# What Google Ads First-Shown and Last-Shown Dates Can Tell You

First-shown and last-shown dates do not prove that an ad ran continuously between them. The elapsed span between reported endpoints is distinct from actual active delivery. Keep source dates and precision, and leave gaps or daily totals unknown unless source evidence proves them.

An Actor is a hosted data collection program. A Task is a saved input preset. Google Ads Transparency build 0.2.7 saved targets use apple.com and samsung.com, requested region US, platform ALL, adFormat ALL, includeDeliveryDetails false, and maxItems 10. Change the whole target list by replacement to one intended known creative or advertiser input, and use maxItems 2 shared across targets. The response includes advertiserId, creativeId, creativeUrl, firstShownAt, lastShownAt, approxDaysShown, requestedRegion, requestedPlatform, requestedFormat, observedAt, status, and statusReason. Current two direct-creative rows prove date-field presence only, not domain discovery, regional search, daily delivery, or full history. A domain can resolve multiple accounts, so compare exact advertiserId and creativeId.

## Complete endpoint-evidence method

1. State the period claim before using dates. Date interpretation is a supported prerequisite. Do not infer campaign duration from a long date range.
2. Match the exact advertiser and creative identity with its source link. Keep region, platform, format filters, and observation time. The same creative ID in a different context may show different evidence. Do not merge records by brand name alone.
3. Preserve firstShownAt and lastShownAt and optional approxDaysShown as separate source fields. Check public source date meaning and precision. Parsed ISO timestamps may be calendar labels or derived values; do not assert all are calendar labels or certified second-level delivery events. A missing endpoint remains unknown.
4. Calculate an endpoint span only under a declared rule. For known timestamps, elapsed duration is last minus first. For calendar-day labels in one time zone, inclusive calendar span is the date difference plus one. These units differ, and neither is an active delivery count. Do not name the result active days.
5. Identify what evidence is needed for continuity: a source delivery record for each day. Two endpoints leave the interior unresolved. A region filter returns public observations, but it does not prove daily spend or private impressions. The Actor does not fill missing days.
6. Keep approximate source day labels distinct from the computed span. Do not explain a mismatch as pauses or tracking failures without evidence. If a label is null, record that state. Even a label equal to the span is not proof of every day's delivery.
7. Handle contradictions before a report. A last date before a first date, an unknown time zone, or a scope mismatch needs review. A last date before today does not prove the campaign ended. A last date near collection time does not guarantee live delivery.
8. Deliver a period table with source endpoints, precision, declared span, optional source day label, continuity state, and unresolved reason. Use observed endpoint range when that is all the evidence supports. If uninterrupted delivery matters, use appropriate authorized owned evidence only when available; competitor public records do not give private delivery access.

## Fictional endpoint-evidence examples

The table below shows fictional records to test date logic. All dates and identities are fictional.

| Record | First Date | Last Date | Approx Days | Declared Span | Continuity State | Unresolved Reason |
| --- | --- | --- | --- | --- | --- | --- |
| C1 | Oct 1 | Oct 5 | 3 | 5 days inclusive | unknown | interior gaps possible |
| C2 | Oct 4 | null | unknown | unknown | unknown | missing end date |
| C3 | Oct 8 | Oct 6 | unknown | invalid | needs_review | last date before first |
| C4 | Oct 7 | Oct 7 | unknown | 1 day | unknown | single calendar endpoint |
| C5 | Oct 1 | Oct 5 | 5 | 5 days inclusive | unknown | label definition unknown |

Record C1 has known calendar dates in UTC. Inclusive calendar span is five days. Elapsed time between midnights is four days. Neither measure proves ads served on October 2 through 4. Two schedules fit these endpoints: running on October 1 and 5 only, or running every day. Both match, so continuity remains unknown.

Record C1 also has an approximate source day label of three with unknown basis. Preserve this separately. Do not assert the exact schedule or overwrite it with five. Record C2 has a null last date, making span and continuity unknown. Record C3 has an invalid date order where the last date is before the first date, which needs review. Record C4 has matching first and last dates, creating a one-day endpoint span, not a continuous 24-hour claim. Record C5 has endpoints and a label of five, but equality does not prove every day's delivery without the label rules. A valid endpoint calculation describes known boundaries, not a performance estimate.

## Next action

To record one source-linked endpoint range with daily continuity unknown while keeping approximate-day basis unknown, source timestamps versus observation time, single-day not 24 hours, and current delivery unknown, check the [Apify Google Ads Transparency Scraper Actor](https://apify.com/chronometrica/google-ads-transparency-scraper) and the [Google Ads Advertiser Monitor Task](https://apify.com/chronometrica/google-ads-transparency-scraper/examples/google-ads-advertiser-monitor) to run bounded checks.

## Sources

- [Google Ads Transparency Policy Help:](https://support.google.com/adspolicy/answer/13733850?hl=en-CA)
- [Competitor Paid Ads Analysis Workflow:](https://www.reddit.com/r/slavelabour/comments/1g3c5ep/task_competitor_paid_ads_analysis/)