Agent Analytics has always shown you how often AI systems cite your pages. What it could not tell you is whether those numbers are good. Is 400 citations a month strong for a site like yours, or far behind the pack? Benchmarking answers that question by placing every one of your pages inside the distribution of the whole Flowgen Network.
Key takeaways
- Benchmarking compares your AI citation performance against more than 800,000 pages in the Flowgen Network, with the network growing every day.
- Every page is classified into a tier, Poor, Fair, Good or Great, based on its percentile position; benchmarks refresh weekly.
- Three views: a Citation Health Overview, a Page Distribution across tiers, and a per-page percentile histogram.
- All comparisons use anonymised, aggregated data, and any segment with fewer than 20 sites is never shown.
What Benchmarking enables
Benchmarking evaluates your AI citation performance relative to more than 800,000 pages across the Flowgen Network, updated every week. Rather than a raw count, each page receives a tier, Poor, Fair, Good or Great, based on where it sits in the percentile distribution of pages that AI engines actually cite. Three dashboards make the data usable:
- Citation Health Overview shows the overall tier for your site, with the thresholds for the next tier visible, so you know exactly how far you are from "Good" or "Great".
- Page Distribution reveals the proportion of your pages in each tier, which immediately shows whether performance is concentrated in a handful of pages or spread across the site.
- Per-page percentile view plots each page against a histogram of the network, so you can see where an individual URL lands and what the pages above it look like.
How customers are using Benchmarking
The first thing teams do is calibrate. A citation count that looked flat suddenly has context: it is either comfortably in the Good tier, or it is Fair and the gap to Good is a specific number of citations away. Both are useful; only one was visible before.
The second use is prioritisation. Sorting cited pages by tier and reading the histogram makes under-performing content obvious. Teams pick the pages sitting just below a tier boundary and focus optimisation effort where a small improvement changes the classification.
The third is reporting upward. When an executive asks "how are we doing in AI compared to everyone else?", the answer is now a tier and a percentile rather than a feeling.
How we protect your data
Benchmarking runs exclusively on anonymised, de-identified data across the Flowgen Network. No individual site is ever exposed, and competitive comparisons are built from aggregated figures only. To protect participants further, any segment that would contain fewer than 20 sites is excluded from the product entirely, so a small industry or company-size cut can never be reverse-engineered into a specific competitor.
You see where you stand in the distribution. You never see who else is in it.
Get started
Benchmarking is available now to all eligible Agent Analytics customers in the Benchmarking tab of the dashboard. Participation is in your control: you can opt out at any time and have your data removed from the network. Read more on the Benchmarking feature page or talk to our team.