PostHog mobile funnels for subscription apps: find the leak, then fix it in-app

PostHog mobile funnels are excellent at answering where users drop off. Install to signup to trial start to paid: the chart turns red at the step that hurts. What PostHog does not do by itself is ship a fix on Tuesday without an App Store release.

The gap between diagnosis and deployment is where subscription teams lose weeks. You find the leak in PostHog on Wednesday. Engineering picks up the ticket the following sprint. QA blocks on a build. Review adds days. The funnel you measured is not the funnel users see when the fix ships.

This post covers how to use PostHog funnels for subscription apps, which events to instrument, how to read mobile-specific drop-offs, and how to close the loop by fixing onboarding and journey steps in-app without waiting on releases.

What PostHog funnels do well on mobile

PostHog aggregates events from your iOS, Android, or React Native app into funnel steps. You define the sequence. PostHog calculates conversion and time between steps.

Strength	Subscription use case
Flexible event model	Custom steps beyond SDK defaults
Cohort breakdown	Compare organic vs paid
Retention and paths	Post-trial engagement
Session replay (where enabled)	Watch confusion on a screen
Self-host or cloud	Data residency choices

Typical subscription funnel in PostHog:

1. app_installed or first_open
2. onboarding_step_completed (with step index property)
3. signup_completed
4. paywall_viewed
5. trial_started or purchase_completed
6. subscription_renewed

The leak between steps 3 and 5 is where most growth teams live.

Instrumentation checklist for mobile subscription funnels

Garbage in, garbage out. Before trusting funnel colors, confirm event quality.

Event	Properties to include	Pitfall
onboarding_step_viewed	step_id, step_index, flow_version	Missing version breaks A/B readouts
onboarding_step_completed	Same as viewed	Counting skip as complete
paywall_viewed	offering_id, placement	No placement = cannot fix timing
trial_started	product_id, source	Duplicate fire from SDK + server
purchase_failed	error_code	Swallowing failures hides leaks

Identity: merge anonymous pre-signup events with identified post-signup users. PostHog alias patterns matter for mobile where install id differs from account id.

Flow version: when you run experiments, every event should carry flow_variant or experiment_id. Otherwise PostHog blends control and variant into one misleading funnel.

Reading mobile funnel drop-offs

Not every red bar needs a paywall redesign. Common mobile patterns:

Drop-off pattern	Likely cause	Fix layer
Cliff after step 2	Value prop weak or too early	Onboarding copy / order
Gradual bleed steps 3-6	Flow too long	Shorten journey
Spike at permission step	ATT or push timing	Branch or pre-prompt
High reach paywall, low trial	Journey qualified users poorly	Upstream messaging
High trial, low D7 retention	Product, not funnel	Outside funnel scope

PostHog tells you where. It does not tell you which variant of a multi-step native flow to ship without either feature flags in code or a remote flow platform.

The diagnosis-to-deployment gap

A typical workflow today:

1. PostHog funnel shows 35% drop at onboarding step 4.
2. PM writes spec for shorter path.
3. Engineering implements new navigation branch.
4. Release train + store review.
5. Two weeks later, measure again.

During those two weeks, ad spend sent users through the broken step 4. Funnel analytics aged out of relevance.

What teams need is a closed loop:

Phase	Tool	Output
Diagnose	PostHog funnel	Step id with worst conversion
Hypothesize	Growth doc	One change per experiment
Deploy	Rheo (or equivalent)	Publish variant without store
Verify	PostHog + Rheo step analytics	Variant comparison

PostHog remains the system of record for product-wide analytics, retention, and cross-platform views. Rheo owns the native journey graph and step-level experiment assignment for onboarding and lifecycle flows.

Connecting PostHog to journey fixes

You do not need to replace PostHog. You need events from both systems to tell a coherent story.

Option A: Rheo events forwarded to PostHog

Fire Rheo step events into PostHog with the same step_id and experiment_id properties you use in dashboards. Build funnels in PostHog that mirror Rheo flow structure.

Option B: Dual read

Use Rheo for step completion during active experiments (fast, flow-native). Use PostHog for longer retention and revenue cohorts. Align on naming conventions in a shared tracking plan.

Question	Best source
Which variant won step 5 this week?	Rheo experiment view
D30 retention by install cohort	PostHog
Cancel flow completion rate	Rheo lifecycle channel
Feature usage after trial	PostHog

Example: fixing a subscription leak end to end

Symptom: PostHog shows 50% drop between signup_completed and paywall_viewed. Only 40% of signups reach paywall within session one.

Diagnosis: Step analytics (Rheo or granular PostHog events) show users exit on a permissions screen between signup and paywall.

Hypothesis: Moving push permission after paywall increases paywall reach without hurting trial rate.

Deploy: Publish variant B in Rheo with reordered steps. No App Store release. Assign 50/50 on the first-launch channel.

Verify: Compare paywall_viewed rate by experiment_id in PostHog. Confirm trial start rate did not collapse.

Note on paywall: Rheo changed when users arrive at the paywall step. RevenueCat or your paywall integration still owns paywall UI and purchase. If trial rate moves after reach improves, run a separate paywall UI test in your billing tool.

Funnel definitions that mislead subscription teams

Bad funnel	Why it lies
Install → paid in one chart	Hides onboarding leaks
Paywall viewed → trial only	Ignores qualification path
No experiment property	Blends variants
Calendar time windows too short	Mobile installs are bursty

Build stage funnels aligned to your journey:

1. Activation funnel: install through first value action
2. Monetization funnel: value action through trial start
3. Retention funnel: trial through paid renewal

PostHog can chart all three. Fix leaks in the stage that actually moved.

PostHog feature flags vs flow changes

PostHog feature flags can toggle boolean gates in your app: show screen A vs B if you already built both in code. That still requires engineering to merge both paths and release the flag infrastructure.

Approach	Best for	Limitation
PostHog feature flags	Simple A/B on existing screens	Both variants must exist in binary
Remote flow platform	Multi-step structure, copy, order	Requires SDK integration
PostHog experiments + custom code	Engineers own variants	Slow for growth-led iteration

For subscription onboarding with five or more steps and frequent copy tests, remote flows outperform flag-only setups. Flags remain valuable for product features outside the journey.

Metrics map: PostHog and Rheo together

Metric	PostHog	Rheo
Install to signup	Yes	Step funnel
Onboarding step drop-off	If instrumented	Native per step
Experiment assignment	If property sent	Built-in
Trial / purchase	Yes (from billing events)	Via integration node
Paywall UI variant	If RC events forwarded	Placement only
Win-back flow	Custom events	Lifecycle channel

Practical workflow for subscription growth teams

1. Weekly funnel review in PostHog on activation and monetization stages.
2. Pick one leak with enough volume to detect change.
3. Ship journey fix via Rheo when the change is order, copy, branch, or placement.
4. Ship paywall UI fix via RevenueCat / paywall vendor when the change is pricing presentation only.
5. Document in tracking plan so next quarter's analyst sees the same step ids.

Summary

PostHog mobile funnels find the leak. They do not close the loop alone. Subscription teams win when diagnosis in PostHog connects to in-app journey fixes without store review:

• Instrument step events with flow_version and experiment_id
• Read drop-offs by stage, not only install-to-paid
• Publish onboarding and lifecycle changes through a remote flow SDK
• Keep paywall UI and billing in your integration layer
• Verify fixes in PostHog with variant-aware funnels

Start for free and close the gap between your PostHog funnel and your next in-app fix.