Decide when to freeze an audience and when to refresh it
Prepared with AI assistance by Alactic Editorial. Examples identified as hypothetical are illustrative, not customer results.
A reviewer approves an audience, but the same filter returns different people when someone runs it again. Some contacts have purchased, others have opted out, and newly active contacts now qualify. The filter may be unchanged, yet the population is no longer the one reviewed.
Freeze membership when approval applies to named records. Refresh membership when approval applies to a rule and an explicitly defined class of later changes. In either case, preserve the original review evidence and recheck conditions that can invalidate eligibility. A frozen list should not preserve permission that has been withdrawn or a need that has been disproved.
The practical choice is therefore not simply static versus dynamic audience membership. It is deciding what approval covers, which facts must remain current, and how additions and removals become visible to the next reviewer. The recommended process below concerns audience preparation and review, not delivery or scheduling.
1. Decide whether the reviewer approved people or logic
Start with one question: what exactly did the reviewer accept?
- A named population: the reviewer inspected particular contacts, exceptions, or supporting evidence. Preserve that population as the review baseline.
- A selection rule: the reviewer accepted criteria that can admit different contacts later. Refreshing may be appropriate, but only within an agreed change policy.
- Both: the reviewer accepted the rule and inspected its current results. Record whether that inspection authorises future entrants or only the current members.
The third case causes many disagreements. One person thinks approval means “this definition is acceptable.” Another thinks it means “these records are acceptable.” Neither interpretation should be left implicit.
A useful approval statement is: Approval covers the reviewed record IDs under rule version R1. Later exclusions must be applied. Later entrants require a separate review. A different, equally valid statement might authorise refreshed membership under an unchanged rule, provided specified evidence checks pass and the membership difference is recorded.
These statements approve different operating choices. Neither should be inferred from a generic “approved” label.
2. Separate the review baseline from current eligibility
A snapshot answers who was in the audience at a particular point. Current eligibility answers who still satisfies the conditions now. Keep those questions separate.
Preserve the reviewed baseline rather than rewriting it whenever a record changes. Then create a later eligibility decision for each affected record. This lets a reviewer see that someone was correctly included earlier but became ineligible after a new purchase, or that earlier inclusion relied on evidence later found to be wrong.
A practical design uses three conceptual groups:
- Reviewed baseline: the record identifiers and evidence references considered during approval.
- Retained members: baseline members who still pass the required checks.
- New entrants: records outside the baseline that qualify under a refreshed evaluation.
For a snapshot approach, later eligibility checks can remove people, but they do not silently add people. For a refresh approach, additions may be allowed, but only under the approved rules for new entrants.
This is an operating model, not an assertion that a particular product stores these groups automatically. In Alactic, check the workspace and current plan for the available audience, export, history, and review functionality before choosing how to implement it. Where necessary, use a controlled review record with restricted editing access.
3. Use a filled snapshot-versus-refresh decision table
Fictional worked example: a team is reviewing an audience for assistance with an unfinished product setup. Its hypothetical rule requires recorded setup need and qualifying activity within the previous 30 days. It excludes contacts who have purchased the assisted setup service, contacts with a relevant recorded exclusion, and contacts without the required recipient permission.
The reviewer approves a hypothetical baseline of 120 records at review point A. The team then evaluates changes at checkpoint B. The filled decision table below is presented as row cards so that each change includes its consequence and review requirement.
Row 1 — New purchase after review
- Changed record: Contact A was included at point A, then purchased the assisted setup service before checkpoint B.
- Snapshot decision: retain the historical record of inclusion, but remove Contact A from the currently eligible subset.
- Refresh decision: exclude Contact A under the unchanged purchase exclusion.
- Does this invalidate earlier inclusion? It invalidates continued inclusion from the purchase onward. It does not mean the earlier decision was wrong if the purchase had not happened yet.
- Evidence to record: purchase reference, purchase time, time recorded, and the rule that makes the purchase disqualifying.
Row 2 — Newly recorded exclusion
- Changed record: Contact B was included at point A. A relevant opt-out is recorded before checkpoint B.
- Snapshot decision: remove Contact B from current eligibility; the frozen baseline is not an exception to the exclusion.
- Refresh decision: exclude Contact B even if activity and need criteria still match.
- Does this invalidate earlier inclusion? Yes for continued inclusion. If evidence shows the opt-out existed before point A but was recorded late, the earlier decision also needs correction.
- Evidence to record: exclusion type, applicable scope, event time where known, recording time, and source reference. If scope is unclear, hold the record for review rather than treating uncertainty as permission.
Row 3 — Corrected need evidence
- Changed record: Contact C was included because a note said setup was incomplete. A correction shows that the note referred to an old account and the current setup was already complete at point A.
- Snapshot decision: remove Contact C from current eligibility and mark the earlier inclusion as based on incorrect evidence.
- Refresh decision: exclude Contact C because the current evidence does not establish the required need.
- Does this invalidate earlier inclusion? Yes. Unlike a later purchase, this correction undermines the basis for inclusion at the original review point.
- Evidence to record: the superseded evidence reference, correction reference, affected account, and the reviewer responsible for resolving the conflict.
