the ga4 alternative for developers who just want the answer

if you are a developer looking for a GA4 alternative, the honest short answer is: pick by what question you actually ask. GA4 (Google Analytics 4) frustrates developers for three concrete reasons. its UI buries simple answers under explorations and dimensions, it samples and thresholds data so the numbers feel untrustworthy, and 30–50% of client-side hits are stripped by ad-blockers. for plain web analytics (visitors, referrers, top pages, no cookie banner) Plausible is the reference cookieless choice and does that job beautifully. if you also ask product questions (funnels, retention, paths, "did the thing i shipped yesterday work"), or you want to ask from inside your editor instead of learning a dashboard, smolanalytics is a commercial tool that does web and product analytics together and answers over MCP, in your editor, through your own model. both are cookieless, both count every event. the split is scope, not privacy.

why do developers leave GA4?

three things, and they compound.

the first is that simple questions take too long. you open GA4 to answer "how many people hit the pricing page this week" and twenty minutes later you are three menus deep in an exploration, still without the number. the interface is built for marketing teams doing attribution, not for a developer who shipped a change and wants to know if it moved anything.

the second is that you stop trusting what you see. GA4 samples and thresholds; the number on screen is an estimate of a subset, and once you know that, you stop reasoning from it. product teams quietly admit they do not trust their own GA4 dashboards.

the third is that the data leaks before it arrives. GA4's client-side script is one of the most-blocked requests on the web, so ad-blockers strip a large slice of hits and you are undercounting by a margin you cannot measure.

the honest version of what happens next: most developers do not switch tools, they just stop opening the dashboard, then stop adding it to new projects, and ship blind. the fix is not "spend more time in GA4." it is a tool shaped like the question you actually have. that is the whole argument.

what should you use instead of GA4?

it depends on one thing: do you only need web analytics, or do you need product analytics too.

if you only need web analytics (pageviews, sources, top pages, goals, and no consent banner) use Plausible. it is the most battle-tested cookieless brand, it is lightweight, and it is a genuinely lovely product with a long public track record. if pageviews and referrers are the whole job, it does it perfectly and you should stop reading here.

if you also need product analytics (funnels, retention cohorts, user paths, "where do people drop off after signup") you have historically needed a second, heavier tool beside Plausible. that is the gap smolanalytics fills: web analytics and product analytics from the same events, in one tool, on every plan. no second tool, no cluster.

if what you want from self-hosting is control without an ops project, be honest about the ops first: PostHog self-hosting means Kafka, ClickHouse, Redis and Postgres, and PostHog itself has said most teams cannot run it reliably. smolanalytics takes the other route to control: it is managed, but every project gets its own isolated instance, nothing phones home, and your data exports in one file any time.

where does Plausible fit, and where does smolanalytics fit?

they share values (both cookieless, both count every event, neither is an ad business) so this is not a privacy fight. it is scope.

pick Plausible when web analytics is the entire job and you want the most proven privacy brand. pick smolanalytics when you also ask product questions, or you want the answer in your editor instead of a browser tab.

how does asking in your editor actually work?

this is the part that is different, so here is the concrete shape. you run npx smolanalytics connect once, and your coding assistant (Claude Code, Cursor, Windsurf) gets the analytics as an MCP server: 97 tools and 15 built-in prompts. then you ask, in the same window you write code:

you ▸ did signups improve this week?
ai  ▸ signups: 212 this week vs 180 last week, up 18%.
      the lift is /pricing traffic. the funnel after
      signup is unchanged: 62% reach activate.

those numbers are the shape of a real answer, not real traffic; this launched recently and there is no traffic worth quoting yet. the important part is the trust story: the MCP tools do not generate SQL. they call the exact same deterministic report functions the HTTP API serves, so your model can ask the wrong question but it cannot invent a number. and that is not a promise in a README. a CI agreement test asks the same question through the HTTP API and through the MCP server and fails the build if the answers differ, and every answer carries a computed_by receipt you can re-run yourself. GA4 has no equivalent, because there is nothing to check its numbers against.

what does the GA4 alternative cost?

Plausible starts at $9/month. smolanalytics is a 14-day trial at Pro limits (no credit card), then $19/month on Pro (2M events, one plan), with $6 per million past that and reads that never lock. that is the honest tradeoff: GA4 is "free" because you are the product and Google keeps the data on its servers; this is cheap-flat because the data is yours and there is no ad business behind it.

the honest caveat

smolanalytics is one person, it launched recently, and there are no testimonials on the site because there are none. none will be faked. so trust cannot come from a track record; it comes from the design. every answer carries a computed_by receipt you can re-run, a CI test pins the AI's number to the dashboard's, nothing phones home, and your data exports in one file any time. worst case, if this is not for you, you leave with everything.

if you want the detail, the docs have the install start to finish. or start the 14-day trial, no card, and get the verdict in your inbox on monday.