PostHog wins the AI answer. Its own site doesn’t get the credit.

What we did

PostHog isn’t a Nivonto client — we picked it because it’s a well-known, well-documented product with genuinely strong content, which makes it a fair test case. On 14 August 2026 we asked ChatGPT, Perplexity and Google three versions of a real buyer question — “best product analytics tools for startups 2026” and “PostHog vs Mixpanel” — from guest or default sessions, with no prompt engineering. Then we checked posthog.com directly: its robots.txt, its llms.txt, and whether its pages carry any structured data at all.

Finding 1: PostHog is already recommended — that part isn’t the problem

All three engines named PostHog, unprompted, as a top pick for startups. That’s the outcome most companies are chasing when they ask for an AEO audit, and PostHog already has it. Strong comparison pages, an honest pricing story, and a well-known open-source brand are doing real work here. The problem isn’t visibility. It’s who gets credit for it.

Finding 2: the citation is going to strangers, not posthog.com

For the broad discovery query — the one a buyer types before they’ve heard of any specific tool — Perplexity cited 10 sources. Every single one was a third-party listicle. Not one was posthog.com.

Google AI Overview did the same thing: it named PostHog, then linked the claim to trackraptor.com — a site PostHog has no control over, no relationship with, and no way to correct if it gets a detail wrong. This is the exact dynamic behind Nivonto’s own research into the “AEO for AEO” pattern: a small set of listicle domains sit between well-known products and the AI engines that cite them, quietly capturing the authority those products already earned.

Finding 3: head-to-head queries are the exception — and the proof

It isn’t hopeless. When we asked the direct comparison question — “PostHog vs Mixpanel” — Perplexity’s answer cited posthog.com itself among its sources. PostHog’s own comparison pages are good enough to win the citation when the query is specific enough to surface them. The gap is entirely in the broader, earlier-stage discovery queries, where third-party content still beats PostHog’s own site to the punch.

Finding 4: the technical foundation isn’t helping

We checked posthog.com directly rather than taking anyone’s word for it.

What this means

PostHog’s AEO problem isn’t “get mentioned by AI” — it already is. It’s “make sure the citation lands on a page you control.” That’s a narrower, cheaper fix than most companies need: structured data (Organization + SoftwareApplication + FAQPage) so engines can lift facts directly from posthog.com instead of a summary of a summary; more of the direct comparison-style content that already wins citations, aimed at the broader discovery queries it currently loses; and treating llms.txt as what it is — a developer-integration doc, not an AEO deliverable, so buyer-facing content has to live somewhere engines actually cite from.

Methodology & honesty note

This is an independent audit, not a case study — PostHog has never engaged Nivonto, and nothing here implies otherwise. All quoted answers are verbatim from live sessions on 14 August 2026; nothing was cherry-picked or re-rolled for a better result. AI answers are not static — the same prompts run today may return different names, different sources, or a different ranking entirely. The value here is the pattern (strong brand, weak citation ownership, no structured data) more than any single number, and it’s the same methodology we run for every Audit Sprint client.