Row 4 — Newly qualifying activity
- Changed record: Contact D was outside the baseline because there was no recent qualifying activity. New activity before checkpoint B brings the contact within the hypothetical 30-day rule.
- Snapshot decision: do not add Contact D to the approved baseline. Place the record in a proposed addition set if the team wants it considered.
- Refresh decision: admit Contact D only if refreshed additions were authorised and all other conditions, including need, exclusions, and permission, pass.
- Does this invalidate earlier inclusion? No. It creates a possible new inclusion; it does not change the validity of the records already approved.
- Evidence to record: qualifying activity reference and time, rule version, completed checks, and the authority permitting additions.
The key distinction is whether a change removes eligibility, corrects an earlier mistake, or creates a new candidate. Those outcomes require different records even when all three change the final audience count.
4. Identify the conditions that cannot stay frozen
Not every field needs another review. Recheck facts whose change would alter the decision. Write those facts into the approval record instead of relying on a general instruction to “check the data.”
- Permission and relevant exclusions: identify whether the contact is still permitted for the intended purpose and whether an opt-out or other exclusion applies.
- Disqualifying events: check purchases, completed tasks, or other events that remove the reason for inclusion under the specific audience rule.
- Need evidence: check for corrections, withdrawals, conflicting records, or evidence that has passed an agreed validity period.
- Time-dependent criteria: reevaluate activity windows and other conditions that can expire even without a new event.
- Identity and account linkage: resolve merges or corrections that might attach evidence to the wrong person or account.
For WhatsApp-related audiences, recipient permission and opt-outs are substantive requirements, not optional refresh preferences. Consult the current WhatsApp Business Messaging Policy for precise requirements. Audience approval is also separate from template review and actual delivery.
Document what happens when a required check cannot be completed. A defensible review design holds the affected record out of the confirmed eligible set until the uncertainty is resolved. An empty field should not quietly become a passing result.
5. Choose the trade-off deliberately
Prefer a reviewed snapshot when people were assessed individually, evidence requires judgement, or additions would exceed the reviewer's intended scope. Its advantage is a stable baseline. Its cost is that newly qualifying records need another decision, and the baseline can become less representative over time.
A snapshot still needs current exclusions. Otherwise, its apparent stability conceals stale eligibility.
Prefer refreshed membership when the audience is intentionally defined by current conditions, the rule is sufficiently clear, and the reviewer has authorised additions under that rule. Its advantage is relevance to current facts. Its cost is that the final population may contain records nobody examined during the original review.
A refresh still needs a preserved baseline and a membership difference. Otherwise, reviewers cannot distinguish an expected addition from accidental rule drift.
Use a hybrid when individual inclusion matters but exclusions must remain current: preserve the reviewed members, subtract newly ineligible records, and put new candidates into a separate review set. This does not solve every case, but it makes approval boundaries easier to explain.
6. Record changes as decisions, not just totals
A count can stay unchanged while the audience changes substantially. A hypothetical audience might lose five reviewed members and gain five unreviewed members while still showing 120 contacts. Matching totals do not establish matching membership.
For each checkpoint, record:
- The baseline identifier, review point, rule version, and reviewer.
- The chosen mode: snapshot, authorised refresh, or hybrid.
- The later evaluation point and the evidence available then.
- Retained, removed, added, and unresolved record identifiers.
- A reason and evidence reference for each change.
- Whether the change affects only current eligibility or also corrects the original review.
- Who authorised additions or resolved an exception.
Distinguish the time an event happened from the time it became available in your records. A late-recorded purchase can explain why the original review missed a disqualifying fact. If the event time is unknown, record that uncertainty rather than assigning an invented date.
Changing the rule itself is different from refreshing its results. Expanding the qualifying activity definition or weakening a need requirement creates a new rule version and requires explicit approval of that revised logic.
7. Recover when the reviewed and selected populations diverge
If someone discovers that a later selection differs from the approved audience, stop treating the earlier approval as sufficient evidence for the new population. Preserve both versions before making corrections.
Compare identifiers, not names or counts alone. Separate departures from additions, then classify each difference: expected eligibility change, corrected evidence, rule change, identity change, or unexplained difference.
Remove newly disqualified records from the confirmed eligible set. Keep unapproved entrants separate. If refreshed additions were already authorised, verify that they passed the agreed checks; do not assume that matching the main filter proves this.
Reopen review when the difference exceeds the authority originally granted, when evidence conflicts remain unresolved, or when the rule has changed. The objective is not to force the new population back to the old count. It is to restore a traceable relationship between approval, evidence, and current eligibility.
8. Test the decision policy on one audience
Before the next review, write a short approval contract covering the membership mode, required rechecks, treatment of new entrants, and conditions that reopen review. Assign an owner for unresolved changes.
Then run a membership comparison at the next review checkpoint. Measure how many baseline members remain eligible, how many were removed by reason, how many candidates were added, and how many differences remain unexplained.
A practical acceptance criterion: every membership difference has a recorded reason, and every addition has a documented authority path. If that test fails, fix the review record before treating the changed audience as approved. This makes the choice between freezing and refreshing observable rather than merely a preference.
Put the next conversation to work.
Explore Alactic Communications for WhatsApp conversations, campaigns and AI-assisted workflows. Choose the product and plan that fit your team.
Explore Communications