Data Foundation / Measurement infrastructure

The numbers look defensible.

You know which ones aren't.

Tagging, analytics, server-side tracking, attribution logic. Every number above them depends on this layer. Usually half-built.

Request an audit→

scroll

How we think

Measurement is where marketing stops being a matter of opinion.

Done well, nobody thinks about it. Done badly, every conversation about budget, channel mix, and acquisition cost starts with a disclaimer.

Most stacks sit between the two. Analytics was configured once and never audited. Server-side tracking was scoped but never finished. Consent banners run, but nobody can tell you what signal they're costing. The CRM keeps its own numbers. The numbers don't match the platforms. Someone on the team knows. The people they report to don't.

We rebuild the layer everything else depends on.

What we build

The tracking and measurement layer every channel depends on.

01 · Tagging plan

Tagging plan & data layer that survives the trip to the CRM.

Where one exists, it was built for the analytics tool and forgotten. Where it doesn't, we start there. Either way, we build from the conversion paths that matter to the business: demo starts, demo completes, pricing views, pipeline signals worth bidding on. Implemented through GTM. UTMs structured so campaign, source, and campaign ID land intact in the CRM.

the foundation everyone skips

tagging-plan.csvv2.4

Event name

Key param

Platforms

account_created

action=confirmation

all platforms

subscription_payment

type=invoice-ok

all platforms

page_view_training

path=/cours

GAds · GA4

contact_form_submit

form submit

GA4 only

02 · Server-side tracking

Server-side tracking, wired to every paid channel.

Spinning up a container takes an afternoon. Wiring every platform through it, one at a time, is the work. CAPI for Meta, Enhanced Conversions for Google, LinkedIn CAPI, TikTok Events API. Each has its own payload requirements and its own way of losing signal in transit.

03 · Consent architecture

Consent Mode v2 without breaking measurement.

A consent banner does its compliance job and breaks measurement at the same time. We configure the consent platform, wire it through Consent Mode v2, and let modelled conversions fill the gap where users decline. Bidding algorithms keep getting the signal density they need. Compliance and measurement stop being a trade-off.

04 · Offline conversion pipelines

Offline conversions and value-based bidding, back to the platforms.

Your ads learn from form submits. Your business learns from closed deals. The two need to meet. Offline conversion imports, Enhanced Conversions for Leads, value-based bidding setup. Campaigns stop optimising for proxies and start optimising for outcomes. In B2B, this is the layer most stacks never finish.

ad spend learns from closed deals

05 · CRM data injection

CRM integration, source and campaign intact.

Lead forms firing is the easy part. Getting each lead into the right CRM object with source, campaign, campaign ID, and UTM intact is where most builds break. Fixed here, not left as a data team problem.

UTMs survive

lead · created● live

Form submitdemo_request

→

CRM · Opportunitystage: SQL

utm_source=intactutm_campaign=intactutm_id=intactutm_content=intactgclid=intact

The half-built layer

When measurement is half-built,
every decision above it is too.

Events tracked

What the business actually cares about, or what was easiest to wire up.

Server-side coverage

Every paid channel wired through, or only the platform someone had time for.

Consent and recovery

Modelled conversions filling the gap, or signal density just collapsing.

CRM round-trip

Closed deals feeding the bidding, or stuck in the spreadsheet.

Tracking health · stack diagnostic4 layers· 3 issues

events

server-side

consent

crm-sync

Events tracked

Degraded7 / 17

Server-side coverage

Critical1 / 4 ch

Consent and recovery

Degraded−31%

CRM round-trip

FailingIDs dropped

auditevents×coverage×consent×crm=trust↑ Scaling ad spend on this is just lighting money on fire.

What changes on your side

Once measurement is wired, the conversations change.

Numbers you can defend upward, whether the next meeting is with a founder, a board, or a finance lead.

A tagging plan that survives the next platform migration without rebuilding from scratch.

Server-side architecture that recovers the conversions ad blockers and browser restrictions are eating.

Reporting that keeps post-click and post-view honest, with platform self-reporting stripped back to something usable.

Campaigns that bid toward actual business outcomes, not just form submissions.

Start here

The problem is clear.
The scope rarely is.

Data Foundation Audit

Typically three weeks, scoped to your stack. The deliverable is yours whether we continue or not.