Data sources and limitations
An honest map of where Reputably looks, what it can and can't see on each platform, and why coverage is never total. Read this before you rely on any monitoring tool — ours included.
Reputably reads live Reddit threads, including comment content — not just post titles. That matters because most brand mentions and recommendation requests happen down in the comments, which keyword-only search tools miss. Coverage is strong for public subreddits. It does not include private subreddits, direct messages, or content removed by moderators before it's seen.
YouTube
Reputably monitors YouTube videos and Shorts, and deliberately balances the two so your feed isn't dominated by short-form clips. A video is classified as a Short by its actual duration. Coverage is based on public videos surfaced through search; private, unlisted, and age-restricted videos are not included, and a brand mentioned only in spoken audio with no supporting text may not always surface.
Reputably covers public Facebook groups — the neighbourhood and interest groups where locals ask for recommendations. It does not access private or closed groups, personal profiles, direct messages, or anything behind Facebook's privacy controls, and it never asks you to connect a Facebook account for monitoring. If a conversation is private, it's private; no tool should claim otherwise.
X (Twitter)
X coverage is limited and reads public posts only. Platform access on X has changed repeatedly, so treat it as a supplementary source rather than a primary one — Reddit, Facebook groups and YouTube carry most local buying conversation.
The open web
Beyond the platforms above, Reputably surfaces mentions from public web pages found through search, which catches blog posts, forums and news that name your business. This is index-based, so it reflects what's been published and indexed rather than live conversation.
AI answers, retrieval evidence and traffic
AI-visibility results are produced by asking ChatGPT, Gemini and Claude the buyer questions configured in your workspace and recording their answers. Those answers vary by prompt, model, location, session context and time. Use repeated results and trends rather than treating one run as a universal answer. See Set up AI visibility tracking.
Prompt coverage is not the same as fan-out searches
Tracked prompts are the buyer questions your team chooses to measure. Fan-out searches are extra queries an answer engine may create while researching one of those prompts. Reputably records fan-out searches only when the engine discloses them; it does not invent a trail where none was returned.
Retrieval disclosure varies by engine
Claude can provide a relatively complete, traceable search-and-page trail. ChatGPT may expose searches or results without enough detail to attribute every page to one exact rewrite. Gemini may return an answer without disclosing a retrieval trail at all. A blank retrieval view means not disclosed, not “zero searches”, and it does not prove the engine used no outside information.
Citations, surfaced sites and passed-over sites
A cited page is a page the engine explicitly attached to its answer. A surfaced site appeared in disclosed retrieval results. A passed-over site was surfaced but not cited in that run. These are useful diagnostic states, not proof of a model's private reasoning or a permanent preference.
AI crawler and referral traffic
AI Traffic requires the Reputably site tracker. Crawler records are classified by known bot identity and supporting network checks; referral visits depend on the browser sending a detectable referrer. Some apps, privacy controls and copied links suppress referrer data, so referral reporting is a lower bound. Crawler access, a later citation and a human visit are separate events: the product does not claim one caused another unless the evidence can support that link.
Why coverage is never total
No monitoring tool sees everything, and any that claims to is overstating. Platforms rate-limit and change access; private spaces are genuinely private; a mention with an unusual spelling or no searchable text can slip through; and syncs run on a schedule, so there's a natural lag between a conversation happening and it appearing. Reputably is built to catch the conversations that matter most — public, buying-intent, in the places customers actually talk — not to promise an impossible 100%.
Your data
Reputably stores what it needs to run monitoring, scoring and reporting for your account. You can request deletion at any time — see data deletion instructions — and read how we handle information in the privacy policy and on the security page. Questions? Email contact@reputably.net.
Someone is asking for a recommendation right now. Make sure it's you.
Start your 3-day free trial and see real conversations about what you sell — in minutes, no card, no accounts to connect.