Is a 62 good? Find out against your whole category

No. Benchmarks are computed from aggregated, anonymised figures and protected by k-anonymity, so no individual brand's numbers, prompts or answers are exposed to anyone else and nothing identifies you inside another customer's benchmark. You see a percentile and a rank for your category, not a table of named brands; named comparisons come from the competitors you choose to track, using your own data. Enterprise customers can also opt out of contributing to model training in their order form.

We match you to comparable brands based on your business and the category your prompts describe, and you can refine that peer set if the automatic match is too broad or too narrow. This matters more than it sounds. A percentile against every software company means close to nothing, while a percentile against the brands that actually appear in the same AI answers as you is a target you can work towards. Refining the peer set re-computes your rank.

Enough to satisfy the k-anonymity threshold, and no benchmark is shown at all where the category is too thin to protect the brands in it. That is a deliberate trade: in a narrow niche you may see no benchmark rather than a fragile one derived from three brands. Where that happens, tracked competitors and share of voice give you the comparison instead, using brands you have named rather than an anonymised field.

It turns a score that means nothing on its own into a position in your category. A 62 could be excellent or dismal depending on the field, and the percentile says which. It is most useful as a target and a direction of travel, because a raw score can rise while everyone around you rises faster. Read it over quarters rather than days: the field moves slowly, and so should your interpretation of it.

Yes. Category benchmarks are included on Starter, Professional and Enterprise at no extra cost, as are share of voice and competitor tracking. What changes by tier is the number of competitors you can name, five, ten or custom, along with engine coverage and prompt volume. The benchmark is derived from the tracking you already run, so there is nothing separate to configure and it moves with the work you do.

Yes. Benchmarks appear in the shareable reports, which are white-label on every plan and sent as tokenised links needing no AI Rankly login, so a client or an exec sees your branding and a percentile in context rather than a raw number they have to interpret. Reports can be scheduled, so the quarterly comparison assembles itself rather than becoming somebody's recurring calendar entry.