FirePing analytics / 30 days ending 12 August 2026
From attention
to protection.
What Plausible, PostHog and the product database tell us about acquisition, map utility, activation and alert value.
The central finding
People understand and use the map. The largest loss happens before account creation—not during setup.
Scroll to explore ↓
01 / Executive signal
The map already
delivers value.
FirePing is not struggling to explain the immediate product. It is struggling to turn public utility into durable protection.
02 / The conversion gap
A healthy public experience.
A severe pre-account loss.
13,467visitors
11,691map loads
4,296fire lookups
927signup visits
159accounts created
157verified
1.17%
rough visitor → verified conversion
03 / Acquisition
Search sustains.
Facebook surges.
Organic search is the dependable engine. Facebook creates sharp, incident-led waves with high signup intent.
Organic search7,033 52%
Direct4,967 37%
Social1,436 11%
Facebook accounts for 1,405 of 1,436 social visitors.
04 / Incident-driven demand
Traffic follows the fire.
Calabria, Italy
592 from Facebook · 756 visitors from Italy
Gironde, France
560 from France · 550 direct · search also strong
United Kingdom
208 from Facebook · 309 visitors from the UK
Evidence limit: coordinates reveal the incident region; Facebook obscures the exact shared post and events lack a stable fire ID.
05 / Audience geography
A global audience
that moves with risk.
The “core market” is not static. Incident location can reshape the audience within hours.
UK19.9%
France17.8%
US17.3%
Italy13.2%
Germany6.3%
The UK rose to 37.1% in the latest seven days.
06 / Map comprehension
People do not merely land.
They explore.
86.8%
load the map
median to first lookup
median map engagement
average map-entry visit
“The immediate proposition is understood.”
07 / Mobile reality
Mobile is healthy—
except where the browser says no.
58–63% of use is mobile. Fire and signup rates are slightly better than desktop.
Interaction responsiveness is weaker: 272ms mobile INP p75 vs 200ms desktop.
19 users hit unsupported push; 11 repeatedly clicked “Enable Push Notifications”.
Fix capability messaging—not mobile map fundamentals.
08 / What people find
Most lookups find
no active detection.
That is not failure. It is a high-intent moment: “nothing detected now—monitor this area for me.”
64.6%zero detections
35.4%found ≥1 detection
2.5%of zero-result lookups used the monitor CTA
09 / Intent signals
A small group explores deeply.
24-hour mode95.4%
Timeline mode4.6%
6.5%
of timeline sessions reached verification—roughly 2–3× ordinary lookup sessions.
Correlation, not causation: timeline users may simply arrive with stronger intent.
10 / Activation after signup
Once inside, setup works.
Do not spend the next cycle rebuilding onboarding. Improve the reason to begin it.
11 / Alert value
Protection is broader
than push metrics imply.
Only 70 users have web push, but email means 236 verified users have both an enabled location and an active delivery path.
Eligibility
We can authoritatively measure enabled locations, active delivery channels and generated alert events.
Delivered value
Alert records show that value occurs. Open, click-through, usefulness and noise are not fully connected end-to-end.
First-alert effect
We still need a clean cohort study: first alert received → return, retention and settings maintained.
12 / Measurement limits
The most useful missing field
is a stable fire identity.
- Fire/entity ID
Cannot rank incidents or prove repeated monitoring of one fire.
- Severity + distance
Cannot compare signup intent by FRP, proximity or perimeter.
- Panel interactions
Cannot tell whether people value FRP, rows, coordinates or timestamps.
- Alert lifecycle
Cannot cleanly connect generated → delivered → opened → returned.
- Identity linkage
Historical anonymous-to-account journeys remain imperfect.
- Contribution maturity
Not enough established contribution data to optimise prompts responsibly.
13 / Product priorities
Where to invest next.
- 01Highest leverage
Convert high-intent map moments
Strengthen the contextual route from a viewed incident—or a zero result—to monitoring that area.
- 02Enables learning
Instrument fire identity and alert outcomes
Add stable entity IDs, severity, distance and the delivered/opened alert lifecycle.
- 03Low effort
Remove push dead ends
Detect unsupported browsers early and lead with email as the dependable default.
- 04Then monetise
Earn retention and contributions after value
Measure the first-alert cohort; ask for support only after useful information or protection has been delivered.
The decision
Do not redesign
what already works.
Make the path from “I’m checking a fire” to “tell me if risk changes” feel like one natural action.
Next studyTraffic source → activated notifications
ThenFire inspection → signup conversion
ThenFirst alert → return and retention