Pulsetrack: Know What People Do, Without Selling Them
Posted: 2026 | Category: Product Analytics
The Tuesday Where Your Own Product Is a Mystery
Someone asks a simple question in a meeting: “do people actually use the export button?”
Silence. Then guessing. Then someone opens Google Analytics, spends four minutes finding a report, and reads out a number for page views on a route that has not existed since the last redesign. Somebody else says “we could add an event for that,” which means a ticket, which means next quarter.
Meanwhile your site opens with a cookie banner that 60% of visitors reject, which means the numbers you do have describe less than half your users, and you send the other half to an advertising company anyway.
You are paying a privacy cost for data you cannot use to answer a question about your own product.
Three Problems in a Trenchcoat
The consent problem. Under GDPR, tracking that feeds an advertising ecosystem needs consent. So you show a banner. The banner annoys people, depresses conversion, and blows a hole in your data. You now have a worse product and worse analytics.
The ownership problem. Your usage data is one of the few genuinely proprietary things you own. Sending it to a third party who also sells advertising is an odd choice when you think about it for more than a second.
The question problem. The question is which users reached step three, who came back in week two, what path they took before they cancelled, and whether last Thursday’s spike was real or a bot.
Page views were never the question.
The question is simple. Getting an answer out of the analytics is not.
How We Glued It Shut
Pulsetrack is product analytics you host yourself. It tracks events, identifies users, and answers the questions you actually ask, on infrastructure you control.
Two ways in, both boring. A server-side API with a Bearer key for identify and track, and a browser snippet small enough to paste into your <head> that posts events with sendBeacon. That is the whole integration.
No cookie banner, by design. The browser snippet uses a first-party identifier in local storage, and the data goes to your own domain and stays there. Nothing is shared with an advertising network, because there is no advertising network. This is the technical basis on which a consent banner stops being necessary; your DPO gets the final word, but the argument is a short one.
Identity that survives. An anonymous session becomes a known user the moment you call identify, and the history joins up. Users are resolved by email, carry arbitrary traits, and are linked per product with an external id and a plan, so you can slice by tier without a second system.
Products and rotatable keys. One workspace holds several products, each with its own API keys and labels, so rotating a key is not an outage.
What You Can Actually Ask
Funnels. Define a sequence of events, see where people fall out. Not a guess about where they fall out.
Retention. Cohorts by week, who comes back and who does not.
User flows. The real paths through your product, including the ones nobody designed.
Per-user timelines. Open one person and read what they did, in order. Support tickets stop being archaeology.
Property breakdowns. Every event carries arbitrary JSON properties. Break any metric down by any property key without shipping new code.
Geography. Country-level, derived server-side. Enough to know where your users are, not enough to be creepy.
Every question above is really the same one: where does the drop-off happen, and to whom.
Live view. What is happening right now, for the moments when a launch is going out and refreshing a dashboard is the only thing keeping you calm.
Anomaly detection. Something moved unusually. Pulsetrack says so instead of waiting for you to notice in a month.
Comparison. Any range against the previous one, so “is this good?” has an answer.
AI insights. Point the analytics context at a model and get a written read: metric health, the golden path, where conversions leak, retention patterns, anomalies, and a short ranked list of what to do. It is a briefing, not a chatbot. There is also a copilot for follow-up questions when the briefing raises one.
Export. Events and users out as files, whenever you want. It is your data, and it behaves like it.
Bots are marked. Events are flagged rather than silently mixed in, so your numbers are about humans.
The Boring Parts That Matter
Multi-tenant workspaces with full data isolation. Admin, user and viewer roles. JWT auth with optional single sign-on. Zod validation on every endpoint and rate limiting on the API. Events are indexed for the queries the dashboards actually run, so a funnel over a busy month does not become a coffee break.
Feature flags and notification integrations ship with it, because a product analytics tool that cannot tell you when something broke is only doing half the job.
Self-hosted means the data sits somewhere you can walk to.
Who This Unsticks
SaaS teams in Europe. You need to know what people do, and you need to be able to explain your data processing in one paragraph. Self-hosted analytics makes both of those easy at the same time.
Agencies and studios. One workspace per client, one platform, complete isolation, no per-client Google account gymnastics.
Anyone who removed the tracking and regretted it. You dropped the third-party scripts for good reasons and lost all visibility. This is the version where you keep both.
Teams tired of page views. If your questions are about funnels, cohorts and paths rather than sessions and bounce rate, you have outgrown a website analytics tool. This is a product analytics tool.
What Makes This Different
Plenty of analytics products are privacy-friendly. Fewer are privacy-friendly and answer product questions. Most privacy-first tools are website analytics with a nice chart of page views, and most product analytics tools are hosted in someone else’s cloud on someone else’s terms.
Pulsetrack is the overlap: funnels, retention, flows and per-user timelines, running on your own machines, with your own database, and an export button that works.
- No cookie banner
- No data leaving your infrastructure
- No per-event pricing that punishes you for instrumenting properly
- No enterprise sales call to see a demo
Get Started
Still guessing whether anyone uses the feature you spent a month on? Still apologising to visitors with a banner in exchange for numbers you do not trust?
Visit pulsetrack.glue.blue, paste the snippet, and watch your first events land.
Pulsetrack is built by glue.blue - we stick solutions to problems. Permanently.
Ready to own your analytics? → Get Started → Contact Us
Posted in: Product Analytics, Privacy, GDPR, Self-Hosted