Skip to content
All projects

5 questions, 1 owned platform

built first-party from scratch · 0 third-party trackers on the site

Case study · Small-group fitness · Berlin

How Forvya got reliable data without over-engineering it

Forvya is a premium small-group fitness studio in Berlin, founded by two entrepreneurs. Every new member follows the same path: discover, visit the site, submit a lead, book a trial, become a member. The marketing job is simple - generate qualified leads efficiently. So we built the data to match the job, not the other way around. Here's what right-sized looks like, so you can tell whether you're over-built.

The problem

When companies think analytics, they reach for GA4. For a lead-gen business, that's the wrong default

Forvya didn't need hundreds of auto-collected events, dozens of reports or a complex attribution model. They needed reliable answers to five questions:

The usual move is to bolt on enterprise analytics before anyone knows what they'll actually use. Here's what that quietly costs: every extra report is something to maintain, second-guess and explain. More dashboards rarely make better decisions - they make slower ones. For a small founder-led team, complexity you can't operate is a tax you pay every week and never deduct.

What good looks like

The best analytics setup is the simplest one that answers your real questions - and that you own

Right-sized doesn't mean cheap or basic. It means deliberate. You can check your own setup against three marks:

If you're loading Google Analytics and a stack of third-party trackers to answer questions you could answer with five clean numbers you own, you're not more data-driven. You're more exposed, and harder to maintain.

What we did

We built a first-party data platform around the decisions, owned end to end by Forvya

Four decisions shaped everything:

  1. 01

    Build around business decisions, not a tool. We collected exactly the metrics that drive a lead-gen business and nothing else - so reporting stays understandable and the founders decide fast without wading through dashboards.

  2. 02

    Own the data from the first event. Every interaction is collected directly into Forvya's own Google Cloud and stored raw in BigQuery - their marketing data warehouse, theirs entirely. Retention, reports and downstream sharing are all Forvya's call. Nothing is locked in a third party's tool.

  3. 03

    Privacy by design. The site loads no Google Analytics and no third-party tracking libraries. Visitors get a single first-party cookie holding an anonymous ID - enough to recognise a returning visitor inside Forvya's own setup, with no personal data attached. Marketing tracking and customer records stay separate; identity is only known once someone becomes an actual customer.

  4. 04

    Send ad platforms only the signals they need. Owning the full data flow means Forvya chooses exactly what reaches Meta and Google Ads - just the optimisation signals the algorithms need to train, no more. If performance ever justifies sharing more, the setup expands in a controlled, transparent way. Privacy and ad performance aren't a trade-off; they're an engineering decision.

The result

A right-sized data platform that answers the questions and scales

The founders get reliable visitor numbers, reliable lead tracking and simple business reporting - on data they own completely, built privacy-first. The current limited signal set already delivers strong cost-per-lead performance, and because the data flow is owned end to end, the architecture can expand in a controlled way the moment more would actually pay off. The system stays focused on supporting decisions instead of generating more dashboards.

The takeaway

Good data infrastructure isn't about collecting as much as possible. It's about collecting the right data, and owning it

Most businesses think they need a bigger analytics stack. Usually they need trustworthy data, clear reports and control over their own setup. Technology should adapt to the business, not the other way around. If you're about to buy complexity to answer five simple questions, stop - build the five answers first, and own them.

Let's make your data trustworthy enough to decide on

Talk to Dominic