PostHog

PostHog vs Userpilot is the wrong framing most of the time. They solve different jobs in the PLG stack: PostHog measures behavior, Userpilot directs it. This page is for buyers who got dragged into a head-to-head spreadsheet anyway and need to know which side of the stack they actually have a gap in.

Typical fit: who each tool is built for

Typical PostHog customer - Indie or Series A–B PLG SaaS, under ~1M MAU, founder or eng-led GTM. - Wants one tool covering product analytics, session replay, feature flags, and LLM observability. - Engineering is willing to instrument events and own the SDK. - Budget is sensitive to seat-based vendors; pay-as-you-go preferred. - Operator pattern, not vendor claim: AI-native startups consolidating LangSmith/Helicone-style LLM cost tracking into the same product they use for funnels.

Typical Userpilot customer - Series A–B PLG SaaS, 1k–10k MAU, CS Ops or Product Marketing seat in place. - Already has product analytics elsewhere (often Amplitude, Mixpanel, or Heap). - Bottleneck is shipping no-code in-app onboarding, tooltips, NPS, and a resource center without engineering. - Mid-market B2B SaaS where CS Ops ships guide variants weekly.

Neither if you're… - An enterprise multi-product org needing analytics + guides + roadmap + feedback portal under one procurement contract — Pendo is the right shape. - A sales-led outbound org looking for sequencer + AI SDR coverage — neither tool plays in that lane.

When PostHog wins

PostHog wins when the question is "what are users actually doing, and why didn't the variant we shipped move the metric?" That's a measurement question with a feature-flag and replay tail. Concretely:

• Input: SDK events from web/mobile, optional Segment or RudderStack ingress, server-side capture for backend events.
• AI step: Max AI for natural-language analytics, replay auto-summaries, LLM-obs span tagging on token cost and latency.
• Human review: Engineer or PM validates event names and flag rollouts before prod; RevOps reviews cohort definitions before any CRM writeback.
• Writeback: Cohort exports to Customer.io or HubSpot lists, Slack/Linear webhooks on funnel anomalies, flag toggles in-app, replay deep links pasted into tickets.
• Metric: Activation by cohort, flag-variant conversion lift, $/feature on LLM spans, replay-watch time per CSM.

If the next blocker on your roadmap is "we shipped the variant but don't know if it worked," Userpilot can't answer that — it's not a measurement tool.

When Userpilot wins

Userpilot wins when the question is "users are getting to the right screen but not doing the right thing — can we nudge them without filing an engineering ticket?" That's a guidance and copy problem, not an analytics problem.

• Input: Identified user events, account traits via Salesforce/HubSpot sync, target-cohort definitions imported from Amplitude/Mixpanel/Heap.
• AI step: AI Writing Assistant drafts guide copy, tooltip text, and survey questions from short intent prompts.
• Human review: CS Ops reviews drafted copy before publishing; product owns which surface gets a guide.
• Writeback: In-app guide shown to user; NPS or survey response back to CRM or to the analytics tool; resource center surfaced contextually.
• Metric: Guide completion rate, feature adoption lift, NPS trend, time-to-first-guide vs. previous (eng-built) onboarding.

If your eng team has refused to prioritize onboarding tooltips for the third quarter in a row, Userpilot replaces that ticket queue with a CS Ops-owned workflow.

When you need both

Most Series A–B PLG SaaS readers of this page need both, not one. The stack we see working in practice:

• PostHog as the measurement layer (events, replay, flags, LLM obs).
• Userpilot as the nudge layer (in-app guides, NPS, resource center).
• A shared cohort definition: PostHog defines "users stuck at step 3"; Userpilot targets that cohort with a guide; PostHog measures whether the guide moved the metric.

The integration is real: Userpilot consumes cohorts and events from PostHog (via Segment or direct integration), and PostHog measures the impact of the guide variant. See the CSM onboarding automation playbook for the closed-loop workflow, and the CSM health score playbook for how the same cohort feeds account health rather than just onboarding.

This is also the configuration where the PostHog vs Amplitude decision actually matters more than the PostHog vs Userpilot one — the analytics layer is the foundation, and the guide layer composes on top.

Pricing and per-account math

PostHog pricing is on the pricing page; Userpilot pricing is on the pricing page and is MAU-tiered.

