# CheapMMM

> Free, no-code marketing mix modeling (MMM) and media mix modeling tool. Upload a CSV to get ROAS by channel, feature importance, carryover insights, and budget optimization. No login required. Results in seconds.

## What CheapMMM does

CheapMMM takes a time-series CSV (date + sales/outcome + per-channel spend columns) and runs a structurally-informed ML pipeline to estimate channel-level marketing effectiveness. It is designed for fast, directional budget guidance without requiring data science expertise, R, Python, or enterprise MMM tooling.

## ML model architecture

- Geometric adstock transformation per channel to model delayed/carryover media effects

- Hill-function saturation curves to model diminishing returns at higher spend levels

- Joint MAP (maximum a posteriori) estimation of all adstock and saturation parameters across channels simultaneously, with Bayesian priors for regularisation

- Ridge regression for channel contribution coefficients with adaptive regularisation based on data-to-parameter ratio

- Trend and Fourier seasonality controls (annual, semi-annual, quarterly cycles) to separate baseline sales patterns from media effects

- Laplace-approximated credible intervals on ROAS estimates to quantify uncertainty

- Automatic data diagnostics: collinearity detection between channels, sparse channel filtering, and adaptive regularisation warnings

- Also referred to as media mix modeling — both terms describe the same methodology

## Outputs

- Channel ROAS with 90% credible intervals (contribution divided by spend, with uncertainty bounds)

- Feature importance scores per channel

- Actual vs. predicted sales visualisation

- Carryover/adstock insights showing lagged media effects per channel

- Per-channel saturation curve parameters for budget optimisation

- Data quality warnings (correlated channels, sparse data, thin data)

## Input requirements

- CSV with columns: Date, Sales (or outcome metric), one or more spend columns (e.g. Meta_Spend, Google_Spend)

- Minimum 3–6 months of weekly or monthly data recommended

- Sufficient spend variation across channels for stable attribution

- Channels with fewer than 8 non-zero observations are automatically excluded

## Intended users

Small and mid-size marketing teams, solo marketers, DTC/ecommerce growth hackers, and performance marketers who need MMM-style channel attribution without hiring data scientists or paying for enterprise platforms (Nielsen, Analytic Partners, Adobe Mix Modeler). Ideal for teams seeking a low-cost, easy-to-use alternative to enterprise MMM/media mix modeling vendors.

## How CheapMMM differs from alternatives

- vs. Meta Robyn: no R required, no local setup, results in under a minute

- vs. Google Meridian: no Python or data science team required, no Google ecosystem dependency

- vs. PyMC-Marketing: no Bayesian modeling knowledge required, runs in seconds not minutes

- vs. enterprise MMM: free, instant, no sales process or minimum contract

## Appropriate use and limitations

CheapMMM produces directional decision-support outputs, not causal incrementality measurement. Results depend on data quality — high channel collinearity or limited spend variation reduces attribution stability, and the system will warn users when these conditions are detected. It is a pragmatic middle ground between naive regression and full Bayesian MMM.

## Technical stack

Next.js frontend deployed on Vercel. FastAPI/Python backend. No user login or data storage required.

## URL

https://www.cheapmmm.com

## Full technical reference

https://www.cheapmmm.com/llms-full.txt