Fathom Analytics
website analytics · Sep 5, 2026
The short version
This scan shows an emerging presence — visible, but not yet consistent.
10 questions · 3 engines · 3 samples · Sep 5, 2026
What the answers show
Competing answers leaned on self-hosted control/compliance and SaaS cost trade-offs.
from Gemini (grounded) · “What are the trade-offs between a cloud-hosted analytics service and a self-hosted one?” · Open in Questions — “What are the trade-offs between a cloud-hosted analytics service and a self-hosted one?”
Fathom messaging: cookieless analytics, minimal personal data (IP+UA), GDPR/CCPA framing—some statements appear inconsistent.
from ChatGPT (API-grounded proxy) · “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?” and 1 more · Open in Questions — “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?”
Where you lose the slot
On an unaided category question — “What are the trade-offs between a cloud-hosted analytics service and a self-hosted one?” — Fathom Analytics is not named by 3 of the engines that answered and the slot goes to Google Analytics, Plausible, Matomo, Mixpanel, Umami.
Read the answers behind this →Start here
- 1usefathom.com — already working in your favor — keep this page fresh and linkable
- 2github.com — developer-trusted territory — answer the questions in your problem space with working code and a plain statement of what your product does, without pitching
- 3matomo.org — a third-party source AI models trust — get listed, reviewed, or cited there
AI visibility, at a glance
Measured against: Google Analytics, Plausible, Matomo, Mixpanel, Umami
AI Visibility
43%
range 43–43% across samples
% of collected answers that mention or cite you at least once — questions naming you excluded. What this means
Share of Voice
19%
range 19–20% across samples
First-mention rate
10%
of answers where you appear first, ahead of the named competitors
Share of voice vs your competitors
Wider ranges mean less stable answers — we show you the spread across samples, not a false-precision single number.
You vs each competitor
Click a column to sort. Share of voice shows the observed range across samples — never a false-precision single number.
| Fathom AnalyticsYou | 19%(19–20%) | 60% | 30% | 1.6 | 165 |
| Google Analytics | 24%(23–24%) | 51% | 24% | 1.7 | 0 |
| Plausible | 17%(17–17%) | 30% | 11% | 2.1 | 0 |
| Matomo | 11%(10–11%) | 20% | 10% | 2.3 | 0 |
| Mixpanel | 13%(11–15%) | 20% | 1% | 2.6 | 0 |
| Umami | 16%(15–17%) | 29% | 0% | 3.0 | 0 |
“Avg. position” is the brand’s mean rank among tracked names in answers where it is named; “—” means never named.
Engine by engine
Where AI recommends you — and where a competitor holds your slot.
| Top competitor | ||||
|---|---|---|---|---|
| Gemini (grounded) | 30/30 | 70% | 30% | Google Analytics (93%) |
| ChatGPT (API-grounded proxy) | 30/30 | 60% | 40% | Google Analytics (30%) |
| Perplexity Sonar | 30/30 | 50% | 20% | Google Analytics (30%) |
Are you in the shortlist — and who took your slot.
Every question below asks about website analytics.
Every ✓ / ✗ is clickable — read the full captured AI answers behind each cell, with the sources they cited.
Strong = mentioned consistently and engines agree · Noisy = engines disagree between runs · Emerging = mentioned, but not consistently · Not mentioned = never appears.
| Prompt | Perplexity Sonar | ChatGPT (API-grounded proxy) | Gemini (grounded) | Signal |
|---|---|---|---|---|
| “What features should I prioritize if I want minimal page-speed impact from an analytics script?” Present instead: Google Analytics, Plausible, Umami | Not mentioned | |||
| “What are the trade-offs between a cloud-hosted analytics service and a self-hosted one?” Present instead: Google Analytics, Plausible, Matomo, Mixpanel, Umami | Not mentioned | |||
| “Can I use a simple analytics tool for tracking conversions and events, or do I need a more advanced product?” | Noisy | |||
| “For a privacy-first, lightweight setup, what are the differences between Fathom, Plausible and Umami in terms of script size, features, and ease of use?” | Noisy | |||
| “How do privacy-focused tools differ from traditional analytics platforms?” | Emerging | |||
| “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?” | Strong | |||
| “Can I self-host Fathom Analytics, and how does that change pricing and maintenance?” | Strong | |||
| “How easy is it to migrate from Google Analytics to Fathom Analytics without losing conversion tracking?” | Strong | |||
| “How does Fathom Analytics’ data accuracy and sampling compare to Google Analytics and Mixpanel for event-heavy sites?” | Strong | |||
| “If I need self-hosting and full data control, how does Fathom compare to Matomo in terms of maintenance burden and available integrations?” | Strong |
All 10 prompts sampled in this scan are shown above.
