Most AI Max accounts have the same 5-10 problems. Here's a systematic checklist to find them before they cost you budget.
Conversion Tracking Audit (Critical)
1. Are conversions firing correctly?
Navigate to Tools > Conversions. Check "Status" for each conversion action — look for "Active" with recent data. "No recent conversions" may indicate a broken tag, not no conversions. Use Google Tag Assistant to verify the tag is firing on thank-you pages.
2. Is Enhanced Conversions enabled?
In each conversion action settings, check if "Enhanced conversions" is enabled. For any conversion action with fewer than expected recordings (especially from mobile), enhanced conversions improves accuracy. Go to Conversions > Settings > Enhanced Conversions and enable at the account level.
3. Are you tracking micro-conversions?
For accounts under 50 conversions/month, AI Max can't learn from macro-conversions alone. Add at least 2-3 micro-conversion events: pricing page visit, demo booking form opened, time-on-site thresholds. Set these as "Observation only" so they inform the model without distorting your primary optimization goal.
4. Do conversion values reflect actual business impact?
If you're running target ROAS, are you passing actual transaction values with each conversion? Open a recent conversion in the Conversions table and verify the conversion value is populated, not zero. Zero values break target ROAS bidding.
Asset Quality Audit
5. Are any assets rated "Low" for more than 30 days?
In each asset group, check asset performance ratings. Remove assets that have been consistently "Low" — they're dragging performance.
6. Do you have minimum recommended asset counts?
Each asset group should have: 15 headlines, 4 descriptions, 1 landscape image, 1 square image, and at least 1 video. Accounts below these minimums have less creative diversity for AI Max to optimize.
7. Are auto-generated assets enabled and have you reviewed them?
In campaign settings, check "Asset automation." If enabled, go to the Asset Library and filter for "Auto-generated" assets. Review each one — AI Max generates creative that may not match your brand voice or may make claims you haven't approved. The July 2026 ToS (https://yositeup.com/blog/google-ads-tos-july-2026-ai-automation-what-changed) requires review within 30 days of first serving.
8. Is there asset relevance between groups?
Asset group 1 should have assets specific to topic 1; asset group 2 to topic 2. If all asset groups are using the same generic headline pool, you're not using asset groups correctly — AI Max can't differentiate intent signals.
Bidding Audit
9. Is bidding strategy appropriate for conversion volume?
Target CPA requires 50+ conversions/month. Target ROAS requires 50+ conversions. Below these thresholds, switch to "Maximize conversions" without a target. Having a tCPA target with 20 monthly conversions is a learning-mode trap — the campaign never stabilizes.
10. Is your target CPA/ROAS set above your current actual performance?
This is the most common mistake. If your current CPL is $75, don't set tCPA at $60. Set it at $75-90 initially. Underbidding your actual performance starves the campaign of impressions. The model needs volume to improve.
11. Is smart bidding portfolio configured correctly?
If using a portfolio bidding strategy, audit which campaigns are in the portfolio. Campaigns with very different conversion qualities (high-ticket B2B + low-ticket e-commerce) in the same portfolio dilute the model.
URL and Landing Page Audit
12. Is URL expansion enabled, and have you audited the expanded URLs?
In campaign settings, check URL expansion status. If enabled, review which URLs AI Max has expanded to in the Search Terms report and URL insights. Remove any expanded URLs that lead to irrelevant pages.
13. Do landing pages match asset group intent?
Navigate to each asset group's final URL. Does the landing page content match the headline themes in that asset group? Poor message match is a major conversion rate killer that AI Max cannot compensate for.
14. Have you reviewed landing page quality scores?
In the old keyword table (if any keywords are present), check Landing Page Experience column. But even without keywords, check Google PageSpeed Insights for your main landing pages. Core Web Vitals below "Good" thresholds directly reduce expected CVR.
Audience Signal Audit
15. Are Customer Match lists active and recent?
Customer Match lists older than 90 days may have significant decay (email addresses change). Refresh your Customer Match upload monthly. Lists that haven't been updated in 6+ months should be re-uploaded before relying on them as AI Max signals.
16. Are remarketing tags firing?
Check your remarketing audience sizes. If "All website visitors" has fewer users than your monthly sessions, the tag is missing pageviews. Common cause: tag not firing on SPA (Single Page Application) page changes.
17. Do you have both positive and negative audience signals?
Positive signals: your best customers, high-intent visitors, in-market audiences. Negative exclusions: competitor employees, existing customers (from prospecting campaigns), unqualified segments. Both types improve AI Max efficiency.
Reporting and Attribution Audit
18. Are you using data-driven attribution?
In Conversions settings, check the attribution model for each conversion action. If using last-click for cross-channel AI Max campaigns, switch to data-driven attribution if eligible (300+ conversions, 3,000+ interactions/month). Last-click undervalues upper-funnel AI Max placements.
19. Is GA4 connected and are GA4 audiences imported?
In Linked Accounts, verify GA4 is linked. In Audience Manager, check for GA4 audiences. If you're not using GA4 behavioral audiences as AI Max signals, you're leaving significant signal quality on the table.
20. Have you adjusted for the June 2026 data deletion?
The June 2026 reporting deletion (https://yositeup.com/blog/google-ads-reporting-data-deleted-june-2026) removed historical data that many accounts used for benchmarking. Verify your tCPA and tROAS targets are based on post-deletion data. Pre-deletion CPL benchmarks may no longer be accurate reference points.
Additionally, check that the DSA migration (https://yositeup.com/blog/google-ads-dsa-ai-max-migration-february-2027) didn't break any conversion tags or audience configurations. DSA-to-AI Max migrations have introduced tracking regressions in approximately 20% of migrated accounts.
When You've Fixed the Basics
After completing this audit and fixing identified issues, wait 30 days before drawing conclusions. AI Max's learning phase means changes don't produce immediate results — the model needs time to adapt to new signals, conversion data, and bidding parameters.
After 30 days, re-audit against these same criteria. Most accounts need 2-3 audit-fix cycles before reaching stable efficiency. The goal is not perfection — it's systematic improvement that compounds over time.
Top comments (0)