Measurement needs a counterfactual

What would have happened anyway?

Measurement Press explains the gap between confident media claims and what the evidence can actually support. We publish practical case studies, field guides, and checklists for people who need to make better decisions from imperfect data.

Read the case studies Find a guide Evaluate a claim

Start with the evidence lane

Four recurring problems explain a surprising amount of weak advertising measurement and distorted media analysis. The map below points readers to the right public guide or campaign evidence workflow before a claim hardens.

Selection

The group that clicked, claimed, searched, or redeemed was already different before the campaign.

Attribution

The last measurable touch gets credit for demand that was already moving toward conversion.

Survey Lift

Brand studies often measure who answered and remembered, not what the campaign truly changed.

Framing

Headlines, comparison classes, and missing base rates can shape belief before facts are evaluated.

Choose the right starting point

Use this map to separate evidence pages written for public judgment from campaign and inventory pages written for media buying, reporting, and package QA.

Reader job	Best lane	Start with	Question it answers
Audit a media frame	Reader-facing evidence	Claim confidence rubric, media claim audit worksheet, worked media claim example, source triangulation checklist, anecdote checklist, case-study generalizability checklist, timeline checklist, source library	Does the wording match the source trail, example selection, chronology, denominator, comparison class, and evidence level?
Choose a measurement method	Reader-facing evidence	Measurement method selector, budget decision evidence ladder, budget decision worksheet, next-method explainer, campaign baseline comparison checklist, uncertainty interval checklist	Should the decision use a test, model, survey, attention metric, attribution report, cleaner baseline, or bounded budget action?
Study a misleading campaign result	Reader-facing evidence	Case studies, case-study generalizability checklist, audience selection checklist	What comparison group would show whether the campaign changed behavior, and can the lesson travel?
Compare a contextual media proposal	Campaign and inventory evidence	RFP evidence checklist, PMP readiness index, package proof sheet, package brief template, deal review checklist	What reader intent, inventory scope, creative fit, reporting fields, and renewal evidence should be visible before launch?
Check ad inventory before launch	Campaign and inventory evidence	Programmatic inventory QA checklist, inventory readiness matrix	Are slots, labels, sizes, reporting keys, and package boundaries ready for clean delivery?
Read a campaign or renewal report	Campaign and inventory evidence	Campaign readout QA checklist, campaign evidence triage tree, campaign baseline comparison checklist, creative and destination troubleshooting matrix, campaign issue register, renewal memo template, renewal evidence archive, renewal follow-up tracker, status-window closeout checklist, status-window closeout register, PMP campaign walkthrough, renewal scorecard	Does the report separate delivery, traffic quality, outcome status, baseline fit, creative and page friction, triage branch, open issues, final action language, archive needs, follow-up gates, status-window maturity, closeout register rows, and causal boundaries?

Calibrate confidence before choosing a claim

When a headline, dashboard, report, or campaign readout sounds conclusive, start by matching the sentence to the evidence level. This keeps strong language for strong evidence and turns weaker signals into better questions.

What the reader has	Use first	Decision it improves
A strong-sounding claim with unclear support.	Claim confidence rubric	Decide whether the evidence is decision-grade, strong but bounded, directional, weak, or unsupported.
A sentence says something caused, drove, reduced, or changed an outcome.	Causal claim review protocol	Check the counterfactual, timing, alternate explanations, and the strongest supportable verb.
A sentence that needs safer wording.	Evidence-to-claim language matrix	Rewrite the verb and scope so the claim matches the source trail, denominator, comparison, and causal design.
A meeting or vendor review that needs a record.	Source and vendor evaluation worksheet	Capture the claim, source, denominator, comparison, caveats, confidence level, and next evidence request.

Advertising Measurement Routes

Campaign evidence gets easier to read when planning, launch, reporting, and renewal are kept in separate lanes. Start with the route that matches the decision already on the table.

Before method choice

Name the decision

Use when a team is deciding whether the question needs MMM, a lift test, a geo test, a brand study, attention diagnostics, or attribution reporting.

Before launch

Protect the baseline

Use when the campaign needs package proof, inventory contracts, source fields, creative approval, and landing-page readiness before traffic starts.

At readout

Separate signals

Use when delivery, viewability, qualified visits, leads, matchbacks, and attribution are being compressed into one performance story.

After readout

Choose the next test

Use when the next action is renewal, creative revision, budget shift, a cleaner holdout, or a designed incrementality test.