Analysis
Everything in this block is the model's reading of the answers this scan stored — not a measurement and not a fact-check. Every statement carries a verbatim quote from an answer you can open; statements without one were dropped (see Methodology).
Why you lose the slot
For the questions where no engine named you, what the answers said instead — the model's reading of the stored answers, each point tied to a quote you can open.
1 “What are the trade-offs between a cloud-hosted analytics service and a self-hosted one?”
Who took your slot: Google Analytics, Plausible, Matomo, Mixpanel, Umami · on Gemini (grounded)
Competitors detailed self-hosted control/compliance and cloud cost trade-offs.
“As your traffic scales, SaaS pricing can increase dramatically (often tier-based on pageviews or monthly active users).”
from Gemini (grounded) · “What are the trade-offs between a cloud-hosted analytics service and a self-hosted one?” · Open in Questions — “What are the trade-offs between a cloud-hosted analytics service and a self-hosted one?”
What AI says about you
What the engines state about you when a question names you — paraphrased by the model from the stored answers, never checked against reality. A Verify chip marks a specific claim (a price, date, feature or availability) worth checking yourself.
- PositiveVerify
Fathom is cookieless for analytics.
Check: Verify docs state no analytics cookies.
“It does not use analytics cookies.”
from ChatGPT (API-grounded proxy) · “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?” · Open in Questions — “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?” - NeutralVerify
Fathom processes IP address and User-Agent as personal data.
Check: Confirm current data-processing docs mention IP and User-Agent.
“Yes — Fathom does process a small amount of personal data (IP address + User‑Agent).”
from ChatGPT (API-grounded proxy) · “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?” · Open in Questions — “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?” - PositiveVerify
Fathom claims GDPR/CCPA alignment and provides a DPA.
Check: Check current DPA and compliance pages.
“Fathom publishes a Data Processing Agreement (DPA) (last updated June 20, 2025) and describes technical measures such as EU‑isolation and one‑way hashing of EU IPs before they leave EU servers.”
from ChatGPT (API-grounded proxy) · “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?” · Open in Questions — “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?” - NeutralVerify
Fathom describes hashing a per-site user signature (IP+UA+hostname+salt) and rotating salts.
Check: Verify hashing inputs and salt-rotation details in tech docs.
“Instead of storing raw IPs/UA with page data, Fathom computes a SHA‑256 “user signature” (IP + User‑Agent + hostname + site salt) and stores the hash to count uniques and visits.”
from ChatGPT (API-grounded proxy) · “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?” · Open in Questions — “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?”
Citation gap
URLs the engines cited in answers naming a competitor and never in an answer naming you — counted by code from the stored sources, not by the model. Top 8, at least two competitor citations each.
| URL | Domain | Ownership | Cited with competitors |
|---|---|---|---|
| https://aaronjbecker.com/posts/umami-vs-plausible-vs-matomo-self-hosted-analytics/ | aaronjbecker.com | Third-party | 3× |
| https://apiscout.dev/guides/plausible-vs-umami-vs-fathom-analytics-2026 | apiscout.dev | Third-party | 3× |
| https://bootstrap.build/articles/best-web-analytics-tools/ | bootstrap.build | Third-party | 3× |
| https://dev.to/alanwest/migrating-off-google-analytics-umami-vs-plausible-vs-fathom-1bnj | dev.to | Third-party | 3× |
| https://dev.to/alanwest/migrating-off-google-analytics-umami-vs-plausible-vs-fathom-36gc | dev.to | Third-party | 3× |
| https://gitnux.org/best/event-analytics-software/ | gitnux.org | Third-party | 3× |
| https://homelabaddiction.com/best-self-hosted-analytics-tools-for-homelabs-in-2026-plausible-vs-umami-vs-matomo/ | homelabaddiction.com | Third-party | 3× |
| https://mapmyvisitors.com/b/analytics-script-size | mapmyvisitors.com | Third-party | 3× |
Counted from the sources each answer returned; a URL counts once per answer.
Action plan
Tickets the model derived from the evidence above. Each names the question it targets, who owns it and the effect to expect — a starting point, not a forecast.
Effort: S = under a day · M = days · L = weeks or more.
- 1
Clarify data handling (IP/UA) vs “no personal data” marketing
Owner: PR · Effort: S
Expected effect: Reduce confusion; align claims across pages; improve trust and conversion.
“Yes — Fathom does process a small amount of personal data (IP address + User‑Agent).”
from ChatGPT (API-grounded proxy) · “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?” · Open in Questions — “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?” - 2
Publish technical paper on hashing, salt rotation, and EU isolation
Owner: Dev · Effort: M
Expected effect: Answer legal/technical due‑diligence questions; support enterprise adoption.
