AdCopilot docsby Atromx
Troubleshooting

An audit returned partial results

Why a full_audit lists checks as not evaluated, what it means when it reached its time limit or timed out, and how to get a complete run.

View as Markdown
Open inClaudePerplexity

An audit that could not read everything says so. A check that did not get its data is listed as not evaluated, with the reason, and is left out of the health score on both sides. It is never reported as a pass. Read the coverage block before reading a result as a clean bill of health.

The audit reached its time limit

Since v2.15.1, a deep full_audit asked for with depth="deep" stops waiting for Google after 45 seconds and returns what finished. A call that names no depth also runs deep, but stops waiting at 27 seconds and returns what finished in the same way:

This audit reached its time limit before Google answered every read, so these sections are incomplete: … Every check that did not get its data is in the not-evaluated list; nothing was scored from missing data. Running the audit again is usually faster, because the Google Ads reads that finished are cached.

What to do:

  • Run it again. The Google Ads reads that finished are cached, so the second run usually completes. A Tag Manager or Analytics read the time limit cut is named on the check it cost, with the time limit as the reason; it does not mean that connection is broken.
  • Narrow it. Ask for a deep audit of one area at a time, for example sections=["wasted_spend"]. The six areas are conversion tracking, wasted spend, structure, keywords, ads and assets, and settings and targeting.

The audit timed out

If a tool gets nothing back within 60 seconds, it returns no result at all:

full_audit did not finish within 60 seconds because Google Ads was slow to answer, so it returned no result. Nothing was changed in the account.

Try once more; the cached reads make the second run faster. Or narrow it with sections=[…].

The audit could not be scored

What you see.

This audit could not be scored: N checks were not read on this run, so too little was measured to grade the account. The not-evaluated list gives the reason for each. This is not a verdict on your account — run the audit again once those reads succeed.

Why it happens. Google refused or did not answer the reads the checks needed. Each not-evaluated check marked UPSTREAM_ERROR carries what Google said, in its reason or its refused_reads.

Fix. Read the reasons before running it again:

  1. USER_PERMISSION_DENIED, CUSTOMER_NOT_ENABLED, NOT_ADS_USER or another access error: running it again gives the same result. Fix access first; see Google refused access. For USER_PERMISSION_DENIED on an account that is listed and switched on, start with USER_PERMISSION_DENIED.
  2. Google busy, or a shared limit reached: run it after the wait the message gives; see Busy, slow or rate-limited.
  3. The time limit: see The audit reached its time limit.
  4. Anything else: run it once more. If the same reads fail again, write to support@adcopilot.cloud with the reasons shown.

Other reasons checks are not evaluated

MessageMeaningWhat to do
"A quick audit runs N of the 75 checks and does not grade the account."A quick audit is three checks by design.Ask for a deep audit.
"The deep audit was not available this time, so this ran as a quick audit."A call that named no depth ran a deep audit that had returned nothing by 36 seconds, so the quick one answered. A deep audit asked for with depth="deep" never falls back, nor, on Free, one narrowed with sections to four or fewer areas.Ask for a deep audit again in a few minutes.
"The deep audit was not available this time, so this narrower audit did not run and counts nothing."On Free, an audit narrowed with sections to four or fewer areas runs deep or not at all, and the deep audit could not run this time.Ask for it again in a few minutes.
"This first audit ran N of the 75 checks …"The first audit is kept short.Ask for a deep audit when you want the score.
"No health score: only N of the checks that apply to this account could be evaluated (under half) …"The account lacks the data for most of the checks that apply to it, often because it is new or has little traffic. A check that does not apply, such as the Performance Max checks on an account with none, does not count against it.Nothing is wrong; the findings are still measured.
"On this account, N of the checks had too little or nothing to measure in this window …"Those checks found too little or nothing to measure. That can be no spend yet, too few impressions or keywords Google has not rated, or something not set up, such as no enabled primary conversion action. They are left out of the score, not counted against it.Read the not-evaluated list of a deep full_audit: each row there says what that check found missing. Spend, impressions and ratings come with time and traffic; a missing setup needs to be set up.
"This audit could not be scored: N checks were not read on this run …"Google refused or did not answer the reads this run needed. Each not-evaluated check gives the reason.Read the reasons first; see The audit could not be scored.
"This audit hit its query budget …"One audit has a ceiling on the reads it makes.Run it with sections narrowed.
"Google Ads requests were paused, so these findings come from cached data."Google Ads requests were paused, so no fresh reads could be made.Fresh audits resume when the pause ends. Try again later; see Busy, slow or rate-limited.
"Spend vs plan cap was not checked: no monthly cap is saved …"No monthly cap in your business profile.Tell AdCopilot your monthly cap.
"Competitor search terms were not checked …"No competitors saved. AdCopilot never guesses one.Tell AdCopilot your competitors, or add them under Competitors on the Business profile page.

A quick audit ignores sections; it always runs the same checks.

On this page