Measurement worksheet

Advertising budget decision meeting worksheet

Organization byline: Measurement Press editorial desk. This page does not list an individual author or reviewer. Read the editorial method and corrections policy.

About 8 min read

On this page

Use this worksheet when a campaign readout, attribution report, MMM, lift test, brand study, or attention dashboard is being used to justify a renewal, budget shift, scale, cut, repair, or stronger test.

The worksheet turns a meeting into an evidence record. It does not decide whether a campaign worked by itself. It makes the requested action, evidence rung, included rows, excluded rows, threshold, uncertainty, and next test visible before the recommendation is written.

Editorial desk scene showing mixed campaign reports organized around one central budget decision worksheet before a meeting. — A budget meeting usually starts with too many signals and one decision to make. The worksheet forces the team to name the action first so the evidence can be judged against that action instead of against a vague sense that the campaign looked good or bad.

Use the evidence ladder Choose the next method

1. Decision context

Decision requested today

Budget, package, channel, market, or flight affected

Meeting owner, date, and final approver

Decision deadline and consequence of waiting

Action under consideration

Repair tracking, creative, destination, pacing, or lead routing. Renew the same scope with bounded language. Shift mix among packages, placements, audiences, devices, or creative. Increase, reduce, or pause material spend. Add a lift test, geo test, holdout, calibration check, or better baseline.

2. Evidence rung

Select the highest rung the evidence clears. If a critical evidence lane is weaker, use the weaker rung for the final recommendation.

Evidence ladder board showing campaign report cards feeding into stronger or weaker budget action choices. — The evidence rung is the meeting's speed limit. Delivery and response evidence can justify repair or cautious renewal, but bigger budget actions need stronger comparison evidence before the recommendation moves up the ladder.

0. Missing source trail: totals or summaries without the fields needed to inspect the claim. 1. Delivery and exposure evidence: clean delivery, viewability, frequency, pacing, and quality checks. 2. Observed response evidence: qualified visits, leads, conversions, matches, or survey responses after quality filters. 3. Directional comparison: prior period, matched context, benchmark, model estimate, or survey-control result with caveats. 4. Preplanned comparison: holdout, matched market, fixed baseline, or readout rule defined before results were visible. 5. Repeatable decision-grade evidence: several credible tests, calibrated model evidence, stable operations, and clear uncertainty ranges.

Weakest material evidence lane

Strongest supportable claim today

3. Rows used and rows excluded

Worksheet review board separating included rows, excluded rows, threshold markers, and uncertainty ranges before a decision. — Rows, thresholds, and uncertainty belong in the same view. If excluded rows, immature outcomes, or a low-end estimate would change the decision, the meeting record should show that before the final sentence is approved.

Rows used in the decision

Rows excluded from the decision

Quality flags that may change interpretation

Result after exclusions or sensitivity checks

4. Method and limit check

Evidence source	Can support now	Needs more before	Limit to write down
Attribution report	Tracking, path, and timing diagnosis.	Budget lift or channel incrementality.	Pre-existing demand and unobserved touches are not resolved.
Campaign readout	Delivery fit, response quality, renewal with caveats, or next-test planning.	Broad causal impact language.	Observed response depends on reporting grain and comparison quality.
Brand study	Surveyed awareness, recall, favorability, or consideration movement.	Sales, profit, or pipeline claims.	Sample source, balance, wording, and outcome relevance govern the decision.
Attention metric	Exposure-quality and creative diagnostics.	Business impact or universal value scoring.	Attention is not the same as incremental outcome change.
Lift or geo test	Bounded causal readout inside the assigned population, market, outcome, and window.	Generalizing to every future audience, season, or spend level.	Leakage, balance, compliance, and uncertainty still matter.
MMM	Planning ranges and scenario discussion when controls, priors, calibration, and uncertainty are visible.	Treating modeled contribution as settled causal truth.	Calibration and sensitivity determine how much budget language the model earns.

5. Threshold and uncertainty

Business threshold for acting

Uncertainty range or confidence statement

Decision if the true effect is near the low end

Decision if the true effect is near the high end

6. Meeting closeout

Final budget action board with checked evidence cards, uncertainty range, and repair, retest, or renewal action icons. — The closeout should leave one bounded action, not a stronger story than the evidence earned. A useful meeting record says what to do now, what language is allowed, and what must be measured before the next larger budget move.

Final action

Exact language allowed in the readout

Next uncertainty to measure

Evidence request before the next budget change

Primary source trail and a simulated decision gate

This worksheet treats attribution as a credit-allocation record, not as proof that advertising caused the outcome. Google Analytics describes attribution as assigning credit to touchpoints on a path to a key event, while its data-driven model uses converting and non-converting paths. Google also documents that the selected lookback window controls which earlier touchpoints can receive credit. Those settings are important fields in the meeting record, but they do not replace a preplanned comparison or a clear counterfactual.

