Campaign measurement

Outcome quality scorecard for campaign readouts

Published May 31, 2026. Status: evergreen source page.

Outcome volume is easy to count and easy to overread. A serious campaign readout asks whether the reported outcomes were eligible, deduped, qualified, timely, and tied to a comparison that can support the conclusion.

Use this scorecard before a campaign launches, when a report arrives, or when a buyer and publisher need shared language for what a visit, lead, sale, matchback, or lift result can actually prove.

Outcome ladder

Read the strongest outcome by the weakest link in its source trail. A late-stage business metric can still be weak evidence if identity, qualification, or comparison rules are unclear.

Outcome level	Useful when it shows	Quality checks	Do not claim alone
Qualified visit	A real session from the intended context reached a relevant destination.	Invalid traffic review, bounce rule, content depth, duplicate visits, source fields, and device mix.	Incremental demand or lead quality.
Engaged account or household	Multiple signals connect to the same qualified organization or buying group.	Identity unit, match rate, account fit, repeat exposure, excluded internal traffic, and minimum activity threshold.	Sales intent without buyer-side status.
Raw lead	A form, call, signup, download, or demo request was captured.	Required fields, duplicate suppression, spam filter, consent status, time to follow-up, and campaign source.	Pipeline, revenue, or incremental impact.
Qualified lead	The lead matches a prewritten fit, need, role, geography, or readiness definition.	Disqualification reasons, missing fields, sales acceptance, follow-up coverage, and lead-source overwrite rules.	That media caused the lead.
Opportunity or pipeline	A sales process accepted the outcome and assigned a value or stage.	Stage definition, owner assignment, duplicate account rules, source influence rule, and time lag from lead to stage.	Closed revenue or lift without maturity checks.
Matched sale or conversion	A buyer-side outcome was joined back to campaign exposure or traffic.	Match method, unmatched universe, lookback window, deduplication, prior intent, and comparison rule.	Incrementality without a protected baseline.
Incremental outcome	A designed comparison estimates what changed because of the campaign.	Holdout integrity, leakage, outcome window, uncertainty, pre-period balance, and sensitivity checks.	That the same effect holds in every future context.

Score the readout

Give each line a low, medium, or high quality read before judging the headline result. A report with high volume and low outcome quality should lead to cleanup, not a stronger claim.

Scorecard dimension	High quality	Medium quality	Low quality
Outcome definition	The event, stage, or sale is defined before launch and visible in the report.	The definition exists, but it changed or needs interpretation.	The report counts a vague conversion, action, or success event.
Eligibility	The outcome belongs to the planned audience, geography, context, date range, and destination.	Most outcomes appear eligible, with some exclusions or missing fields.	The report pools unlike audiences, placements, windows, or destinations.
Deduplication	User, account, household, lead, and order duplicates are handled with named rules.	Duplicate rules exist for some but not all outcome types.	The same person, account, or order can be counted multiple ways.
Qualification	Quality filters and disqualification reasons are prewritten and reported.	Quality is reviewed manually after results arrive.	Raw volume is treated as useful demand.
Source trail	Campaign, placement, creative, destination, audience, and outcome fields survive into the readout.	Some fields survive, but source or creative detail is incomplete.	Outcomes are detached from the delivery and creative context that produced them.
Timing	Outcome lag, data maturity, and reporting cutoff are stated before comparison.	The report mentions lag but does not show mature cohorts.	Early outcomes are ranked as final performance.
Comparison	The readout uses a holdout, matched baseline, prior flight, or explicit no-comparison label.	The comparison is directionally useful but imperfect.	Attributed outcomes are written as lift.
Decision fit	The recommended action matches the evidence level.	The action is plausible but needs a follow-up test or cleanup.	The report asks for a budget decision the evidence cannot support.

Minimum outcome packet

A useful outcome packet lets a reviewer reproduce how the result became a conclusion. Ask for these pieces before debating whether the campaign worked.

Event dictionary

Defines every counted action, including page events, form submits, calls, downloads, demo requests, account actions, sales stages, purchases, refunds, and exclusions.

Qualification rubric

States what makes a visit, lead, account, opportunity, or sale useful enough to count for the campaign's decision.

Deduplication rule

Names the unit being counted once: person, browser, device, household, account, lead, opportunity, order, or transaction.

Source fields

Preserves campaign ID, package ID, placement ID, creative ID, destination, audience rule, device, date, and content context.

Lag and maturity note

Shows how long outcomes usually take to appear and which cohorts are mature enough for comparison.

Comparison rule

States whether the readout uses a holdout, matched context, prior flight, model baseline, or descriptive attribution only.

Lead and pipeline review

Lead generation reports need a quality view before they need a victory sentence. Volume is only useful when the follow-up process can separate real fit from easy form completion.

Review field	Ask for	Why it matters
Form completion	Required fields, hidden fields, validation rules, duplicate handling, and spam filters.	Prevents form ease or weak validation from masquerading as demand.
Fit status	Role, company size, geography, use case, category need, or other prewritten fit criteria.	Separates relevant demand from broad curiosity or low-fit submissions.
Follow-up coverage	Share contacted, time to first contact, owner assignment, and unreachable rate.	A campaign cannot be fairly judged if many leads were never worked.
Disqualification	Reasons such as duplicate, student, vendor, competitor, wrong region, no need, no budget, or bad contact data.	Shows whether the campaign is producing the wrong audience or the wrong page behavior.
Stage movement	Sales accepted, meeting held, opportunity opened, closed, lost, pending, and aged-out status.	Prevents early lead counts from being treated as mature pipeline evidence.
Source overwrite	Whether later channels, CRM imports, or sales updates can overwrite the original campaign source.	Keeps credit rules from reshaping the readout after results are visible.

Careful outcome language

Evidence available	Use this wording	Avoid this wording
Qualified visits only.	The campaign produced observed qualified traffic from the planned contexts.	The campaign generated qualified leads.
Raw leads with limited qualification.	The campaign produced form activity that needs duplicate, fit, and follow-up review.	The campaign produced pipeline.
Qualified leads with no comparison.	The campaign produced qualified leads under the stated filters and source rules.	The campaign created incremental demand.
Matched conversions with no holdout.	Matched records show observed outcomes inside the selected window.	The campaign caused these conversions.
Holdout or matched baseline with clear outcome quality.	The campaign supports a measured lift read for this population, window, and outcome definition.	The result proves the same lift in all future buys.

Meeting script

Which outcome is the report asking us to believe in: visit, lead, opportunity, sale, or lift?
Was that outcome definition written before launch?
What share of outcomes survived duplicate, eligibility, and quality checks?
Which outcomes are still immature because the lag window has not closed?
What comparison makes this result meaningful, and was it chosen before results were visible?
What would we change if we judged by qualified outcomes rather than total outcomes?

Pair with

Use this scorecard with the campaign readout QA checklist when a finished report needs review, the campaign KPI dictionary when metric language needs boundaries, the campaign data-layer spec before launch, the landing page and lead quality checklist for form and follow-up checks, the identity matchback checklist for joined outcomes, the attribution window and conversion lag checklist for credited outcomes, the campaign status-window closeout checklist when lead, matchback, sales follow-up, attribution, or survey rows need maturity review, and the incrementality test plan template when the decision needs causal evidence.