The Alactic library
Campaign measurement guides
From a better WhatsApp conversation to a clearer team handoff. Find a specific decision, a useful checklist or a workflow your business can put into practice.
20 articles in Campaign measurement
Clear filtersVerify that an apparent campaign winner survives a confirmation test
A top-performing variant is a candidate, not yet a repeatable improvement. Use a fresh cohort, a fixed baseline and a written decision rule to find out whether the apparent winner holds up.
9 min readChoose between campaign variants when reply rate and outcome quality disagree
A higher reply rate is not automatically a better campaign. Use a predeclared decision rule, consistent denominators, and separate guardrails to choose a variant without moving the goalposts.
10 min readKeep campaign outcome coding consistent across reviewers
Use a shared sample, independent labels and a versioned rule log to separate real campaign performance changes from changes in reviewer judgement.
10 min readCompare campaign cohorts only after equal outcome observation time
A recent campaign can look weak simply because its recipients have had less time to act. Use matching observation windows, explicit maturity labels, and a release schedule to make fairer comparisons.
10 min readMeasure repeated campaign exposure without counting people repeatedly
Separate message deliveries, campaign recipients and period-level reach. Use a filled fictional ledger to see who responded across repeated exposures—and why higher response rates do not establish a frequency effect.
10 min readCompare campaigns fairly when recipient mixes differ
A higher response rate may reflect a warmer audience rather than a better campaign. Use this fictional worked comparison to inspect recipient mix, compare like groups, and recognise when you need a randomised test.
10 min readFind where campaign engagement stops becoming useful outcomes
A busy reply inbox can hide a weak campaign. Use a recipient-level outcome chain to separate delivery, response, useful response and completion—and identify the transition that needs investigation.
10 min readInclude negative responses in campaign success decisions
A rise in useful replies is not enough to justify a rollout. Use a paired decision scorecard to measure withdrawals, complaints, wrong-recipient replies and unresolved classifications before expanding a campaign.
10 min readMeasure reply quality instead of treating every response as success
A high reply rate can conceal confusion, wrong recipients, and polite acknowledgements. Use a campaign-specific coding guide to distinguish useful outcomes from activity without hiding uncertainty.
9 min readSet stopping rules before checking campaign test results
Use a fixed analysis schedule, protect recipients without declaring premature winners, and document what happens after pauses, sample shortfalls and conflicting results.
9 min readDecide whether a campaign test has enough evidence to act
A higher reply rate is not automatically a winner. Use a minimum worthwhile effect and an uncertainty interval to decide whether to launch, confirm, or close a test as inconclusive.
9 min readTest one campaign change without confounding the result
When offer, wording, audience, and timing all change, a better result cannot explain what worked. Use a filled experiment brief to isolate a message contrast—or deliberately test a complete campaign package.
10 min readChoose the right randomisation unit for a campaign experiment
When colleagues or household members can influence one another, randomising individual recipients can weaken your comparison. Use this worked decision table to choose a defensible assignment unit and plan the analysis.
10 min readUse a holdout group to estimate incremental campaign outcomes
Campaign attribution does not show what would have happened without the campaign. Use a randomised recipient holdout, fixed outcome window and intention-to-treat analysis to estimate the difference.
10 min readHandle late campaign replies without silently rewriting reports
Keep original campaign reports intact while making late replies and corrections visible. Use a three-clock policy and a filled reporting register to explain every revision.
10 min readAvoid double-crediting outcomes across overlapping campaigns
Use one outcome ledger and a transparent credit rule to reconcile overlapping campaign reports. A worked example compares first-touch, last-touch, and shared credit.
9 min readSet an attribution window that matches the campaign decision
Choose response and outcome windows from observed delays, campaign purpose, and the decision your report must support. Use a filled worksheet to compare cutoffs and see whether the conclusion changes.
9 min readReport campaign results when delivery status is incomplete
Managers need a result even when delivery records are unresolved. Use a fixed reporting cutoff, separate observed responses from delivery evidence, and publish conditional bounds instead of a falsely precise rate.
10 min readChoose the right delivery denominator for campaign reply rates
Selected, attempted, and confirmed-delivered recipients answer different campaign questions. Use this filled worksheet to match each denominator with the right responders.
8 min readCount unique campaign responders without inflating results
Separate unique responding endpoints from reply volume with a clear counting rule, a worked ledger and a reconciliation that exposes duplicates and uncertain attribution.
10 min readTurn a guide into a working customer journey.
Explore the tools and services behind the operating process.