AI Max Testing and Implementation Guide: A Practical Framework

The Core Components

AI Max is not revolutionary new technology. It packages existing Google Ads features into one convenient toggle:

1. Text Customization (Automatically Created Assets)

• Google creates headlines and descriptions using landing page content
• Previously available as a separate feature
• Now bundled into the AI Max setting

2. Final URL Expansion (DSA Technology)

• Allows Google to send traffic to different landing pages
• Based on what the algorithm believes will perform better
• Existing dynamic search ad technology rebranded

3. Keyword Matching (Broad Match Technology)

• Expands targeting beyond exact and phrase match
• Uses intent signals to find related searches
• Forces campaigns into broad match behavior

4. Location of Interest Targeting

• Enables ad group level location control
• Useful for targeting visitors searching for local services
• Example: Targeting US users searching for Los Angeles restaurants

5. Brand Settings

• Include or exclude specific brands
• Now available at ad group level
• Useful for brand protection strategies

Best Candidates for AI Max Testing

Account Profile	Expected Outcome
Large accounts already using broad match	Modest improvement (5-15%)
Accounts with exact/phrase only	Bigger jump (15-27%) but more volatility
Small accounts with limited data	Often struggles—insufficient learning signals

Who Should Wait

• New accounts (<100 conversions): Not enough data for AI to learn
• Tight ROAS requirements: AI Max may overspend during learning
• Single product/service focus: URL expansion may be counterproductive
• Heavily regulated industries: Auto-generated assets may cause compliance issues

The Early Testing Reality

Google's published results (14-27% improvement) come from enterprise advertisers with massive data volumes. Smaller accounts often see:

• 2-4 week learning period with volatile performance
• Increased spend without proportional conversion increase
• Difficulty diagnosing issues due to automation opacity

Phase 1: Pre-Test Setup (Week 0)

1. Document baseline metrics: CPA, ROAS, conversion rate, CTR for last 30 days
2. Verify conversion tracking: Ensure accurate data before testing
3. Set clear success criteria: What improvement justifies adoption?
4. Identify test campaign: Choose one with sufficient volume (50+ conversions/month)

Phase 2: Controlled Rollout (Weeks 1-2)

Day	Action
Day 1	Enable AI Max on single campaign
Day 2-3	Monitor for dramatic changes; disable if CPA spikes 50%+
Day 7	First performance review—compare to baseline
Day 14	Second review—assess learning phase completion

Phase 3: Evaluation (Weeks 3-4)

• Compare 2-week AI Max performance to 2-week pre-AI Max baseline
• Analyze search terms report for targeting quality
• Review auto-generated assets for brand compliance
• Calculate true incremental value (not just reported metrics)

Phase 4: Decision (Week 5)

Result	Action
Clear improvement (>15%)	Expand to additional campaigns
Marginal improvement (5-15%)	Continue testing another 2 weeks
No improvement or decline	Disable AI Max, revert to baseline

How to See AI Max Activity

Important: AI Max does not create new ad groups or visible keyword changes. It operates in the background.

To see AI Max activity:

1. Go to Search Terms report
2. Click the dropdown filter
3. Select "Search term match type"
4. Look for terms triggered by AI Max expansion

Key Metrics to Track

Metric	What to Watch
Search term relevance	Are new terms matching user intent?
CPA trend	Is cost per conversion stable or increasing?
Conversion rate	Are expanded audiences converting?
Asset performance	Are auto-generated assets performing well?

Warning Signs

• CPA spike >30%: Learning phase may be too aggressive
• Irrelevant search terms: AI is over-expanding targeting
• Asset disapprovals: Auto-generated content violating policies
• Impression share drop: Budget being spread too thin

Selective Feature Activation

You don't have to enable all AI Max features. Configure based on your needs:

Text Customization

• Enable if: You want Google to test more headline/description variations
• Disable if: Brand messaging must be strictly controlled

Final URL Expansion

• Enable if: You have multiple relevant landing pages
• Disable if: Single product/service with dedicated landing page

Keyword Matching Expansion

• Enable if: Already comfortable with broad match performance
• Disable if: Tight keyword control is required

Recommended Starting Configuration

Feature	Initial Setting	Rationale
Text Customization	OFF	Test one thing at a time
Final URL Expansion	OFF	Keep landing page control
Keyword Matching	ON	Primary AI Max benefit

Strategy: Enable features one at a time, measuring impact of each.

Pre-Test Checklist

• ☐ Baseline metrics documented (last 30 days)
• ☐ Conversion tracking verified
• ☐ Success criteria defined (target improvement %)
• ☐ Test campaign selected (50+ conversions/month)
• ☐ Budget allocated for potential learning phase volatility

During Test Checklist

• ☐ Daily monitoring for first 3 days
• ☐ Search terms reviewed at Day 7
• ☐ Auto-generated assets checked for compliance
• ☐ CPA/ROAS tracked against baseline
• ☐ Week 2 performance review completed

Post-Test Checklist

• ☐ 4-week performance summary created
• ☐ Comparison to baseline documented
• ☐ Search term quality assessed
• ☐ Decision made: expand, continue testing, or disable
• ☐ Learnings documented for future campaigns