently afterwards. Most reporting collapses all three into one number, and that number is almost always the least defensible of the three.
What campaign reach figures actually mean
"Reach" in platform analytics means an opportunity to see. It does not mean seen, read, understood or believed. A post that appeared in 40,000 feeds and was scrolled past in 40,000 feeds reports identically to one that was read.
Radio reach is an estimate derived from coverage area and survey data, usually old survey data. Print circulation counts copies distributed, not copies opened. Event attendance counts bodies in a room, some of whom are colleagues.
None of this makes the numbers useless. It makes them a measure of distribution, which is worth knowing and should be labelled as such.
The three-tier structure worth adopting
Report in tiers, and never let a number from one tier be presented as evidence of another.
- Distribution — what you sent out and where it could have been seen. Fully within your control, easy to verify, weak evidence of anything.
- Engagement — what people did that required effort. Time on page rather than page views. Calls to a phone-in. Questions asked at a session. Materials requested. Each of these costs the audience something, which is what makes them meaningful.
- Change — what is different now. Service uptake, policy movement, reported behaviour, recall in a follow-up survey. Slow, expensive, and the only tier that answers the question anyone actually asked.
BetterEvaluation is the most practical open resource for designing the third tier without a research budget.
Attribution: say what you cannot know
The hardest question in campaign measurement is whether the change would have happened anyway. Rigorous answers require comparison groups that most organisations cannot construct.
The honest response is not to fake precision but to state the limit. "Service uptake rose 18% in the broadcast area over the campaign period. Uptake in the comparison district rose 6%. We cannot rule out other factors, including the clinic staffing change in March." That paragraph is more credible than any confident attribution claim, and funders who know the field read it as competence rather than weakness.
Metrics that quietly mislead
- Cumulative reach across channels. The same person counted four times is not four people. If you must aggregate, say that overlap is unremoved.
- Impressions. Counts renders, not readers. It is the number most easily inflated by spend.
- Follower growth. Measures account performance, not campaign effect.
- Media value equivalence. Converting coverage into a notional advertising cost invents money that was never spent or earned. Avoid entirely.
- Attendance at your own events. Measures convening ability. Rarely measures persuasion.
Design measurement before the campaign, not after
Retrospective measurement can only use whatever data happened to be collected, which is why so many reports fall back on distribution figures. Deciding beforehand what would count as evidence takes an afternoon and changes what is possible.
Three questions to settle in advance: what specifically should be different, for whom, and by when? What would we observe if it worked, and what would we observe if it did not? What is the baseline, and have we actually recorded it?
That last one is where most measurement plans die. A baseline is only obtainable before you start.
Reporting honestly is a strategic choice
There is short-term cost in reporting a modest, defensible figure while others report large, soft ones. The compensation is that your numbers survive scrutiny, and over several cycles that becomes the reputation that wins work.
It also improves the work itself. A campaign measured by distribution optimises for distribution. A campaign measured by change is forced to think about what would actually produce it — which is the harder and more useful question, and the one running through why humanitarian storytelling keeps failing its subjects.
For channels where reach data is inherently soft, see the field guide to community radio partnerships, and for measuring response under pressure, building a crisis communications plan that survives contact.
Common questions
What is the single most useful metric?
An action that costs the audience something — a call, a visit, a request, a registration. Effort is the cheapest available proxy for attention.
How do you measure a campaign with no digital component?
Short structured surveys at fixed points, service records from partner organisations, and comparison against a similar area not covered. Slower, and often more reliable than platform analytics.
What should a report include when results are weak?
The result, the likely explanation, and what you would change. Funders fund learning; they defund the appearance of certainty that later collapses.
If you want a measurement framework built before your next campaign rather than after it, let us talk.