PostObs

AI that tells you why users fail onboarding.

Sales tells you what users say. We tell you what they actually do— and what they'd say about it if no one was translating. Upload your funnel data, get a diagnosis in plain English within 24 hours.

Works with Amplitude, Mixpanel, Heap, PostHog, or any CSV with event data.

Or talk to me first — book a 30-min call.

Example insights from a real analysis

42% of LinkedIn traffic drops at onboarding step 2

Cause: Expectation mismatch — users expect instant value, see config first.

Fix: Move sample-data demo before configuration step.

Mobile signups convert at 1/3 the rate of desktop

Cause: File picker UX fails on touch devices.

Fix: Detect mobile, offer email-yourself-a-link alternative.

Users who revisit pricing convert 2.4× more

Cause: Pricing is hidden behind 3 clicks from the landing page.

Fix: Surface pricing in the top nav.

We'll also tell you what to track better

Every diagnosis comes with concrete instrumentation gaps — the events, properties, or tags your team isn't capturing that would sharpen the next pass. Useful to PMs, doubly useful to engineering.

No time-on-step events

Why it matters: Funnel transitions are tracked, but not how long users spent on each. Impulsive drops look identical to contemplative ones.

How to add: Emit step_view with duration_ms when the user leaves each screen.

UTM tags missing from LinkedIn share URLs

Why it matters: All LinkedIn-sourced traffic falls into a single bucket. Can't tell which post / which audience converts.

How to add: Add utm_source=linkedin and utm_campaign=<post-id> through a link shortener.

Founding pricing

$499/mo $199/mo — locked in for life.

Reserve founding access

Not sure yet? Book a call instead.