Per-account math is the wrong frame here because the tools price on different denominators. The honest comparison is:

• At 5k MAU PLG SaaS, ~20M events/mo, ~10k replays/mo: PostHog bill is event + replay metered (verify against the calculator); Userpilot Growth is $749/mo flat. They are not substitutes — adding Userpilot on top of PostHog is a net new line item, not a swap.
• At 50k MAU: Userpilot pricing climbs steeply; this is where buyers re-evaluate against Pendo's bundled analytics + guides TCO, not against PostHog.

Do not model a "switch from PostHog to Userpilot saves money" scenario — they don't cover the same surface.

Feature overlap and gaps

Genuine overlap is narrow: surveys and NPS exist in both, and both touch onboarding from opposite ends.

Treat this matrix as "do I have this gap in my stack" rather than "which tool wins the row." Most rows have a clear primary tool.

The buying mistakes we see most

1. Buying Userpilot to replace analytics. Userpilot's path analysis and dashboards look like analytics until you try to define a cohort from raw events. It is a guide-layer product; pairing it with Amplitude, Mixpanel, or PostHog is the design.
2. Buying PostHog to ship onboarding tooltips. Engineering goes quiet for a sprint, ships a half-built in-app guide framework, and the CS Ops owner ends up filing tickets anyway. PostHog has no native guide builder — buy a guide tool, not a flag tool, for that job.
3. Picking one because the other "covers it." Vendor decks on both sides exaggerate adjacency. Userpilot's analytics module is light; PostHog's "surveys" are basic. Buyers who pick one and skip the other usually re-buy the other within 12 months.
4. Ignoring the MAU vs event mismatch. PostHog scales on events (you control via instrumentation hygiene); Userpilot scales on MAU (you don't, unless you stop activating users). A team that wins activation hard sees Userpilot's bill jump while PostHog's stays flat.
5. Treating Userpilot's AI Writing Assistant as a publishing autopilot. It's a drafting tool. Generic AI-drafted onboarding copy ships, users dismiss the tooltip soup, and the team blames the tool instead of the review step that was skipped.

What to test in week 1

PostHog test (engineering-led, ≤5 days):

1. Instrument one activation event tied to revenue ("completed onboarding step 3 within 24h").
2. Build the funnel; sample 5 dropped users from replay; document what actually breaks.
3. Ship one flag-controlled tweak to the broken step (A/B).
4. Measure: activation lift, total elapsed time vs. your prior stack, replay-watch time per CSM.

Userpilot test (CS Ops-led, ≤5 days):

1. Pick one adoption gap tied to expansion ("users on Pro tier not using feature X within 30 days").
2. Define the cohort from your existing analytics (Amplitude/Mixpanel/PostHog).
3. Use the AI Writing Assistant to draft an in-app guide; ship to 50% of the cohort (A/B).
4. Measure: guide completion rate, feature-X adoption lift, time-to-first-guide vs. your prior tool.

If you can't complete either in a week, the bottleneck is upstream — instrumentation hygiene for PostHog, cohort definitions for Userpilot. See the CSM onboarding automation playbook.

Migration and coexistence

For most teams the question is not migration but coexistence. Run them in parallel for 90 days:

• Cohort source: PostHog defines and owns the cohort; Userpilot consumes it via integration or via Segment.
• Event source of truth: PostHog. Do not duplicate event instrumentation in Userpilot's SDK if PostHog already captures it.
• Guide-impact loop: Userpilot ships the guide; PostHog measures the lift on the downstream funnel step.
• Contract risk: Userpilot's MAU tier is the line item that surprises buyers at renewal — watch the MAU curve quarterly.

If you are migrating from an analytics tool (Mixpanel/Amplitude) to PostHog while also adding Userpilot, do not do both at once. Stabilize the analytics migration first; add Userpilot in quarter two.

FAQ

Does Userpilot need a separate analytics tool? Practically, yes. Userpilot ships path analysis and basic event tracking, but defining cohorts from raw events and running funnels is shallow. Most production deployments pair Userpilot with Amplitude, Mixpanel, Heap, or PostHog.

Can PostHog ship in-app guides? Not natively. Engineers can render in-app messages from feature flags, but there is no no-code guide builder, resource center, or NPS tooling on par with Userpilot.

Does PostHog's session replay cover what Userpilot's analytics show? Replay shows individual user sessions; Userpilot's analytics is aggregate funnel and adoption charts at a shallower depth than dedicated analytics tools. They are not substitutes.

If we already have Amplitude, do we still want PostHog? Usually no — pick one analytics layer. Userpilot composes on either. See PostHog vs Amplitude for the analytics layer decision.

Does gtmpod earn commission on either tool? PostHog: yes, disclosed at the top. Userpilot: no.