Marketing Mix Model
New Customer Onboarding
Onboarding runs two workstreams in parallel, and the initial deployment can complete in as fast as two weeks.
| Step | Who | Typical time |
|---|---|---|
| Connect ad platforms | You | A few minutes — a couple of clicks per platform |
| Connect KPI data | You / your technical team | Shopify: about a minute. Database or data lake (typical for SaaS): a few hours by a technical team member, once |
| Additional manual data sources (optional — TV, podcast, influencer) | Your team | Days — depends on your own data readiness |
| Define measurement scope | You + Maxma | One working session plus a few follow-ups after evaluating the data — usually done within a week |
| Model build & validation | Maxma | About a week |
The steps overlap — data connections take minutes and the scope discussion runs while the data lands — which is how the wall-clock total stays around two weeks rather than the sum of the rows.
1. Data Collection
Most of the data arrives through integrations: connect your ad platforms via OAuth in the Maxma web app and Maxma pulls up to a year of history for the initial retrospective analysis — the same connections then feed the ongoing refresh, so this is a one-time setup, not a recurring chore. Outcome data comes from Shopify or read access to your data warehouse.
For channels without a direct integration (TV, podcast, influencer), the team agrees on a simple data drop during onboarding — what's available, at what granularity, and how often.
2. Defining What the Model Measures
The other workstream is a scoping session with your team: deciding what "performance" means for your business before the model is built. Here's an illustrative scope for a subscription business:
| Dimension | Definition |
|---|---|
| Business outcome | New subscription volume |
| Corresponding metrics | Cost per incremental new subscription; ROAS based on LTV per new subscription |
| Markets | Baseline: global as a whole. Split into NAMER, EMEA, and APAC (based on data volume) |
| Time period | September 1, 2023 – September 30, 2024 (initial retrospective window) |
Scoping also defines the tactic cuts — how channels split into the units you actually make budget decisions on:
- Automatic, from platform metadata — Maxma pulls campaign meta information from the ad platforms and differentiates dimensions out of the box: search vs. video vs. display vs. Demand Gen, pMax vs. non-pMax, lookalike vs. prospecting vs. retargeting.
- Custom, from your taxonomy — cuts can also be defined around how your team plans (for example, from your campaign naming conventions), so the model's units match your budget lines.
- Campaign-level drill-down (beta) — Maxma's latest algorithm advancement can go below the tactic level and surface which campaigns contribute most to a tactic's incrementality.
The Scope Isn't Frozen
Business outcomes, markets, tactic cuts, and channels can all be updated after the initial onboarding as your business evolves — the allowed frequency and scope of changes may depend on your tier. When you launch a new channel or a new tactic, it gets included in the measurement in days, not months — so the model keeps up with how you actually run marketing.