Campaign measurement

Campaign readout QA checklist

Published May 30, 2026. Updated June 3, 2026. Status: evergreen source page.

A campaign readout is not just a report. It is a request to believe that the campaign delivered a certain kind of value and that the next budget decision should follow from it.

Use this checklist when a campaign report arrives, before a renewal meeting, or before an attributed result becomes a board slide. It is written for marketers, publishers, agencies, analysts, and operators who need to separate clean delivery evidence from persuasive but unsupported performance language.

Start with the claim

Read the executive summary as a claim, not as a deck. The first job is to identify what the report wants the reader to believe.

Readout move	QA question	Careful conclusion
High delivery volume	Was delivery in the planned context, format, geography, device mix, and window?	The campaign delivered volume only if the eligible inventory definition was met.
Strong viewability or attention	Were measurable impressions representative, and were exposure-quality metrics matched to the campaign objective?	The campaign met an exposure-quality threshold, not a business-impact threshold.
High click or visit response	Were sessions qualified, deduped, and compared with a relevant baseline?	The campaign produced observed response from measured users.
More leads or conversions	Were outcomes defined before launch, filtered for quality, and tied to the correct source?	The campaign produced reported outcomes under the stated attribution rule.
Positive brand movement	Were exposed and control respondents balanced, and are question wording and field dates visible?	The survey measured a movement for this sample and window.
Incremental lift	Was the comparison protected before results were known, and are leakage and uncertainty shown?	The campaign supports a lift claim only within the design's limits.

Minimum report packet

A readout is easier to trust when the source materials are visible. Ask for the packet before debating the conclusion.

Campaign brief

The original objective, eligible context, placement IDs, dates, geography, device scope, creative set, destination pages, and primary signal.

Delivery log

Impressions, measurable impressions, viewable impressions, clicks, spend, pacing, frequency, and exclusions by placement, date, device, and creative.

Outcome definition

The exact event counted as a visit, engaged session, lead, qualified lead, conversion, sale, brand movement, or pipeline step.

Comparison rule

The prior period, matched context, holdout, geo split, exposed-control design, model baseline, or explicit statement that no comparison exists.

Quality controls

Invalid-traffic review, deduplication, bot filtering, conversion-window rules, lead-quality checks, audience leakage checks, and any missing-data notes.

Decision limit

The strongest action the report is allowed to support: delivery acceptance, creative iteration, package renewal, budget shift, or a deeper incrementality test.

Decision QA checklist

Question	Good evidence	Do not accept
Did the campaign run as planned?	Delivery by planned placement, context, date, device, geography, size, and creative.	Total impressions without showing where and how they ran.
What universe is being measured?	Clear denominator for eligible impressions, users, sessions, leads, respondents, or markets.	Percent changes with no starting level or eligible population.
Did the primary metric match the objective?	A preselected signal tied to the campaign's job.	Whichever metric looked strongest after the flight.
Was the comparison credible?	A holdout, matched baseline, prior period, or model baseline chosen before the readout.	A before-and-after chart presented as lift.
Were outcomes clean enough?	Deduped events, quality filters, conversion windows, and buyer-side status where available.	Raw attributed outcomes treated as incremental outcomes.
Are slices stable enough?	Minimum delivery thresholds and uncertainty for placement, creative, device, audience, or time slices.	Ranking small cells as winners and losers.
What else changed?	Concurrent campaigns, seasonality, pricing, promotions, sales coverage, inventory shifts, and site changes.	A report that credits the flight for every movement during the window.
What action follows?	A recommendation tied to the evidence level and the stated decision.	A broad budget claim from a narrow diagnostic signal.

Causal language ladder

The same result can support different language depending on the comparison. Use the strongest wording the design deserves, not the strongest wording the chart can imply.

Evidence available	Stronger wording	Weaker wording
No baseline or holdout.	The campaign delivered measured activity and observed response.	The campaign drove incremental growth.
Prior-period comparison.	Observed outcomes were higher than the selected prior period, with known context changes.	The campaign caused the entire increase.
Matched context or matched market.	The result is directionally stronger than a defined comparison, subject to matching limits.	The result is equivalent to a randomized test.
Designed holdout or geo test.	The campaign produced measured lift for this design, audience, window, and uncertainty range.	The campaign will produce the same lift in every future context.
Calibrated model plus experiments.	Multiple evidence streams support the budget direction, while assumptions still matter.	The model has proven exact channel contribution.

Slice audit

Breakouts are useful when they explain a decision. They are dangerous when a small, noisy cell becomes the story.

Placement

Compare eligible placements only after minimum delivery, viewability, creative mix, and device mix are visible.

Creative

Check whether one message received better inventory, a different audience, or a stronger landing page before calling it the winner.

Destination

Separate landing-page quality from media quality. A weak destination can depress qualified response even when delivery was clean.

Audience

Ask whether high-performing users were already higher-intent, easier to match, or more reachable before the campaign.

Time

Inspect weekday, seasonality, pacing, promotions, and news-cycle effects before assigning causality to a flight window.

Device

Device splits often mix screen size, context, measurability, landing-page behavior, and conversion friction.

Meeting script

What decision was this readout supposed to inform before the campaign started?
Which metric was primary, and was it selected before results were visible?
What comparison does the report use, and what better comparison was unavailable?
Which result would survive if attributed conversions were removed from the summary?
Which slice is useful enough to act on, and which slice is too small or confounded?
What should be renewed, changed, tested, or stopped based on the actual evidence level?

Pair with

Use this checklist with the campaign reporting worksheet when the readout needs a compact decision note, the campaign reporting sample when the readout needs a shared table structure, the campaign reporting terms glossary when delivery, traffic, lead, matchback, attribution, or lift language needs boundaries, the campaign KPI dictionary to keep metrics in the right lane, the creative and destination troubleshooting matrix when weak response needs a fix, narrow, hold, or retest decision, the campaign issue log and renewal register when weak lanes need owners, due dates, retest rules, and claim boundaries, the campaign renewal memo template when those rows need final renew, revise, retest, or hold language, the campaign renewal evidence archive when final limits need to carry forward, the campaign status-window closeout checklist before delayed lead, matchback, attribution, or survey rows are treated as final, the campaign status-window closeout register when final and preliminary rows need durable export fields, the outcome quality scorecard before reported visits, leads, pipeline, or matched outcomes become claims, the campaign data-layer spec, campaign tagging QA checklist, and creative asset acceptance checklist to preserve the source trail before launch, the private marketplace campaign measurement checklist to plan the readout before launch, the private marketplace readout export sample when the report needs row-level package, placement, device, creative, traffic-quality, and outcome-status fields, the private marketplace renewal scorecard before turning those fields into renewal, creative-change, budget-shift, or lift-test decisions, the incrementality test plan template when the decision needs causal evidence, and the measurement method selector when the report is using the wrong tool for the question.