Featured Case Studies

Each case study walks through the business incentive, the tempting read, the statistical break, and a better test design.

Brand lift measurement

The brand lift study that mistook survey recruitment for persuasion

How exposed and control respondent pools can make ad recall or consideration look stronger than a respondent-balanced read supports.

Read the case

Retargeting measurement

The retargeting campaign that counted cart returners as new demand

Why exposed visitors who recently browsed, saved, compared, or abandoned a cart can make attributed ROAS look stronger than true lift.

Read the case

Lead generation

The lead-gen campaign that counted form fillers as new demand

Why attributed form fills and CRM matchbacks can confuse pre-existing buying intent with incremental pipeline.

Read the case

Geo lift testing

The geo lift test that mistook seasonal markets for media lift

How uneven pre-period demand and market selection can turn a clean-looking post-launch gap into overstated incrementality.

Read the case

Cash-back retail media

The cash-back claim test that rewarded pre-existing purchase intent

Why comparing buyers who claimed an offer against everyone else can make engagement look like incremental sales.

Read the case

Search advertising

The search campaign that confused intent capture with incrementality

How paid search can look extremely profitable when brand demand and commercial intent are already present.

Read the case

Card-linked offers

The card-linked offer that targeted people already ready to spend

Transaction-history targeting can make a promotion look stronger than its true causal effect.

Read the case

Practical Guides

Short references for the recurring questions analysts and editors should ask before trusting a claim.

Best first pass: use the claim confidence rubric to decide how strong a claim deserves to sound, the worked media claim example to see one claim move from source trail to supportable sentence, the source triangulation checklist when several citations may trace to the same origin, the anecdote and exemplar checklist when one vivid example may be standing in for a broader pattern, the causal claim review protocol when a sentence says something caused or drove an outcome, then use the evidence-to-claim language matrix when the final sentence needs cleaner wording.

Source and vendor evaluation desk

Choose the right checklist for news claims, research reports, dashboards, vendor pitches, and budget evidence.

Source and vendor evaluation worksheet

Print a structured review sheet for scoring evidence trails, denominators, comparisons, measurement design, caveats, and next actions.

Advertising measurement glossary

Align terms like incrementality, attribution, MMM, lift, brand study, attention, and confidence before a readout becomes a decision.

Advertising measurement method selector

Choose between MMM, lift tests, geo tests, brand studies, attention metrics, and attribution reports based on the decision.

Advertising budget decision evidence ladder

Match repair, renewal, budget shift, scale, cut, or test actions to the strongest evidence rung the report can support, then use the meeting worksheet to record rows, exclusions, thresholds, and next uncertainty, or the next-method explainer when stronger evidence is needed.

Brand lift study readout checklist

Review survey balance, respondent quality, outcome meaning, uncertainty, and decision limits before trusting a brand lift report.

Attention measurement decision guide

Use attention metrics as exposure-quality diagnostics without treating attention as proof of incremental value.

The MMM causal validity checklist

Controls, priors, calibration, uncertainty, and why predictive fit is not the same as causal proof.

Media claim audit worksheet

Check the evidence trail, denominator, comparison class, source mix, causal language, and disconfirming context before trusting a frame; use the worked example to see the pattern applied.

Headline and source-mix framing checklist

Check headline scope, causal upgrades, source roles, quote weighting, and missing counter-context before trusting a story.

Quote weight and response standard checklist

Inspect quote roles, rebuttal placement, response parity, and verb choice before accepting a balanced-looking story.

Incrementality test plan template

Define the decision, counterfactual, assignment, outcomes, power, leakage checks, and readout rules before results are visible.

Geo lift test design checklist

Plan market-level experiments with cleaner baselines, pre-period trend checks, outcome rules, and sensitivity tests.

Retail media incrementality checklist

Separate closed-loop attribution, iROAS, clean-room matches, and buyer-would-have-bought-anyway selection from true lift.

Where lift tests and brand studies quietly lose truth

Holdout leakage, survey recruitment bias, short windows, proxy outcomes, and platform measurement incentives.

How subtle media bias enters before explicit opinion appears

Base-rate omissions, verb loading, source imbalance, comparison drift, and context laundering.

A source quality scorecard for media claims

Grade proximity, completeness, independence, uncertainty, and replicability before trusting a frame.

Private marketplace campaign measurement checklist

Plan and read private marketplace, sponsorship, and contextual display results without mistaking delivery signals for causal proof.

