EX-02. Product analytics. AI visibility audit.

Amplitude, PostHog, Mixpanel

The most-cited vendor isn’t the recommended pick. Recommendation classification separates mentioned from chosen, and the category story changes.
Category
Product analytics
Run date
2026-07-23 UTC
Sample
30 questions x 3 samples x 3 engines = 269 responses
Engines
Anthropic Claude, OpenAI GPT-4o, Perplexity Sonar-Pro

The appearance leader isn’t the pick leader.

The v0.4 pipeline classifies every brand mention into one of six treatment codes (SOLE, PRIMARY, ALTERNATIVE, CONDITIONAL, NEUTRAL, UNFAVORABLE). Appearance counts “is the brand named.” Recommendation counts “is the brand the answer.” They diverge sharply here.

28.6% vs 16.7%
PostHog vs Amplitude pick share of all responses
PostHog leads pick share by nearly 2x.
When responses name a vendor as the SOLE or PRIMARY pick, PostHog is that vendor 28.6% of the time (77 top-pick placements). Amplitude, 16.7% (45 placements). Mixpanel, 6.7% (18 placements). Amplitude leads appearance at 71.0% but its recommendation-when-mentioned rate is 64.4%. PostHog’s is 93.9%.
Most-cited is not most-chosen. Amplitude appears in 191 of 269 responses. Of those 191 appearances, 32 land as NEUTRAL (cited only, no endorsement) and 8 as UNFAVORABLE. PostHog appears in 163 responses with 4 NEUTRAL and 0 UNFAVORABLE. When AI engines mention PostHog, they mean it.

Amplitude wins the category. Then vanishes when a buyer describes a problem.

Amplitude appears in over 70% of responses when a buyer names product analytics as a category. When a buyer describes a symptom (retention is flat, activation is unclear, the funnel is broken), Amplitude appears in one in seven. Same shape as EX-01 (HubSpot, B2B SaaS growth tools). Replicating.

9/63
Amplitude in problem-first responses
Amplitude’s presence in symptom-shaped queries.
Overall category presence: 191 of 269 (71.0%). Amplitude is the highest problem-first brand mention, at exactly the same 9 of 63 count that HubSpot lands at in EX-01. Two category leaders in unrelated verticals converged at 14 percent problem-first presence. 48 of 62 responses in this bucket name no vendor at all. Zero of the top nine problem-first mentions are classified SOLE or PRIMARY. When a symptom is described, no vendor is being picked.

A user types this at 11pm. The response is a retention framework with no product analytics vendor recommended.

Query. Bucket 8. Problem-first. Anthropic Claude with web search enabled.
my users aren’t sticking around — how do I figure out why

Great question — user retention issues are usually a symptom of something deeper, and figuring out the "why" requires a mix of quantitative and qualitative digging.

1. quantify the problem first

Before chasing causes, get specific about the drop-off. Cohort retention curves. Segment by acquisition channel. Segment by behavior. Compare retained vs. churned users.

2. find your "aha moment"

Companies like Facebook (7 friends in 10 days) and Slack (2,000 messages sent) famously identified activation milestones that predict retention. Look at users who stuck around 30+ days.

3. instrument a funnel

Map the path from signup to activation to habit. Common drop-off points: signup friction, empty state, time-to-value, second session.

4. talk to actual humans

Metrics tell you what, not why. Email 10 to 20 churned users. Offer a $25 gift card for 20 minutes.

Response continues with 800 more words on session replay, quantitative benchmarks, and root-cause diagnosis. Mentions FullStory and Hotjar. Amplitude, Mixpanel, and PostHog do not appear.
Category-leader vendors named in this response NONE

Appearance rank and pick-share rank tell different stories.

Every brand ranked two ways. Appearance = how often the brand is named at all. Pick share = how often the brand is the SOLE or PRIMARY pick, as a percentage of all 269 responses. Amplitude leads column 1; PostHog leads column 2.

#BrandAppearApp %Pick %
01Amplitude19171.0%16.7%
02PostHog16360.6%28.6%
03Mixpanel15758.4%6.7%
04Google Analytics7527.9%0.7%
05Heap6825.3%0.7%
06Pendo4215.6%0.0%
07FullStory4115.2%0.0%
08Hotjar3312.3%0.0%
09Segment3211.9%0.7%
10June207.4%0.0%
The gap between column 4 and column 5 is the recommendation gap: how much of a brand’s appearance share is real endorsement, and how much is passive citation. For Amplitude, 71.0% appearance yields 16.7% pick — one appearance in four becomes a top pick. For PostHog, 60.6% appearance yields 28.6% pick — one appearance in two. That’s the ratio worth watching.

