TECHPERFORMANCE
Solution · Analytics warehouse

Your traffic data, out of Google.
Raw schema in-house, interactive queries, predictable cost.

Managed self-hosted ClickHouse, first-party ingest from your domain, custom Grafana dashboards. Post-GA4, post-Schrems II, post-Universal Analytics. When you want to ask questions of your data again.

Talk to an architect See the stack
The problem

Three things happening to almost every Italian product and marketing team.

01
// anti-pattern · analytics

GA4 is a black box your DPO struggles to defend.

After Schrems II and the 2022 rulings from the Italian DPA, the GA4 entry in the processing register got complicated. The easy answer would be "we no longer use it". The real one is "we have the data in-house".

02
// anti-pattern · analytics

Events scattered across different SaaS, each with its own bill.

Product analytics, web analytics, ad platform tracking, application logs. Each team has its own tool, each tool its own invoice, each invoice a tier that kicks in at the first peak.

03
// anti-pattern · analytics

Questions you can’t ask because the data is locked in.

You want to know how a user segment behaved 18 months ago. Or join CRM with web events. Or run custom cohort retention. On standard SaaS it’s "no", "not available", "enterprise upgrade".

How we run it

Gradual repatriation, no big bang.

We set up a first-party collection endpoint on your domain, run it in parallel with GA4 (or whatever you use) for a few weeks, import the history where it matters, then disconnect. Meanwhile, you keep seeing the numbers you always saw — and you start asking questions you couldn’t ask before.

  • →Same team from single site to HA cluster.
  • →Response within 1h with an SRE, not a dispatcher.
  • →Entirely EU infrastructure.
stack · analytics warehouse
first-party · EU-only
  • ClickHouse columnar cluster
    compressed storage, interactive queries on billions of rows
  • First-party collection endpoint
    events collected on your domain, no third parties in the user ID, payload under control
  • PostgreSQL for metadata and segments
    segment configuration, user mapping, product entities
  • ETL pipeline with n8n and dedicated workers
    sync towards CRM, ad platforms, billing, BI — versioned orchestration
  • Grafana for self-service dashboards
    panels for product, marketing, finance, each with its queries
  • Configurable backup and retention
    long history without surprise price tiers, periodic restore drill
Expected results

Vendor independence, not as a slogan.

What changes is not the Monday-morning dashboard — that you can replicate. It’s the class of questions the product team starts asking when raw data is in-house.

0 vendor

extra-EU analytics. Raw event schema on your domain, on your racks, under your contracts.

multi-year

configurable retention without surprise pricing tiers. History queryable, not archived.

interactive query

on billions of rows in seconds. New questions the product team had not yet asked.

Reference setup

Italian publisher in post-GA4 repatriation.

Editorial network with tens of millions of pageviews/month, under pressure from its DPO and CFO. Goal: own the event schema in-house, preserve continuity of historical numbers, reduce the number of analytics vendors from four to one.

Talk to an architect about your case
Reference setup
Documented, anonymized
Publisher · 30M+/month
4 → 1
analytics vendors
6 months
history retention
EU-only
data residency

Result in six months: GA4 disconnected with no gaps in historical series, analytics cost reduced to a single canon, raw event schema available to product and marketing for custom queries.

Reference setup, not a case study of a specific customer. The numbers are consistent with the pattern observed across networks managed by Tech Performance.

FAQ

Questions we get from DPOs, CTOs and CFOs.

If yours is different, write to us: we reply within the day.

How hard is it to leave GA4?
Less than it seems technically, more than it seems organizationally. Technically: set up a first-party endpoint, replicate the event schema you need, run GA4 and ClickHouse in parallel for a few weeks, then disconnect GA4. Organizationally: the real work is getting product, marketing and analytics to agree on which events really matter, because GA4 handled them "automatically".
Is ClickHouse Cloud an alternative? Why self-hosted?
ClickHouse Cloud solves the "I want to use ClickHouse without running it" factor. Tech Performance solves that plus "I want every byte in the EU under a single provider + I want predictable billing + I want a person to call if the midnight query gets stuck". It’s a sovereignty and relationship choice, not just a technology one.
Can I also do product analytics, not just web?
Yes. ClickHouse is not limited to web traffic: the real value is when product events, application logs, mobile telemetry, CRM business events all flow in. A unified raw schema, cross-cutting queries, dashboards for every role.
Can I keep some SaaS in parallel during transition?
Yes, and it’s almost always the right thing to do. For a few months you keep GA4 (or another vendor) as a second reference, you compare numbers, you close schema gaps. Then you disconnect calmly, with no gaps in time series.
Free audit

Let’s talk about your stack,
free, no strings attached.

30 minutes with a Romiltec architect. Together we figure out whether Tech Performance is a fit, and if it isn’t, we tell you straight away. No cold pitch, no black-box quote.

Book a call cal.com/romiltec/tech-performance · 30 min call