Hotjar and PostHog both record sessions and draw heatmaps, so they end up on the same shortlist. They answer different questions, though. Hotjar shows you what visitors do on a page. PostHog tells you what users do across your product, then lets you ship a feature flag or an A/B test off the back of it. Pick the wrong one and you'll either drown a marketing team in event schemas or leave a product team squinting at heatmaps when they needed a funnel.
Short answer: Marketing and UX teams working on a website should start with Hotjar. Product teams with at least one engineer should pick PostHog, because the free tier covers more and it replaces three or four other tools. If you only need heatmaps on a handful of landing pages, Hotjar's free plan is enough and PostHog is overkill.
What each tool actually is
Hotjar is a behavior analytics tool: heatmaps, session recordings, on-page surveys and feedback widgets. Contentsquare bought it in 2021. You install one script tag, and recordings start showing up within minutes. A marketer can be productive on day one without touching an event schema. It does have funnels and trends on its top plan, but quantitative analysis isn't what it's built for.
PostHog is an open-source product analytics suite. One dashboard covers event analytics, funnels, session replay, heatmaps, surveys, feature flags, A/B testing and a data warehouse. You can run it on PostHog's cloud (US or EU hosting) or self-host it with Docker Compose or Kubernetes. It's built for product and engineering teams, and it expects someone to think about which events to track.
That difference in audience explains nearly every trade-off below.
Pricing side by side
| Plan | Hotjar | PostHog |
|---|---|---|
| Free | Basic: 35 daily sessions, automatic heatmaps, page-targeted surveys | 1M events/month, 5K session replays, 1M feature flag requests, unlimited team members |
| Entry paid | Plus: $32/month, 100 daily sessions | Pay-as-you-go: no base fee, all features, $0.00031 per event after the first 1M |
| Mid tier | Business: $80/month, 500 daily sessions | None (usage-based) |
| Top tier | Scale: $171/month, unlimited sessions, funnels, trends | Enterprise: custom (SSO/SAML, advanced permissions, custom retention) |
Hotjar prices by daily sessions on a flat monthly plan. PostHog charges for usage past a free allowance, with no minimum. That makes Hotjar easier to budget and PostHog cheaper until you're at real volume.
Some quick math. A site doing 2 million tracked events a month on PostHog pays for the million past the free tier, which comes to about $310 at $0.00031 per event. Session replays and flag requests are metered separately once you pass their free allowances, so run your own numbers through PostHog's calculator before you commit. On Hotjar, a team that needs more than 100 recorded sessions a day jumps from $32 to $80, and unlimited sessions cost $171 a month. Neither is cheap at scale. They just get expensive in different ways.
Where Hotjar wins
It's usable by people who don't write code. That's the main reason to pick it. A UX researcher or growth marketer can install it, watch ten recordings of the signup page, and come back with three fixes before lunch. PostHog can do that too, but the interface asks you to understand events, properties and insights first.
Heatmaps are the core product, not an add-on. Click, scroll and move maps aggregate cleanly, and they're easy to screenshot into a slide for a stakeholder who'll never log in. For landing page work that matters more than most analysts admit.
Surveys and feedback widgets sit right on the page. NPS prompts and one-question intercepts are built in, and page-targeted surveys come with the free plan. They're basic, but they put the "why" next to the "what".
Frustration signals on Business. Rage clicks, dead clicks and console errors get flagged, so you don't have to sit through hours of footage to find the broken button.
Where Hotjar falls short
The free plan runs out fast. 35 recorded sessions a day is nothing for a site with real traffic. You'll see a sample, not your audience.
It slows pages down. Our product data puts the script at roughly 50KB, loading semi-synchronously by default, and it's the most common complaint we see from developers. If your Core Web Vitals are already marginal, Hotjar makes them worse.
It isn't product analytics. Our plan data lists funnels and trends on the $171 Scale plan, and even there they're thin next to a dedicated tool. If someone asks "what percentage of trial users created a second project within a week?", Hotjar won't answer that.
Privacy needs configuring. Data is hosted in the EU, and PII masking is available, but IP anonymization and consent mode take manual setup. That's not unusual for this category, but don't assume the defaults pass your DPO.
Where PostHog wins
The free tier is very generous. One million events, 5,000 session replays and a million feature flag requests a month, with unlimited team members. Plenty of early-stage startups never pay.
It replaces several tools. Event analytics, session replay, heatmaps, surveys, feature flags and A/B testing in one place means one SDK, one bill, and one data model. You can trace a drop in a funnel, watch the replays for that step, and ship a flag to test a fix without exporting anything.
You can self-host it. For a security team that's the headline feature. Self-hosting keeps behavioral data on your own infrastructure, which simplifies GDPR and data residency conversations a lot. If you'd rather not run it, EU Cloud covers most residency requirements. PostHog states it doesn't train AI on customer data and holds SOC 2 Type 2.
Developers like it. SQL access, data portability and transparent usage-based pricing make engineers far less likely to route around it.
Where PostHog falls short
It needs an engineer. Not for the snippet, which installs in minutes, but for deciding what to track and keeping event names consistent. A marketing team without technical help will get less out of it than out of Hotjar.
Self-hosting is real work. PostHog runs on Postgres and ClickHouse, and both need proper hardware once event volume climbs. "Free and self-hosted" still costs servers and someone's time.
The polish gap is real. The UI has improved a lot, but it's still less polished than Amplitude or Mixpanel, and the mobile SDKs and very large datasets are where users report the most friction. Feature flags and experiments work well for most teams, but they aren't at LaunchDarkly's level yet.
Usage pricing can surprise you. Pay-as-you-go is great until a tracking bug fires an event on every scroll. Set billing limits on day one.
Who should pick which
Pick Hotjar if your team is marketing, UX or content, your main question is "why aren't people clicking this", and nobody wants to maintain a tracking plan. Plus at $32 a month is a fair price for a few landing pages under active optimization.
Pick PostHog if you're building a product with logged-in users, you have an engineer who can own instrumentation, and you'd otherwise pay separately for analytics, replay and feature flags. Start on the free tier and set a spending cap. If you're also comparing it to the big product analytics players, our PostHog vs Mixpanel vs Amplitude comparison goes deeper on funnels and retention.
Use both if marketing owns the website and product owns the app. That combination is common and it works. Just don't load both scripts on the same pages without checking the performance hit.
Skip both if you only want traffic numbers: visits, referrers, top pages. A lightweight tool does that with far less script weight, and our Plausible vs Google Analytics refresh covers that decision.
My take
If I were advising a typical SaaS company with an engineering team, I'd pick PostHog. It covers more ground for less money, and self-hosting or EU Cloud makes the data protection conversation easier than with almost any other analytics vendor. I'd only reach for Hotjar first if the people doing the work are marketers who need answers this week and don't have engineering time to spare. In that case, it's the faster tool to get value from, and for a website that's what matters.