GrowthMad

What did each channel really contribute?

Marketing mix modelling with Google’s open-source Meridian. A statistical view of what each channel adds, including the ones nobody clicks on, and where the next dollar should go.

Ask about this What’s involved

Based in Perth. Delivered remotely across Australia and overseas.

Click-based attribution only credits what it can see. Television, radio, outdoor, sponsorship and a good part of social never produce a click, so they look worthless in a dashboard while quietly doing much of the work.

Marketing mix modelling takes a different route. It uses your history of spend and sales, week by week, to estimate statistically what each channel contributed, with no cookies involved. It used to belong to very large advertisers. Open-source tools such as Meridian have put it within reach of mid-sized ones.

I also build the tooling. My own modelling application sits on top of Meridian. It walks a team through mapping and checking their data, trains the model on cloud hardware, and shows the diagnostics, results and budget scenarios in a browser. Less time goes on spreadsheets, and the results stay open for your team to question after the read-out.

This is for you if

  • You spend meaningfully across several channels, some of them offline.
  • You have at least two years of weekly spend and sales data, or can assemble it.
  • The board wants evidence for the marketing budget.
  • Platform reports each claim credit for the same sales.

How the work runs.

  1. Feasibility

    A check on whether your data can support a model, before you pay for one.

  2. Data

    Gathering spend, sales and outside factors such as price, season and promotions into one clean set. My tooling checks it for gaps and mismatches before any modelling starts.

  3. Model

    Building and validating the model, including tests of how well it predicts periods it hasn’t seen.

  4. Read-out

    Contribution and return by channel, with the uncertainty shown.

  5. Scenarios

    What happens if budget moves, and a recommended split, in a planner you can try your own numbers in.

Deliverables

What you end up with

  • Feasibility assessment
  • Clean modelling dataset
  • Validated model
  • Channel contribution and return report
  • Budget scenarios
  • Refresh plan

Where AI comes in

Faster, with a person checking

The model is statistics, not a chatbot. I built the application around it with AI coding tools, working to written acceptance criteria with a separate review of every change. AI also helps prepare data. The outputs are then checked against what’s plausible in your market.

How I use AI

Tools I work in

  • Google Meridian
  • Python
  • BigQuery
  • Google Cloud

Fair things to ask.

Something else? Ask me directly.

Is our business big enough for this?

It depends more on data than on size. The feasibility step answers that cheaply. If the answer is no, I’ll say so and suggest something simpler.

How often should the model be refreshed?

A few times a year suits most businesses, or after a major change in spend.

Do you use your own software?

Yes. I’ve built a modelling application on top of Meridian that handles the data checks, model training, diagnostics and scenario planning. The statistical engine is Google’s open-source work, so the method is public and can be checked by anyone.

Does this replace attribution in GA4?

No. The two answer different questions. Attribution helps you tune campaigns week to week. A mix model informs how the whole budget is divided.

What decisions does a marketing mix model support?

How to split the budget between channels, whether to spend more or less in total, and what a change is likely to do to sales. Marketing mix modelling services are worth paying for when one of those decisions is live. I work as a Google Meridian consultant, so the method is open and can be checked.

Advertising

Media planning and strategy

Media planning and buying advice from a former media agency strategist: channels, budget, audiences and expected results. Forecasts come as ranges with the conditions attached.

Read more

Data

Dashboards and reporting

Looker Studio dashboards built on clean, reconciled sources and designed around the questions your team asks every month.

Read more

Marketing

Digital marketing strategy

Digital marketing strategy from a consultant who still does the hands-on work: channel priorities, budget split and a 90-day plan, built from your numbers.

Read more

All data services

Something driving you mad?

Tell me what’s going on. I’ll come back with questions or a scoped proposal, and I’ll say so if I’m not the right person for it.

Get in touch