Measurement you own — not a vendor dashboard

1. Home
2. Services
3. MMM-Lite Engagement

Privacy-first marketing measurement, built on your data in six weeks. A media-mix model, a live incrementality test, a self-serve dashboard — and the playbook your team runs afterward. You keep everything, including the code.

Seed–Series B teams ready to trade last-click for real answers.

Seed–Series B teams spending roughly $15–20K+/month across 3+ channels, with 26–52 weeks of history, who no longer trust last-click attribution — and who want answers, not a vendor dependency. If your spend has barely varied or your history is thinner, start with the free Starter Kit and the Marketing Budget Planner; the sprint will be worth more in two quarters.

A six-week Diagnose → Model → Test → Handoff cycle.

1. 01 Week 1

   Diagnose

   Data audit against the schema, gap fixes, promo/launch flagging, unit-economics inputs locked.

2. 02 Week 2

   Model v1

   Adstock + diminishing-returns regression with out-of-sample validation; first mROAS read.

3. 03 Week 3

   Design

   Sensitivity runs, channel deep-dive, reallocation plan (10–25% moves), test selection.

4. 04 Week 4

   Test launch

   Geo-holdout incrementality test, pre-registered and locked before spend moves.

5. 05 Week 5

   Dashboard

   The self-serve "what happens if I add $1K to…" view, wired to your data.

6. 06 Week 6

   Handoff

   Decision-rules playbook, team walkthrough, monthly refresh procedure. The code is yours.

7. 07 Wks 8–10

   Test readout

   Async: verdict on the live test + final reallocation memo (included).

Six deliverables. All yours.

Model + code Adstock + diminishing-returns regression, documented and yours to keep — no vendor lock-in.

Live incrementality test Geo-holdout test, pre-registered before spend moves — no peeking, no goal-moving.

Self-serve dashboard The "what happens if I add $1K to…" view, wired to your data.

Reallocation decision rules Small, reversible moves (10–25%). Never zero a channel on model evidence alone.

Monthly refresh procedure The steps your team runs each month to keep the model honest.

Test readout memo The verdict on the live test plus the final reallocation memo, delivered async in weeks 8–10.

Illustrative of the method — measurement outcomes vary by data quality and spend patterns.

Three rules that keep the work trustworthy.

Out-of-sample validation only

No flattering fit statistics — the model is judged on data it has never seen.

The test outranks the model

Wherever the live incrementality test and the model disagree, the test wins.

Small, reversible moves (10–25%)

Never zero a channel on model evidence alone. Learn from measured shifts, not swings.

Fixed fee. Fixed scope. Weekly standups.

Everything in "What you get" is included — model, code, dashboard, test, and playbook. No lock-in, no recurring seat fees.

The fixed fee is published on the pricing page. The scope is confirmed on the call, once your data readiness is checked.

Take them free — or have us run them with you.

These are the same frameworks we use inside the sprint.

• Free kit MMM-Lite Starter Kit Open
• Free template Marketing Budget Planner Open
• Free tool Paid Media Budget Allocator Open
• Playbook Paid Media Budget Allocation Open
• Blog MMM-Lite for Startups Open
• Blog Incrementality Testing for Startups Open

MMM-Lite Engagement — FAQ

What if we don't have 26 weeks of clean data?

Week 1 exists for exactly this — most teams' data needs assembly, not creation. If it's genuinely too thin, we'll tell you on the scoping call and point you to the free kit until it isn't.

Do we keep the model?

Yes — code, dashboard, playbook, everything. The sprint's goal is that you don't need us afterward.

MMM vs our attribution tool?

They answer different questions. Attribution credits touchpoints per user and degrades as tracking signal disappears; MMM reads aggregate spend-to-outcome patterns with no user tracking, and answers the budget question directly.

What does the incrementality test require?

One priority channel, a geo split we design together, and 4–8 weeks of patience. It's pre-registered before launch — no peeking, no goal-moving.

Own the measurement system.

Six weeks from kickoff to a live incrementality test, a working dashboard, and the playbook your team runs on their own.