Marketing engineering · Gräfelfing near Munich
We build the data and tracking architecture your marketing runs reliably on.
When GA4, Ads and your CRM report three different revenue numbers, your marketing is not the problem. The architecture underneath is.
First call within 48 h · report in 2 weeks · from €2,400 net
Discrepancy detected · 3 sources
Fictional example, one month of e-commerce · no live data
Symptom → Service
You recognise the problem. We name the service.
On the left, the sentence we hear in the first call. On the right, the service that fixes it: fixed scope, concrete deliverables.
Symptom
Ads reports fewer conversions than your shop or CRM actually sees.
Measurement & Privacy Engineering
Server-side architecture that recovers lost tracking signals.
Symptom
Meta, HubSpot, and Stripe report three different acquisition costs.
Revenue Intelligence & Executive BI
One source across marketing, CRM, and backend. True CAC, real LTV.
Symptom
Client-side tests flicker and drag down the Core Web Vitals.
Experimentation & Conversion Intelligence
Edge-based. Statistically significant. Without client-side flicker.
Symptom
Shop, ad systems, and CRM show three different revenue numbers.
Data Platform & Governance
Your Marketing Data Lakehouse plus governance, on BigQuery/Snowflake instead of expensive all-in-one CDPs.
Symptom
Smart Bidding, Advantage+, or HubSpot AI run without use-case documentation.
AI Strategy & Data Readiness
From EU AI Act to custom LLMs: your data infrastructure for the AI era.
For full-cycle scopes we build together with our engineering partner Saloid, a brand of the same GmbH: the same senior people, one contract, one sign-off.
Compare scopesWe work data-contract-first
The Data Contract Model.
A data contract works like a real contract: it fixes what data comes in, what goes out, and who is liable when something is off. So instead of stacking tools, we guarantee that your numbers carry decisions.
Input
Website events
Consent signals
Campaign data
CRM data
Contract layer
schema.checkout_purchase
v2.4.1 · owner: growth-eng
Naming
The same event names across web, app and server. No three truths in three tools.
Validation
Events that break the contract are blocked before they distort the reporting.
Ownership
Every number has an owner. Changes are documented, not shouted across the chat.
Governance
You hear about tracking changes before they happen, not when revenue goes missing from a dashboard.
Quality checks
If event volume drops or the conversion rate flips, the system raises an alarm.
Output
Dashboards
Numbers you trust in the meeting.
Audiences
Segments that are actually right.
Automations
Less manual work, fewer errors.
Business decisions
Decisions on solid ground.
Free checks · no login
Check your setup before you talk to us.
Tracking Check
One URL, one scan: you see what share of your conversions reaches Google Ads and Meta, and where the signals get lost along the way.
Consent Loss Calculator
What your consent rate costs in conversions, in euros per year.
Open the tool →EU AI Act Quick-Check
Five questions to the risk class of your AI use case, duties included.
Open the tool →GDPR Stack Calculator
Which analytics tool fits: Plausible, Piwik PRO, Matomo or GA4.
Open the tool →All resources
Checks, methodology and blog articles in one place.
Overview →Who works here
The person who sells it, builds it.
No account manager as a firewall, no juniors sold as seniors. The person who plans your setup builds it.
Open Analytics · live
Our own stack runs in public.
Plausible Community Edition, self-hosted in the EU, with a public dashboard at /open. A transparency example for cookie-free analytics, not our standard recipe for every client.
Client projects are EU-first: EU servers, EU tools, EU contracts. US tools only with a DPA and SCCs, when the business question demands them.
datascale.de
livelast 30 daysVisitors
4,812
Pageviews
11,207
Cookies
0
Sample view · full dashboard at /open/
How the Audit Sprint runs
Ten working days to a loss rate per channel.
Not a generic analytics audit: we follow the conversion all the way into Google Ads and Meta.
Days 1–2
Access & inventory
Read-only on GA4, GTM web and server, Google Ads conversion actions, Meta Events Manager and your CMP. We list every conversion action and every event currently armed.
Days 3–5
Measuring the signal path
How much arrives? Blocked requests, consent rate, Enhanced Conversions coverage, Event Match Quality at Meta. The result is a loss rate per channel, not a hunch.
Days 6–8
Reconciling the counts
Ads against GA4 against the shop backend: duplicates, attribution windows, web/CAPI deduplication, missing transaction_id. This is where the three different revenue numbers get their explanation.
Days 9–10
Report & priorities
Findings with severity, effort in person-days, and a sequence. A 90-minute walkthrough with your team. Then you decide, with no follow-up obligation.
from €2,400 net · fixed price depending on scope
From €2,400 net, fixed price depending on scope (Audit Sprint / Audit Sprint Plus), two weeks. More complex setups with several brands or app, web and shop take three weeks. We settle that in the kickoff.
Yes. This website tracks cookie-free with Plausible Community Edition on our own EU server; the dashboard is public at /open/. Client projects are EU-first: EU servers, EU tools, EU contracts. US tools only with a DPA and SCCs, when the business question demands them.
Read-only is enough: GA4 as Analyst, GTM read access, consent tool read access, optionally Google Ads and BigQuery if an export is active. No admin rights, no production credentials.
For 200-hour discovery processes, when a brand name for the board slide is the point, for campaign execution (creative, media, ads ops) or a single 8-hour ticket. Better providers exist for all of that, and we name them.
No. On the paid side we are the data foundation your agency builds on, not its competitor. And if your in-house team is in place and the volume justifies it, in-house is almost always better. We build the foundation underneath and hand it over.
Architecture audit
In two weeks you know which numbers you get to keep.
We audit tracking, consent, data quality and reporting, and tell you which decisions your existing data can carry and which it cannot.
2 weeks
from kickoff to roadmap
5 layers
tracking, consent, data quality, reporting, stack
from €2,400
net, fixed price, no follow-up obligation