TL;DR
The most expensive Black Friday mistake is not a broken pixel. It is a healthy account whose owner read the dashboard too early, concluded the ads were failing, and made a structural change at noon that put every ad set back into learning for the rest of the weekend. The platforms themselves tell you their numbers run behind: Google documents its freshness tiers precisely [1], puts offline imports at up to 72 hours [3], and Meta's reporting continues to shift for days as delayed and modeled conversions land. Even the reporting systems themselves are not immune to the day: on Black Friday 2022, Amazon's ads console showed advertisers spend figures roughly half of what they expected, from Friday afternoon until Amazon acknowledged and resolved the lag that Sunday, while ads kept serving: a reporting failure, not a delivery one [4]. This article is a reading schedule: what each number means at each hour of the day, which comparisons are safe, and the two conditions that actually justify touching a campaign mid-peak.
Key Takeaways
- Shopify's order count is the only number on your screen that is correct in real time all day. Anchor every other reading to it.
- Google Ads conversion columns refresh on documented schedules: about 3 hours behind for last-click and up to 15 hours for non-last-click attribution models [1]; offline conversion imports typically process in under 12 hours but can take up to 72 when keyed on the newer click identifiers [3].
- Meta publishes no overall Ads Manager freshness figure; the delay numbers in its documentation are case-specific. Its conversion counts for a given hour keep being revised upward for days as delayed and modeled events settle; treat same-day Meta ROAS as a floor, not a fact.
- The safe same-day comparisons are directional: hour-over-hour order volume, spend pacing, and click traffic. The unsafe ones are absolute: same-day ROAS, cost per purchase, and any platform-to-platform comparison.
- Only two conditions justify a mid-peak campaign intervention: spend flowing with zero clicks (a delivery failure), or a platform gap that is still widening 24 hours later (a tracking failure). Everything else is lag wearing a costume.
- A structural edit made in a panic restarts platform learning and is paid for across the entire remaining weekend, a cost that argues for freezing campaign structure well before the day.
What the platforms actually publish about their own lag
Google is the only major platform with concrete public freshness numbers, and they are worth reading precisely because they are longer than most advertisers assume [1]:
| Number you are reading | Documented freshness | On Black Friday, current as of |
|---|---|---|
| Google Ads conversions, last-click model | about 3 hours behind [1] | roughly breakfast, by mid-morning |
| Google Ads conversions, other attribution models | up to 15 hours behind [1] | yesterday evening |
| Offline conversion imports (incl. ones keyed on newer click identifiers) | typically under 12 hours, up to 72 [3] | the day before, at worst Tuesday |
| Meta purchases and ROAS | no overall published figure; revised for days | unknown; treat as a floor |
| GA4 standard reports | processing runs on its own delay; same-day numbers are provisional | late morning at best |
| Shopify admin orders | real time | now |
Meta deserves its own paragraph, because the absence of an overall freshness number gets misread as "real time." Meta's reporting restates recent history as delayed events arrive and as modeled conversions (its statistical fill-in for unobservable users) are computed. Practitioners commonly observe revisions for up to about three days; Meta commits to no overall number, so neither will we. The operational consequence is simple either way: the purchases column for this morning will be higher when you look again tomorrow, and higher again on Sunday. Reading it at noon and calling the campaign a failure means judging it on the one version of the number guaranteed to be an undercount.
And when volumes get extreme, the reporting pipes themselves can fail independently of delivery: on Black Friday 2022, Amazon's ads console under-reported advertisers' spend by roughly half for the better part of the weekend while ads continued serving and selling [4]. The lesson generalizes to every platform: a dashboard reading zero is a claim about the dashboard until Shopify's order count corroborates it.
The reading schedule
Here is the timetable we would actually run on Friday, November 27, 2026. Times are your store's local time; adjust the anchors to your own morning.
7 a.m., before coffee: record three numbers in a note: Shopify orders since midnight, total ad spend by platform, total clicks by platform. This is your baseline row. Do not record ROAS.
10 a.m.: second row of the same three numbers. The comparisons that mean something now: is spend pacing to budget, are clicks scaling with spend, are Shopify orders growing hour over hour. If spend is flowing and clicks are near zero on one platform, that is the rare genuine delivery emergency: check the platform's status page and your ad account before touching anything.
