Experiments

The Lab is where Sill measures whether changes to your content actually improve your AI visibility. Instead of guessing whether a content update helped, you get measured results.

How Experiments Work

Sill continuously monitors your website for content changes. When it detects a meaningful change — a new page, updated copy, added structured data, or a revised product description — it automatically proposes an experiment.

Each experiment tracks:

1. What changed on your site
2. Your AI visibility before the change (baseline)
3. Your AI visibility after the change (measured over subsequent monitoring cycles)
4. The measured impact — did visibility go up, down, or stay flat?

Experiment Lifecycle

Experiments move through three stages:

Proposed

When Sill detects a content change, it creates a proposed experiment. You can review the detected change and confirm it to start tracking, or dismiss it if the change isn't meaningful (for example, a typo fix).

Active

Once confirmed, the experiment is active. Sill measures your AI visibility across monitoring cycles to determine the impact. Active experiments need time to collect enough data — the more cycles, the more reliable the measurement.

Completed

After enough data has been collected, the experiment completes with a measured result showing the change in your AI visibility metrics. You can see whether the change helped, hurt, or had no measurable effect.

The Visibility Timeline

The Lab includes a 90-day visibility trend that overlays your experiments on a timeline. This lets you correlate visibility changes with specific content updates and see how your overall trajectory is evolving.

Events and experiments are plotted on the same timeline, making it easy to spot cause-and-effect relationships.

How Experiments Bundle Changes

Unlike traditional A/B testing, AI visibility doesn't respond to isolated tweaks. Unless your brand has strong topical authority and your pages are crawled frequently, a single content change is unlikely to move the needle on its own.

Sill accounts for this by bundling multiple related changes into a single experiment. For example, an experiment might include adding schema markup to your pricing page, updating your product description, and publishing a new comparison article — all tracked together as one coordinated effort.

This gives you a realistic measurement of impact. When the experiment completes, you'll see the combined effect of everything you changed during that period, which is how AI visibility actually shifts in practice.