AI Optimize reports

An AI Optimize report includes many of the same metrics you’ve already seen in traditional test and manual personalization reports. You’ll still see things like:

• Sessions
• Conversions
• Conversion rate
• Performance over time
• Traffic allocation
• Audience insights

The main difference is that AI-optimized optimizations add a metric called Strength.

[IMAGE: AI Optimize report with familiar metrics highlighted, plus Strength called out.]

Instead of waiting for one permanent winner, use the report to understand how each variation is performing while AI continues adapting delivery over time.

Read the core metrics first

Start with the same fundamentals you’d use in any optimization report.

• Sessions: How many browsing sessions were attributed to each variation.
• Conversions: How many times visitors completed the action tracked by the selected goal.
• Conversion rate: The number of conversions divided by the sessions associated with that variation.
• Users: The number of unique visitors who saw the variation.

[IMAGE: Results table with Base and several AI-optimized variations. Call out Sessions, Conversions, Conversion rate, and Users.]

If your optimization tracks multiple goals, the target goal appears by default, and you can switch goals to review other outcomes. You can also use filters to narrow the results by criteria like device type or new versus returning visitors.

These metrics tell you what has happened so far.

Strength helps you understand what that performance may mean.

What the Strength metric tells you

The Strength metric is the probability that a variation will continue to perform better or worse than the base based on historical results.

New variations start at 50% Strength. As Optimize gathers enough data, that value can move up or down.

Note: Strength won’t begin calculating until both the variation and the base have recorded at least 10 conversions. Until then, you may see the default 50% value even while traffic is coming in.

[IMAGE: Variation A — 82% Strength | Variation B — 53% | Variation C — 24%.]

In general:

• Higher Strength means the variation is more likely to continue outperforming the base.
• Lower Strength means the base is more likely to perform similarly or better.

Webflow suggests treating values around 70–100% as stronger-performing ideas worth leaning into, while values around 0–30% may be candidates to pause or rethink.

Strength considers more than the current conversion rate. Optimize also looks for:

• A meaningful difference in conversion rate
• Enough traffic to both the variation and the base
• A difference that persists over time

So a variation with a temporarily high conversion rate won’t necessarily have high Strength.

Read Strength alongside traffic

AI Optimize changes variation delivery as it learns, so you shouldn’t expect every variation to receive equal traffic.

The traffic allocation chart shows how much traffic each variation received over time, while the performance chart shows how conversion rates changed over the same period.

[IMAGE: Strength values beside a traffic allocation chart, showing traffic shifting across variations.]

Read those views together.

A variation may begin with a neutral Strength value, gather evidence, and then become stronger or weaker as visitor behavior changes. Its traffic allocation can change along with that learning.

That’s different from a traditional test, where you intentionally keep traffic allocation fixed while collecting evidence toward a result.

Look at performance over time

Strength isn’t permanent.

Visitor behavior can change, and conversion rates can fluctuate, which means a variation’s Strength can rise or fall over time.

[IMAGE: One variation’s Strength changing over several dates.]

Use the report to look for patterns rather than reacting to a single snapshot:

• Is a variation becoming consistently stronger?
• Is a previously strong idea losing momentum?
• Does a change line up with a campaign, promotion, or other event?
• Are audience insights showing that an idea resonates especially well with a particular group?

Audience insights can help surface which ideas are resonating with different visitor subsets, while the performance and traffic charts help you see how those results evolved over time.

Use what AI learns

An AI Optimize report doesn’t end with one permanent winner.

Instead, use stronger and weaker variations as clues about what your visitors respond to.

For example, if several high-Strength variations use benefit-focused messaging, that may be a signal worth exploring further. You could create new variations that build on that idea or carry the learning into a future traditional test. Webflow specifically recommends leaning into concepts from high-Strength variations and reconsidering low-Strength ones.

A useful read isn’t just: Which variation has the highest Strength? It’s: What is AI learning about the kinds of experiences that work for these visitors?

Ready to keep going?

You now know how to read the familiar metrics in an AI Optimize report, interpret Strength alongside traffic and performance over time, and use those signals to shape what you try next.

Next, you’ll zoom out from individual optimizations and learn how to monitor the impact of your optimization efforts across the whole site.