Google Analytics: Get started with attribution - defines attribution and explains how data-driven attribution uses converting and non-converting paths.
Google Analytics: Select attribution settings - documents the reporting model, creditable channels, and key-event lookback window.
Google Ads: About attribution models - explains that the model changes conversion counting and can affect conversion-based bidding.
Google Ads: About the Experiments page - describes splitting budget or traffic between an original campaign and an experiment for a defined comparison period.

Simulated worked example - not a live campaign: assume a meeting packet contains 12,000 qualified visits and 360 reported conversions, but only 300 are mature under the team's stated outcome window. A comparable prior period contains 10,000 qualified visits and 250 mature conversions. The current mature conversion rate is 300 / 12,000 = 2.50%; the prior rate is 250 / 10,000 = 2.50%; the observed rate ratio is 2.50% / 2.50% = 1.00x. The 60 immature conversions, any change in traffic mix, and the non-preplanned comparison cap the written recommendation at "retain or repair while planning a stronger comparison," not "the campaign generated lift" or "increase budget."

Assumptions and limits: the example assumes the visit definition, maturity rule, period comparability, attribution settings, and deduplication rule are known and stable. The conclusion changes if any of those fields change, or if an experiment shows a credible difference between comparable arms. Use the worksheet's Weakest material evidence lane, Result after exclusions or sensitivity checks, and Evidence request before the next budget change fields to expose those changes.

What the verified sources establish

Google Analytics documents attribution as a way to assign reporting credit across touchpoints; its model, eligible channels, and lookback-window settings can change how key events and conversions appear in reports. Google Ads likewise explains that an attribution model changes how conversion credit is counted and used for bidding. Its Experiments guidance describes splitting budget or traffic between an original campaign and an experiment for a defined period, then evaluating the comparison before applying a change.

Source-use boundary: those platform mechanics do not validate the simulated visits, conversions, maturity rule, business threshold, or any claim that the campaign caused incremental outcomes. A prior-period comparison is not an experiment merely because it is numerically comparable. The meeting owner must attach the underlying export or study record, state the attribution and lookback settings, and write the comparison design before treating a budget action as supported.

Simulated packet acceptance pass - not a benchmark

Score the example against the six worksheet gates below using 0, 1, or 2 only after writing the evidence used. A score of 2 means the field is inspectable and fit for the proposed meeting action; 1 means it is present but limited; 0 means it is missing.

Gate	Example evidence	Score
Decision and scope	The meeting asks whether to retain or repair while planning a stronger comparison; the larger increase-budget action is not supported.	2
Source trail and row scope	Qualified visits and mature conversions are visible, but placement, creative, and excluded-row details are not attached.	1
Outcome maturity	300 of 360 conversions are mature, but the example does not demonstrate qualification or deduplication.	1
Comparison design	A prior period is shown, but it was not a preplanned holdout or experiment.	1
Threshold and uncertainty	No interval, minimum effect, cost ceiling, or decision threshold is supplied.	0
Closeout and reversibility	Retain/repair is bounded, but an owner, due date, and post-action readout are not yet recorded.	1

The reproducible total is 2 + 1 + 1 + 1 + 0 + 1 = 6 / 12 = 50%. Because the threshold-and-uncertainty gate is zero, the packet supports a retain-or-repair handoff and a stronger comparison request, but it does not support scale, cut, lift, or a new planning assumption. Re-score the same six gates after attaching the missing rows, threshold, owner, and closeout readout; the percentage is a packet diagnostic, not a universal quality bar.

Downgrade rules

If the meeting finds	Downgrade to	Better next action
The recommendation depends on attributed conversions without a comparison group.	Observed response evidence.	Renew cautiously, repair tracking, or plan a holdout.
The strongest result is a small slice without minimum delivery or quality thresholds.	Directional comparison or weaker.	Retest with minimum cells and fixed slice rules.
A model rank changes materially when priors, exclusions, or calibration anchors change.	Directional planning evidence.	Use ranges, name assumptions, and require sensitivity review.
A survey result lacks sample source, balance, question wording, or uncertainty.	Weak survey evidence.	Request method detail before budget language relies on the result.
Outcome rows are still immature or missing buyer-side status.	Preliminary observed response.	Close the status window before final renewal language.

Pair with

Use this worksheet after the advertising budget decision evidence ladder sets the evidence rung. Pair it with the measurement method selector before choosing a design, the next measurement method guide when the final action needs stronger evidence, the campaign evidence triage tree when the first weak lane is unclear, the uncertainty interval checklist when a threshold needs a range, and the campaign renewal memo template when the meeting result needs final action language.

Keep reading

Choose the next measurement check

Move from this page into method choice, baseline review, and uncertainty language before the evidence is overread.

Method choicePick the evidence designChoose between MMM, lift tests, geo tests, brand studies, attention metrics, and attribution reports. Baseline checkName the comparisonCheck prior-period, matched, holdout, and modeled baselines before observed response becomes lift. UncertaintyBound the conclusionRead intervals, thresholds, noisy slices, and decision language before calling a result meaningful.