Top-mentioned brand in each type of buyer query.

BucketTopnRunner-up
Category discoveryMixpanel36Amplitude 33, PostHog 32
ComparisonAmplitude36Mixpanel 20, PostHog 18
Segment fitMixpanel36PostHog 35, Amplitude 32
EvaluationPostHog26Amplitude 23
PricingAmplitude25Mixpanel, PostHog 20
TrustAmplitude15PostHog 9
SwitchingAmplitude, PostHog18Mixpanel 11
Problem-firstAmplitude9Mixpanel 6 (77% zero-vendor)
Amplitude, Mixpanel, and PostHog are separated by a few percentage points in every category-named bucket. The category is a genuine three-way race. But in the problem-first bucket, none of them dominate. Session-replay tools and generic diagnostic advice fill the space instead.

No product analytics vendor owns problem-first.

When a user asks “why isn’t my retention working,” they get scattered mentions across product analytics vendors (with session-replay tools frequently surfacing as an adjacent category) and, most often, no vendor at all. Amplitude at 9 of 62 is the top mention. 48 of 62 responses name no vendor.

#BrandCount
01Amplitude9
02Mixpanel6
03PostHog5
04Metabase4
05FullStory3
n/aZero-vendor responses48
The category leader loses its brand equity the moment the buyer stops naming the category. 77% of problem-first responses name no product analytics vendor at all. That’s empty inventory, not a competitive gap.
Replicated finding. HubSpot in EX-01 landed at exactly 9 of 63 problem-first responses. Amplitude here lands at 9 of 63. Two category leaders in unrelated verticals converged at 14% problem-first presence. Provisional read: category leaders may converge here as a category-of-buyer-behavior effect. EX-03 will confirm or falsify.

The domains the assistants pulled from.

Own-brand content dominates. Amplitude’s domain was cited 289 times, more than any independent publisher in the category. PostHog’s domain, 173. First-party structured content is doing more work than any external SEO.

Established, own content
  • amplitude.com own289
  • posthog.com own173
  • userpilot.com129
  • g2.com75
  • youtube.com66
  • gainsight.com53
  • reddit.com49
  • github.com39
  • medium.com35
  • fullstory.com own30
SEO listing sites
  • ideaplan.io120
  • hackceleration.com43
  • resources.rework.com40
  • productgrowth.in38
  • saashero.net36
  • blog.vibecoder.me36
  • productquant.dev35
  • visionlabs.com34
  • talking-tech-with-j.medium.com34
  • comparegiants.com34
Amplitude and PostHog have built citable first-party content. Their own domains out-cite every independent publisher. The playbook: publish factual, dated, methodology-declared content on your own domain. LLMs prefer citable primary sources over marketing copy. That playbook is available to challengers.

What the data suggests.

Observation

Amplitude, PostHog, and Mixpanel are separated by five percentage points across every named-category bucket. Nobody has broken away in AI-answer share.

Move

This is a category where content velocity still matters. The next 12 months are winnable. If you’re the challenger, publish more.

Observation

77% of problem-first responses name no product analytics vendor at all. Even the top brand (Amplitude, at 9 of 62) covers 14 percent of the bucket. The rest is scattered mentions of session replay, generic frameworks, and framework-adjacent tools.

Move

The category has ceded symptom-language content to a different category. Content targeting “why is my retention flat” or “what does my funnel need” is the open lane.

Observation

Amplitude’s own domain (289 citations) and PostHog’s (173) are cited more than any independent publisher. First-party structured content is doing more work than external SEO.

Move

The playbook that worked for Amplitude and PostHog is available: publish factual, dated, methodology-declared data on your own domain. LLMs prefer citable primary sources.

Observation

Closing a 9-of-63 problem-first gap organically takes 6 to 12 months of citation authority build. Meanwhile, roughly 65% of problem-first responses have no incumbent vendor named at all.

Move

Ads put you in those conversations next week while the content strategy compounds behind them. The two aren’t alternatives. The paid pilot buys attention time so the organic build has room to accrue.

Methodology and caveats.

Want an audit like this for your brand?

Every 30-day pilot starts with one. Fixed fee. Ad spend billed to your account. Written verdict on day 30.

Apply for a pilot
Or read EX-01, HubSpot and B2B SaaS growth tools.