1 p.m.: Google's last-click conversion column now roughly covers your morning [1]. It remains blind to the last 3 hours and to every non-last-click model. You may now sanity-check Google's reported conversions against the Shopify orders from your 7 a.m. and 10 a.m. rows, expecting Google to show less. A Google column at, say, 60 to 80% of Shopify's morning orders is ordinary composition (other channels, direct traffic, attribution model differences), a topic we unpacked in why platforms disagree on ROAS.
4 p.m.: compare the platform-to-Shopify gap against the 1 p.m. reading. Narrowing gap means lag is doing its thing; keep shopping running and walk away. This is also the moment to resist the day's biggest trap: the platform-to-platform comparison. Meta at a revision-pending floor versus Google at a 3-hour lag versus GA4 mid-processing produces three different numbers for the same morning, all of them defensibly wrong, none comparable, for reasons that hold on every ordinary day too and that we detailed in the attribution windows and reporting delays explainer.
Saturday morning: Google's non-last-click models have now caught up to Friday afternoon [1]; Meta's Friday is one revision cycle in. This is the earliest hour at which a ROAS reading is worth writing down, and it will still drift upward.
Monday and beyond: standard online conversions' 15-hour tail [1] clears each day the following morning, so by Tuesday, December 1 the Friday-through-Sunday picture is largely in. If you import conversions offline, give Cyber Monday's 72-hour tail [3] until Thursday, December 3. Then reconcile platform conversions against Shopify orders, per platform, and log the final numbers. That reconciliation, not Friday's adrenaline reading, is what should feed next year's budget.
When is touching a campaign mid-peak actually justified?
Two conditions, and their shapes are distinctive:
- Spend without clicks. Money flowing, click column near zero, on one platform while others behave. That is delivery, not measurement, and it will not self-heal. Diagnose in the ad account (disapprovals, billing, account flags) and on the platform's status page.
- A widening gap after 24 hours. Anchor to Shopify orders, measure the platform gap twice, several hours apart, on Saturday. Still widening: something in the delivery chain is genuinely failing, and it is time for a structured check of each layer, starting from a pixel audit rather than a campaign edit.
Everything else, every scary morning ROAS, every platform disagreement, every conversion column that looks light while the order printer runs hot, is lag, and the correct intervention is the 4 p.m. re-read. The asymmetry that makes patience the right default: waiting costs you nothing because delayed conversions backfill with their original attribution, while a structural edit costs recalibration across the platforms' learning systems for days. On the biggest weekend of the year, the campaign that gets left alone outperforms the campaign that gets managed.
FAQ
Why does Google Ads show fewer Black Friday conversions than Shopify all day?
Three stacked reasons: documented reporting freshness of 3 to 15 hours depending on attribution model [1], conversions from other channels that Google never sees, and click-window mechanics. The steady-state version of this question has its own article in why Google Ads reports fewer conversions than Shopify; on Black Friday, add the lag tiers on top.
When is Meta ROAS from Black Friday final?
Meta does not publish a completion time. In practice the number keeps revising upward for one to three days as delayed and modeled conversions land, so read Friday's ROAS on Monday or Tuesday, and treat anything earlier as a floor.
Is a zero in my dashboard ever real on the day?
Sometimes. Check Shopify orders first. Orders flowing plus one platform at hard zero for over an hour suggests a real problem with that platform's event delivery or its reporting pipeline; large-scale reporting failures have happened before [4]. Orders flowing plus all platforms merely low usually means you are early on every lag clock at once.
Should I pre-schedule the reconciliation?
Yes, literally on the calendar. December 1 covers most setups; December 3 if you import conversions offline, so Cyber Monday's 72-hour processing tail has cleared [3]. Per platform, against Shopify's order truth. It is the only Black Friday number that deserves to influence money decisions, and scheduling it in advance is what stops the Friday-morning version from doing so.
Sources
- Google Ads Help, Data freshness for conversion reporting. https://support.google.com/google-ads/answer/2544985
- Google Ads Help, Data discrepancies: factors and troubleshooting. https://support.google.com/google-ads/answer/7457111
- Google Ads Help, Fix discrepancies and errors in offline conversion imports. https://support.google.com/google-ads/answer/13321563
- eMarketer, Amazon Ads reporting mishap on Black Friday. https://www.emarketer.com/content/amazon-ads-reporting-mishap-on-black-friday-gives-retail-media-temporary-black-eye
WeltPixel Conversion Tracking uploads orders to your ad platforms as individual calls the moment the pipeline has its attribution data, so the gap you watch on the day is the platforms' documented processing, or Shopify's own delivery on the wildest days, never a batch queue on ours.