Campaign KPI dictionary for cleaner readouts

Separate delivery, attention, traffic quality, lead quality, brand movement, and incrementality before reporting language hardens.

Campaign data-layer spec for cleaner readouts

Preserve placement, creative, destination, audience, test-cell, and outcome fields before reporting becomes a claim.

CTV and video campaign measurement checklist

Read video reports without confusing completion, attention, attribution, or brand lift with incrementality.

Identity matchback measurement checklist

Check match rates, identity joins, offline outcomes, clean-room overlap, and comparison rules before trusting matched conversion claims.

Topic taxonomy and reader intent map

Find the right desk, guide, case study, or sponsor-fit context for the judgment a reader needs to make.

Source library for media and measurement claims

Find official references for citation trails, source quality, measurement methods, attention metrics, MMM, and programmatic inventory.

Survey and poll claim checklist

Check sample source, wording, weighting, field dates, uncertainty, and comparisons before trusting a poll or survey headline.

Public records and denominator checklist

Check official records, incident counts, percent changes, missing populations, and rate bases before trusting a public-data frame.

Campaign readout QA checklist

Audit delivery, comparisons, outcome quality, slices, and causal language before a report becomes a budget decision.

Viewability and invalid traffic measurement checklist

Check measurability, viewability, invalid traffic, frequency, and exposure-quality denominators before trusting campaign outcomes.

Landing page and lead quality measurement checklist

Separate campaign traffic, page conversion, form quality, sales follow-up, and incrementality before a lead-gen report shapes budget.

Reach and frequency measurement checklist

Check deduplicated reach, identity units, frequency caps, high-frequency tails, and incremental reach before trusting delivery claims.

Before-and-after chart and trend claim checklist

Check chart windows, denominators, seasonality, concurrent changes, and comparison rules before treating a trend break as proof.

Creative testing readout checklist

Check message, offer, placement, audience, landing page, and outcome quality before calling a creative winner.

Audience selection bias checklist

Check whether retargeting, lookalike, CRM, retail media, or high-intent audience results were already different before exposure.

Sponsored research and vendor report checklist

Inspect sponsor role, method, sample, denominator, comparison, and limitations before trusting a research asset.

Campaign reporting sample for contextual media buys

Use a sample report structure for delivery, context, creative, traffic, lead quality, and evidence-level language.

Campaign reporting worksheet for contextual media buys

Fill a compact readout worksheet for delivery quality, outcome status, comparison rules, missing fields, and next-decision language.

Campaign reporting terms glossary

Define delivery, exposure quality, traffic, lead quality, matchback, attribution, incrementality, and renewal terms before a report shapes budget.

Denominator framing examples for media claims

Translate counts, percentages, rates, survey shares, public records, and reported incidents into fairer denominators before trusting a frame.

Composition mix and average claim checklist

Check whether an average, rate, share, survey result, or campaign metric moved because the underlying mix changed.

Media buying RFP evidence checklist

Ask for context proof, inventory scope, reporting fields, comparison rules, and bounded claim language before proposals are scored.

Contextual package proof sheet

Use a one-page package handoff for reader job, inventory, creative fit, reporting grain, and renewal decisions.

Private marketplace readout export sample

Check package, placement, device, creative, traffic-quality, outcome-status, and conclusion fields before renewal language is written.

Private marketplace renewal scorecard

Turn readout fields into renewal, creative-change, budget-shift, or lift-test decisions.

Campaign issue log and renewal register

Track weak evidence lanes, owners, retest rules, and claim boundaries before renewal language is written.

Campaign renewal memo template

Convert scorecard results and issue closeout rows into final renew, revise, retest, or hold wording.

Campaign renewal evidence archive template

Preserve memo decisions, excluded rows, issue IDs, next-test requirements, and follow-up dates across flights.

Campaign renewal follow-up tracker

Carry archived issue IDs, owners, due dates, next-test requirements, and launch gates into the next flight.

Campaign status-window closeout checklist

Decide when delayed lead status, matchback, attribution, survey, and follow-up fields are mature enough for final language.

Campaign status-window closeout register template

Preserve closeout IDs, maturity thresholds, excluded rows, allowed claims, and next review dates across flights.

Evidence-to-claim language matrix

Rewrite media, research, dashboard, and campaign claims so the wording matches the evidence.

Timeline and event-order framing checklist

Separate sequence, lag windows, prior state, and concurrent changes before treating chronology as cause.