Data & dashboards

One source of truth for growth decisions

ADEX centralizes channel, MMP and CRM data in BigQuery and builds CPA, CAC, ROAS and LTV dashboards in Looker Studio, Metabase or Supabase — so the team decides from one set of numbers instead of ten.

The problem

Every platform ships its own dashboard, and none of them agree. Marketing quotes platform ROAS, finance quotes blended CAC, product quotes activation, and half the meeting is spent arguing about which figure is real.

The cost is not just wasted time. Decisions get anchored to whichever number is loudest, not the one that reflects the business — and the same disagreement resurfaces every month because nothing was reconciled at the source.

How ADEX runs it

  1. 01

    Centralize the sources

    Pipe Meta, Google, TikTok, MMP, CRM and backend data into BigQuery so everything lives in one warehouse instead of ten exports.

  2. 02

    Define the logic once

    Agree on how a customer, a conversion and a cost are counted across channels, MMP and CRM — then encode it, so the numbers are consistent by construction.

  3. 03

    Model the decision metrics

    Build CPA, CAC, ROAS, LTV, payback and new-vs-returning as modeled fields, separated from the tactical platform metrics.

  4. 04

    Surface in the right tool

    Expose the layer in Looker Studio, Metabase or Supabase depending on the team, with views for operators and for leadership.

  5. 05

    Add alerting

    Flag anomalies — spend spikes, conversion drops, tracking gaps — so problems surface before the monthly review, not during it.

Proof

  • Built on BigQuery with Looker Studio, Metabase or Supabase, using the stack the company already runs where possible.
  • The same measurement discipline ADEX applies to MMP attribution flows straight into the warehouse, so channel and app data reconcile.
  • Managed across $50M+ in ad spend and 18 countries — the modeling holds at scale.

Scope and timeline depend on how many sources you connect, the state of your tracking, and how much historical data needs backfilling. We scope that before committing.

What you get

Pipeline

  • BigQuery warehouse setup
  • Channel, MMP and CRM connectors
  • ETL and modeling logic

Dashboards

  • CPA / CAC / ROAS / LTV views
  • Operator and leadership layouts
  • New-vs-returning and cohort cuts

Operation

  • Anomaly alerting
  • Documentation and handover
  • Optional ongoing maintenance

When this is right

  • Your teams argue about which number is correct in every review.
  • You run multiple channels plus an MMP or CRM that never reconcile.
  • You want CAC, LTV and payback modeled once, consistently.

When it isn't

  • You run a single channel with clean native reporting — a warehouse is overkill.
  • Your tracking is fundamentally broken — fix attribution first, then centralize.
  • You want a static PDF report — this is a living decision layer.

Frequently asked questions

How do you centralize Meta, Google, TikTok and MMP data?
Each source is connected into a BigQuery warehouse, normalized with consistent logic for customers, conversions and costs, then modeled into CPA/CAC/ROAS/LTV fields exposed in a dashboard.
Which dashboard tool do you use?
Looker Studio, Metabase or Supabase depending on the team and stack — the warehouse is the same underneath, so the tool is a presentation choice, not a lock-in.
Why not just use each platform's dashboard?
Platform dashboards only see their own data and count it their own way. A central layer reconciles them so a single, defensible number reaches the review.
Can you connect our CRM and backend?
Yes — CRM and backend data are what turn platform metrics into business metrics like qualified pipeline and validated revenue, so they are part of the model.
Do you set up alerting?
Yes. Anomaly alerts for spend spikes, conversion drops and tracking gaps surface issues before they distort a monthly review.
How long does a dashboard build take?
It depends on the number of sources, the state of your tracking and any historical backfill. We scope a realistic timeline before starting.

Tired of ten dashboards that disagree?

Build my growth dashboard

Keep exploring