Campaign Measurement: What to Set Up Before You Launch

Campaign measurement is decided before the campaign runs, not after. The metrics themselves are standard: impressions and reach for exposure, click-through and engagement for interest, conversion and cost per acquisition for outcome, and incremental revenue for contribution.

What determines whether you can report any of them is instrumentation: whether the campaign has a consistent name across every platform, whether its tracking parameters were applied correctly, and whether the creative variants are distinguishable. A campaign that ran well but was tagged inconsistently produces a report nobody can defend.

What is campaign measurement?

Campaign measurement is assessing what a single campaign achieved against what it was meant to achieve.

The second half of that sentence is the one teams skip. A campaign report that lists what happened is not a measurement; it becomes one when there is a stated intention to compare against. Without it, every number is simultaneously good and bad, and the conversation defaults to whoever is most confident in the room.

Campaign measurement is also narrower than it sounds. It covers one campaign, on its own terms. Questions about which channels deserve more budget, or whether the portfolio is working, are marketing performance measurement and need a different grain of data.

What to instrument before launch

Three things must be settled before launch: the campaign’s name, its tracking parameters, and how creative variants are distinguished.

None of the three can be added afterward, which is what makes them different from everything else on this page.

1. The campaign’s name, in every system. Not a name per platform. One value that appears identically in the ad platform, the email tool, the CRM and the analytics platform. This is the join key for every report that will ever be run on this campaign, and it is set once, by whoever sets up each platform.

2. The tracking parameters, built from agreed values. Source, medium and campaign at minimum, drawn from a permitted list rather than typed. Email, email and e-mail are three mediums in most analytics platforms, and the difference is invisible until reporting.

3. Creative variant distinction. If two creatives run under one identifier, no report can separate them, ever. The data to answer “which creative worked” either exists from the first impression or does not exist at all.

A useful test before launch: write down the three questions you will be asked about this campaign when it ends, then confirm the instrumentation can answer each. If one of them requires a field nobody is capturing, that is a five-minute fix now and an impossible one in six weeks.

Get the tagging right first: UTM parameters explained — what each parameter does and how to build them.

The metrics that answer each question

Exposure, interest, outcome and contribution each have their own metric set.

Experian’s guide to measuring campaign effectiveness (accessed 2026-09-13) works through a comparable set, including incremental lift as the measure that separates organic activity from what the campaign actually drove.

The rows are not interchangeable and the common failure is substitution upward. A campaign with a brand objective gets reported on conversions because conversions are available; a campaign with a pipeline objective gets reported on click-through rate because the pipeline number takes six weeks to land. Both reports are accurate and neither answers the question the campaign was funded to answer.

Pick the row before launch, from the objective. Report the others as context.

Engagement metrics

Engagement measures whether the audience did anything beyond seeing the campaign.

• Engagement rate = engagements ÷ impressions (or ÷ reach, if the platform reports it). Which denominator is in use matters more than the number, and platforms differ.
• Click-through rate = clicks ÷ impressions. The narrowest engagement signal and the most comparable across channels.
• Video completion rate = completed views ÷ starts. Note what each platform counts as a “view”: thresholds range from two seconds to full completion.
• Save, share and comment rates. The strongest interest signals on social, and the least standardized between platforms.

Engagement is the level where cross-platform comparison is most tempting and least valid. Each platform defines an engagement differently, so a higher engagement rate on one than another is not evidence of anything until both definitions are on the table. Compare a platform against itself over time; compare across platforms only on a metric both define identically, which in practice usually means click-through rate.

Paid media measurement

Paid media adds cost-side metrics (CPM, CPC, CPA) and a platform-reporting discrepancy problem.

The cost metrics are arithmetic and uncontroversial. DemandScience’s breakdown of campaign success measures (accessed 2026-09-13) maps them to funnel position: CPM for top-of-funnel reach, CPC for driving a specific action, CPA as campaign cost divided by conversions.

The discrepancy problem is the part worth planning for. The ad platform reports conversions it attributes under its own rules and windows; your analytics platform reports conversions it observed on site. These two numbers will not match, and the gap is structural: different attribution windows, different definitions of a conversion, and view-through conversions that never appear in analytics at all.

The right response is to pick one as the reporting source of truth before launch, state which one, and characterize the expected gap rather than trying to close it. A campaign report that switches source mid-flight to whichever number looks better is the fastest way to lose the room. Where several channels run together, cross-channel measurement compounds this, because each platform claims the conversions the others created the audience for.

How to report a result that survives scrutiny

A defensible campaign report states the metric, the source system, and the known gaps.

Four elements, and the third is the one that gets omitted:

1. The objective, as stated before launch. Quoted, not paraphrased after the fact.
2. The metric and its source system. “Conversions: 1,240 (GA4, last-click, 30-day)” rather than “1,240 conversions.”
3. The known gaps. View-through conversions not counted, one channel’s tagging incomplete for the first four days, an agency-trafficked placement missing its campaign value.
4. What you would do differently. Which is only credible if element three exists.

Stating the gaps is counterintuitive and it is what makes the report hold. A number presented without caveats invites the audience to find the caveat, and they will, usually mid-meeting and usually the one you knew about. A number presented with its limits already named moves the conversation to the decision.

“The campaigns that report cleanly aren’t the ones with better dashboards. They’re the ones where the name was decided before anything launched.” — Ethan Lowe, Senior Sales Engineer, Claravine

Why campaign reports get challenged

Reports are challenged when platform numbers disagree, which usually traces to inconsistent campaign values.

The disagreement people notice is between the ad platform and analytics, and that one is expected and explainable. The disagreement that does real damage is between two systems that should agree — analytics and the CRM, or two analytics views of the same campaign — because there is no innocent explanation available in the meeting, and the measurement function absorbs the doubt.

Underneath, it is almost always the same thing: the campaign is recorded under different values in different systems, so the two reports are grouping different sets of records. Each is internally correct. Neither is comparable.

One analytics team described the workaround this generates before it is fixed at the source.

“To download our campaign IDs (CIDs) for display advertising, we had a whole deduping process that we had set up with queries. We’d send it to our ad agency. They would append information, send it back to us, and we’d upload it to Adobe.” — Kimberly Whitehead, marketing technology manager, Vanguard

That is a recurring, multi-party process existing solely to reconcile identifiers that could have agreed at creation. It is also completely invisible in any campaign report, which is why it rarely gets funded as a problem.

Frequently asked questions

Sources

• Experian, “Measuring Marketing Campaign Effectiveness” (accessed 2026-09-13) — the metric set, including incremental lift.
• DemandScience, “Campaign Measurement: Ways to Measure Campaign Success” (accessed 2026-09-13) — CPM/CPC/CPA definitions and funnel mapping.