The Ecomm Analyst

PostHog has quietly become the default answer for engineering-led teams that want product analytics, session replay, feature flags, and experiments in one place instead of buying four tools. It is open source, you can self-host it, and the free tier is genuinely generous (a million events, five thousand session recordings, and a million feature flag requests every month before you pay anything). For a lot of Shopify and DTC teams that is enough to run real analysis without a bill. But it is not the right fit for everyone, so here are the alternatives I actually compare it against.

Before the list, one honest caveat about the whole category. Product analytics answers “what are people doing on my site or in my app,” which is a different question from “which ad drove this sale.” If you are a store owner trying to figure out where revenue comes from, event analytics is the wrong shelf. It shines when you have a real onboarding flow, an account area, or a configurator, and you want to see where people drop off. If your site is a storefront and a checkout, you may get more from your ad and attribution tooling than from any of these.

The alternatives worth looking at

Mixpanel is the one most people weigh against PostHog first. It is easier for non-technical PMs and growth folks to navigate quickly, and self-serve funnel and retention analysis is its strength. Worth knowing that Mixpanel cut its free tier from twenty million to one million events in late 2025, which narrowed the gap that used to make it the obvious free pick. If your priority is fast analysis for marketers rather than a developer platform, it is still the friendliest option.

Amplitude is the enterprise end of this. If your organization runs on behavioral analysis, stakeholder dashboards, and formal reporting workflows, it goes deeper than PostHog on niche behavioral cohorts and specific funnel analysis. It also carries enterprise pricing and a sales motion to match, so it earns its keep at scale and feels like overkill below it.

Heap deserves a mention for its autocapture-first approach. It records everything by default so you can answer questions retroactively without having instrumented the event ahead of time, which is useful when you do not know yet what you will want to measure. The tradeoff is the usual autocapture one: a lot of noise to sort through, and costs that can climb as volume grows.

Matomo is the pick if privacy and data ownership are the priority. It is open source, GDPR-friendly, and can be self-hosted so the data never leaves your infrastructure. It leans more web analytics than deep product analytics, so treat it as a GA4 alternative with product features rather than a full Amplitude replacement.

Plausible is the lightweight end of the spectrum. If you do not actually need event-level product analytics and just want clean, privacy-first web metrics, it is a fraction of the cost and a fraction of the setup. I mention it because plenty of teams buy a heavy product analytics tool and use ten percent of it, and a nine-dollar-a-month web analytics tool would have covered what they needed.

Where does attribution fit in all this? It does not, cleanly, which is the point. None of these tools tell you which campaign earned a purchase. That is a separate job, and for e-commerce it usually matters more than in-app behavior. If that is the gap you are really trying to close, an attribution platform like ThoughtMetric (a sponsor of this site) is built for the “where did this order come from” question that product analytics leaves untouched. I would run one of the tools above for on-site behavior and something purpose-built for attribution, rather than expecting either to do both.

A word on the self-hosting pitch, because it is the thing people get seduced by. Yes, PostHog is open source and you can run it on your own infrastructure for the price of the servers. In practice, self-hosting only pencils out at very high volumes, and even then the DevOps overhead of patching, scaling, and handling incidents tends to run one and a half to three times the visible infrastructure cost. For almost every store I have seen, the cloud version is cheaper once you count the engineering hours you would otherwise spend babysitting it. Reach for self-hosting because you have a hard data-residency requirement, not because it looks free.

To make the “do you even need this” question concrete: the DTC stores that get real value from product analytics are the ones with something to analyze beyond a linear checkout. A subscription brand watching where people fall out of the manage-subscription flow. A store with a product configurator or a quiz, wanting to see which step loses people. A brand with a logged-in account area and repeat behavior worth understanding. If that is you, PostHog or Mixpanel earns its place. If your site is a homepage, some collection pages, a product page, and a checkout, you will instrument three events, look at them twice, and wonder what you bought. Match the tool to whether you actually have a funnel worth studying.

The short version. PostHog if you want the whole product stack in one open-source tool and have engineers to run it. Mixpanel for friendly self-serve analysis, Amplitude if you are operating at enterprise scale, Heap if you want autocapture, Matomo or Plausible if privacy and simplicity win. And do not ask any of them to do your attribution.