Set the conversion observation window before a creator campaign begins

Direct answer

Fix the rule for the conversion observation window after the team has named the outcome and its evidence source, but before the first creator post or paid ad is live. Write what starts the reporting period, what ends it, the time zone, which event is counted, which business system confirms it, and when the report will be read again for late or corrected data. The actual times can be filled from verified publication and event receipts as they occur. Do not wait to see a favorable conversion curve and then choose the window that makes it look best.

This campaign rule is a decision made by the team. It is not the TikTok Ads Manager ad attribution window, and it is not Google Analytics 4's key-event lookback window. Those products use their own rules to decide which prior touchpoints may receive credit. They do not choose the client report's calendar boundary or prove that the campaign caused an outcome. Use the blank window-decision card to keep the three definitions together without merging them.

Three windows that answer different questions

Google defines its key-event lookback window as how far back a touchpoint can be eligible for attribution credit. For most key events the documented default is 90 days, with 30- and 60-day alternatives; the first_open and first_visit events default to 30 days and engaged-view key events to three days. Those are product defaults to verify, not a recommendation to watch every creator campaign for 90 days. Google says lookback changes apply going forward across the property, and the chosen key-event lookback also applies to session attribution. Record the property's current value and the event type rather than silently changing a shared setting for one campaign.

For paid TikTok activity, TikTok Attribution Analytics distinguishes actions after clicks and views and provides a comparison of results under different attribution windows. Its Performance Comparison guide also offers an ad-impression-time or real-time date filter. Those are choices about TikTok attributed ad reporting. TikTok's September 2026 auction guidance describes different web attribution behavior depending on Advanced Matching, so record the effective setup for the particular ad and event source instead of importing an app-specific setting or an old default. If the effective rule is not visible in the account record, mark it unresolved and ask the ad operator for its source.

Name the outcome and its authority first

The window has no meaning until “conversion” names an observable action. An accepted lead, a booked consultation, a paid order and a qualified opportunity have different dates and record owners. Write one primary outcome and the system that confirms it. A GA4 generate_lead key event may show that a configured event fired; it does not by itself confirm the CRM accepted and qualified the lead. A tagged visit does not confirm an order. If the business question is revenue, keep the reconciled order or payment record, refund rule and currency beside any analytics report. The GA4 key-event guidance distinguishes an important event collected in Analytics from the separate ad-conversion reporting surface.

State the unit before choosing a duration: unique qualified inquiries, accepted bookings, orders net of returns, or something else. Name how duplicates, cancellations and later corrections will be treated. This is an editorial measurement method, not a claimed GA4 or TikTok feature. It keeps a report from turning a tracking event into an invented business outcome.

If the outcome cannot be observed in the chosen system, record it as unknown and choose a narrower available measure. The decision about *which objective is measurable* belongs before this one; choose the objective and its evidence path first. This article starts once the team can name the outcome source. The window-decision card has a field for the authority and its exact event definition.

Freeze a date rule, not a convenient result

Write the start and end as a reproducible rule in the measurement plan. For example: “include confirmed lead submissions timestamped at or after the first dated public-post observation and before 00:00 UTC on the day after the fourteenth complete calendar day following the last verified post date.” The start is inclusive and the end is exclusive, so every timestamp on the last included day still counts. Fourteen days here is merely a possible team choice; it is not a platform standard or a conversion benchmark. Starting at the first observation conservatively excludes any unobserved earlier exposure; it does not prove the post first appeared at that time. The right duration depends on the campaign question, decision deadline and the brand's evidenced buying or qualification cycle. If those inputs are unknown, state the uncertainty and choose a clearly provisional rule before launch.

Keep planned dates apart from observed dates. A scheduled creator post is not a public post; a configured Spark Ad start is not evidence of delivery. Save the exact public post URL and dated verification. The first observation proves only that the post was public by then. If a reliable earlier publication timestamp later appears, preserve both receipts and record the revised start rule explicitly rather than quietly replacing the original.

Use one named report time zone. GA4's property setup guidance says its reporting time zone determines the date assigned to a visit; TikTok's account time-zone guide says the account zone affects report data and delivery settings. A creator's local publication clock can cross midnight differently from either report. Store the original timestamp and zone, then derive the report date in the declared zone. Do not compare daily rows until both windows are expressed on the same basis.

