Competitors — how we choose them and what 'competitor mention' means
How your competitor list is picked, how you can override it, and how competitor visibility is measured on the same axis as yours.
Competitors — how we choose them and what "competitor mention" means
TL;DR — We auto-pick a first competitor list from same-category signals in AI answers themselves. You can override any of it. Competitor visibility is scored on the exact same axis as yours, so numbers are directly comparable.
How the initial list is built
At onboarding, we run a first-pass scan across your generated prompts and log every distinct brand the engines named. The brands that appear most frequently — filtered by category fit and de-duplicated — become your starting competitor set.
This is deliberately empirical: your competitors aren't who you think they are, they're who the AI engines think they are. The two lists are often different, and that difference is a strategy signal in itself.
The typical count
Most projects run 5–10 competitors. Fewer than that misses the shape of the landscape; more than that dilutes the signal (and the dashboard).
If your industry has a very long tail (agencies, local services), we lean toward the top 10 by mention share. If it's concentrated (enterprise SaaS with three real players), we lean toward the top 5.
Editing the list
Every competitor can be:
- Added manually — by domain or brand name
- Removed — with history preserved if you re-add later
- Renamed — if we captured a display name you don't like
Bring your gut. If you know a competitor matters even though the engines haven't caught up yet, add them.
How "competitor mention" is measured
Same formula as your own visibility:
competitor_visibility = (competitor_mentions ÷ responses) × 100
That means one bar on the competitor chart is directly comparable to your bar. If your score on a prompt is 22% and Competitor A is 44%, the AI names them about twice as often as you.
De-duplication rules
- Brand name variations collapse to one entity (
OpenAIandOpen AIare the same competitor) - Parent/child brands are separated by default (
MetaandInstagramare distinct) - Domain-level detection catches white-labelled mentions that use a different name
Reading competitor data
Three questions the tab is designed to answer:
- Who's rising? Look at the 30-day trend, not the absolute score. A competitor gaining 4 points a week is a bigger threat than one sitting at 40%.
- Where are they winning that I'm not? The per-prompt overlay shows exactly which prompts each competitor owns. Those become priority gap prompts.
- What sources are they citing? The Sources tab, filtered by competitor domain, shows the specific pages of theirs that AI engines love — great input for your own content roadmap.
What we won't do
We don't publicly compare you against competitors on our marketing site. Everything competitor-related is inside your dashboard, private to your account.