“Instead of storing raw IPs/UA with page data, Fathom computes a SHA‑256 “user signature” (IP + User‑Agent + hostname + site salt) and stores the hash to count uniques and visits.”
from ChatGPT (API-grounded proxy) · “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?” · Open in Questions — “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?” “Fathom publishes a Data Processing Agreement (DPA) (last updated June 20, 2025) and describes technical measures such as EU‑isolation and one‑way hashing of EU IPs before they leave EU servers.”
from ChatGPT (API-grounded proxy) · “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?” · Open in Questions — “Does Fathom Analytics collect any personal data, and is it compliant with GDPR/CCPA?”
What to fix
- usefathom.com — cited 165× across your promptsalready working in your favor — keep this page fresh and linkableYours
Seen on: “Can I use a simple analytics tool for tracking conversions and events, or do I need a more advanced product?” and 164 more
- github.com — cited 60× across your promptsdeveloper-trusted territory — answer the questions in your problem space with working code and a plain statement of what your product does, without pitchingThird-party
Seen on: “If I need self-hosting and full data control, how does Fathom compare to Matomo in terms of maintenance burden and available integrations?” and 59 more
- matomo.org — cited 46× across your promptsa third-party source AI models trust — get listed, reviewed, or cited thereThird-party
Seen on: “If I need self-hosting and full data control, how does Fathom compare to Matomo in terms of maintenance burden and available integrations?” and 45 more
- plausible.io — cited 32× across your promptsa third-party source AI models trust — get listed, reviewed, or cited thereThird-party
Seen on: “Can I use a simple analytics tool for tracking conversions and events, or do I need a more advanced product?” and 31 more
- analytics-alternatives.com — cited 19× across your promptsa third-party source AI models trust — get listed, reviewed, or cited thereThird-party
Seen on: “What are the trade-offs between a cloud-hosted analytics service and a self-hosted one?” and 18 more
Scored by prompt type
General category prompts
Unprompted category questions — do you come up when nobody names you?
- Visibility
- 8%
- First-mention rate
- 0%
- Coverage
- 100%
% of planned engine calls that returned an answer.
Branded prompts
The model is asked about you by name here — presence is expected and is NOT a visibility signal. This lens reads sentiment and factual accuracy, not top-of-mind recall.
- Brand-fact presence
- 100%
- First-mention rate
- 78%
- Coverage
- 100%
Share of answers to questions naming you that actually discuss you — not a fact-check of what they say.
% of planned engine calls that returned an answer.
How AI portrays you
44% positive · 44% neutral · 11% negative
Judged on 27 answers that mention you — the model's tone toward you, not a fact-check of what it says.
Comparative prompts
Head-to-head questions naming you against your competitors.
- Visibility
- 89%
- First-mention rate
- 22%
- Coverage
- 100%
% of planned engine calls that returned an answer.
Prompt coverage by stage
| Stage | Prompts sampled | Your visibility | Top competitor (share of voice) |
|---|---|---|---|
| Awareness | 2 | 6% | Google Analytics 23% |
| Consideration | 5 | 41% | Google Analytics 32% |
| Decision | 3 | 83% | Umami 22% |
Methodology
This is a sample, not a census: we ran 10 questions across each engine 3 times and show you the range, so you see both where you stand and how stable that standing is.
- Perplexity Sonar — sonar (best-effort — echoed by the provider, not a dated snapshot)
- ChatGPT (API-grounded proxy) — gpt-5-mini-2025-08-07
- Gemini (grounded) — gemini-3.5-flash
Analysis — gpt-5 (gpt-5-2025-08-07), Sep 6, 2026. The model read the stored answers and explained them; it added no facts and verified none. 10 statements dropped for lacking a verbatim quote.
What the engine labels mean
- API-grounded proxy — consumer ChatGPT and Claude.ai can't be queried programmatically, so we call the same underlying OpenAI/Claude models with web search enabled — the closest measurable proxy.
- Grounded — the model answers with live web search enabled, the same way its consumer product does.
- Pinned — each engine runs one default model for every customer so results stay comparable; a premium model, if you opted into one, is listed on its own line above.
- Best-effort version — Perplexity and Grok echo back the model name we requested instead of resolving one, so their version line records what we asked for, not what answered. ChatGPT, Gemini and Claude each return their own dated snapshot.
Where you stand today
This scan shows an emerging presence — visible, but not yet consistent.
What to expect next: most brands' AI visibility holds fairly steady between quarters — fewer than 10% of brands shift 30–50% in a single quarter. That's a general benchmark, not a prediction for your brand specifically — the only way to know if yours is moving is to re-scan and compare.
Get the Audit — $149