Home / Symptoms / All platforms

Attribution data delayed by a day or more

Attribution data that fills in a day or two late is normal. See the delays Google, Meta and Apple SKAN document, and how to stop acting on incomplete days.

SymptomInstalls and conversions from yesterday or today look low, then fill in over the next one to three days.
Most likely causeAd platforms and Apple add processing and privacy delays before conversions reach reports.
First checkCompare a date range that ended at least 72 hours ago and see whether the gap is still there.

Short answer: recent data is incomplete by design. Google says app conversions can take up to 24 hours to process, Apple adds at least 24 to 48 hours before a SKAN postback, and Meta says SKAN results are delayed by at least 24 hours. Judge campaigns on data at least two to three days old.

Why attribution data arrives late

Reporting delay: the time between an install or event and the moment it appears, attributed, in a report.

Some delay is processing. Some is deliberate: Apple adds a random delay to SKAN postbacks so that installs cannot be tied to individuals by timing.

How long should I wait before trusting the numbers?

Wait at least 48 to 72 hours for Android and web-to-app data, and longer for iOS SKAN conversion values. Later SKAN postbacks can take up to 144 hours after their window closes.

Likely causes

  • Google processing. Google says app conversion processing can take several hours, sometimes up to 24 hours. After linking GA4 or Firebase, allow 24 to 48 hours.
  • SKAN random delay. Apple says the minimum time from impression to postback is 24 to 48 hours. For SKAN 4, the first postback arrives 24 to 48 hours after its window ends and the second and third 24 to 144 hours after theirs.
  • Meta's SKAN delay. Meta's help centre lists SKAN delays of at least 24 hours among the reasons AEM and SKAN figures differ.
  • Invalid conversion removal. Google says some invalid conversions are removed later, so totals can drop as well as rise.

How to confirm it

  1. Pull the same report for the last 7 days and again two days later. If earlier days rose, you are seeing delay.
  2. Compare a range that ended at least 72 hours ago with the same range in the ad platform.
  3. For SKAN, check the dates your postbacks were received against the install dates.

How to fix it

  1. There is nothing to repair; set expectations instead.
  2. Grey out or label the last two days on dashboards as provisional.
  3. Make campaign decisions on data at least 72 hours old.
  4. For SKAN, judge conversion values on cohorts at least a week old, and longer for second and third postbacks.
SourceDocumented delay
Google app conversionsSeveral hours, up to 24 hours
Google after GA4 or Firebase linking24 to 48 hours to propagate
SKAN minimum, impression to postback24 to 48 hours
SKAN 4 second and third postbacks24 to 144 hours after window ends
Meta SKAN resultsAt least 24 hours

If late data is landing in organic first, see paid installs showing as organic. For SKAN timing in detail, read the SKAN 4 postbacks guide or second and third SKAN postbacks arriving late.

FAQ

Common questions

Why are yesterday's installs lower than usual?

Because conversions are still being processed and claimed. Google, Meta and Apple all document delays of a day or more before data is complete.

Can I speed up SKAN postbacks?

Not in production. Apple's testing profile shortens the delay to about 5 to 10 minutes on a test device, which helps with debugging but not with live reporting.

Why did last week's numbers go down?

Google says some invalid conversions are removed after the fact. Totals can fall slightly as that clean-up happens.

Should I pause campaigns based on today's data?

No. Today's numbers are the least complete. Wait at least two to three days, and longer for iOS SKAN results.

What is mobile attribution?

Mobile attribution is the process of crediting each app install or in-app event to the ad, link or channel that drove it. A mobile measurement partner (MMP) does this across all your channels so each install is counted once.

Why do ad networks report more installs than my MMP?

Each network counts every install it touched within its own windows, including views, while an MMP gives each install to a single source. Adding the networks' numbers together therefore overstates real installs.