Record platform attribution without calling it the campaign window

If the campaign has only creator organic posts and tagged links, there is no TikTok Ads Manager ad-group attribution window for those posts. The team can still set its own outcome observation period and record what the site or business system sees. If it later runs a Spark Ad, add the exact ad identity, paid start evidence, effective click-through and view-through rules, event source and report view to a new version of the card. Decide the paid lane's own outcome period, or explicitly revise a combined team rule, before ad delivery; retain the original organic rule and report instead of silently extending its end date.

For the same paid activity, TikTok and GA4 can assign credit under different rules. TikTok's Performance Comparison guide explicitly lets the operator view by ad-impression time or the actual event time; it recommends ad-impression time when comparing with Ads Manager. GA4's attribution settings specify the model, channels eligible for credit and key-event lookback. In GA4, the reporting attribution model affects event-scoped traffic dimensions, while session-scoped dimensions such as Session source are unaffected by that model change. Write down the exact report and dimensions used, not merely “GA4 conversions.”

This does not require forcing the two platforms to agree. An ad exposure can qualify for a platform's credit while the event falls outside the campaign's chosen outcome dates, or while GA4 credits another touchpoint. Conversely, an outcome in the business window might have no supported TikTok attribution. Neither mismatch alone measures incremental effect. Do not add the two platform totals or treat one as a person-level reconciliation of the other.

Decide when a number is preliminary and how to revise it

Record an as-of export time and the source's completeness state. Google's data-freshness guidance says GA4 processing can take 24–48 hours and key-event attribution credit can change for up to 12 days after a key event is recorded. Those are platform processing limits, not an automatic declaration that every order or CRM outcome is final after 12 days. A client report issued earlier can be useful if it is labeled preliminary and its coverage is explicit.

If a creator posts late, an ad starts later, a link is repaired, or the client changes the outcome definition, append a dated version of the window rule with the reason and effective period. Preserve the earlier rule and any already-sent report. Do not quietly slide the end date because later conversions looked attractive, or erase a missed period because the result disappointed. A later export can update the observation while retaining the original as-of value. The blank CSV provides one record for each reporting lane and decision version.

Fictional plan with no campaign result

The following dates and team decisions are fictional teaching data. No Superdirector client, creator, TikTok account or GA4 property was measured. A fictional team wants to count CRM-accepted inquiries for a two-post creator campaign. Before posting, it agrees that the organic report will include inquiries whose CRM acceptance timestamp is at or after the first dated public-post observation and before 00:00 UTC on the day after the fourteenth complete calendar day following the last verified post date. It will list the CRM record as the business authority, keep generate_lead as a site diagnostic, and name UTC as the report zone. The 14-day choice reflects this imaginary team's decision deadline; it is not a recommended duration.

Suppose the two fictional posts are first verified public on 3 November 2026 at 10:00 UTC and 5 November 2026 at 12:00 UTC. Under the stated rule, the outcome period would start at 3 November 10:00 UTC, inclusive, and end at 20 November 00:00 UTC, exclusive; all of 19 November is included. These are observation and reporting boundaries, not proof that either post first appeared at its check time. If the second post were instead first verified on 7 November, the exclusive end would move to 22 November 00:00 UTC, and the team would keep both the planned and observed post times in the card. No lead or sale count is supplied in this example.

If a paid Spark Ad is added, the team records its separate effective attribution rules and the source/date basis of its TikTok export. The fictional organic report does not automatically expand: a paid outcome lane needs its own pre-agreed dates, or an explicitly revised combined rule, before delivery. A hypothetical 7-day click setting would not convert the campaign's calendar report into a “7-day campaign,” and a GA4 property lookback would not extend the CRM period. The values remain separate even if one vendor later credits a conversion to an earlier interaction. The window-decision card shows the three rules side by side; its CSV provides blank fields for a separate record per lane and decision version.

Use the decision card before the first post

Open the window-decision card with the campaign owner and measurement owner. Name the outcome, source and rule first; then copy the TikTok and GA4 settings that actually apply. If a platform does not apply, write “not applicable” with the reason. Save the blank CSV template beside the brief and source receipts; add one record for each reporting lane and decision version. The team should be able to reproduce the same inclusion rule after launch without remembering what the dashboard looked like on decision day.

Editorial source note

Editorial source note: Google's Analytics documentation supports the key-event lookback, scope of attribution settings, property time zone and data-freshness limits. TikTok's help pages support the distinction between click/view attribution windows, available comparison views and account time-zone effects. The campaign outcome window, versioning rule and fictional calendar example are original editorial methods, not a platform feature or an observed campaign result. Sources were checked on 2026-10-28. Search volume and difficulty were not measured; the saved, privacy-limited GSC export has no exact window-query row, which does not establish zero demand. The working phrase is an intent hypothesis.