A statistical technique to measure the impact of each marketing channel on revenue, without cookies.
MMM uses historical data, typically weekly or monthly spend and revenue, to statistically tease out how much each channel contributed. It doesn't rely on user tracking (which breaks in a privacy-focused world). Outputs: contribution curves, saturation points, and cross-channel effects. Good MMM tells you how much more revenue you'd get from an extra $100k in each channel.
In a cookie-less, privacy-first world, click-level attribution is dying. MMM is making a comeback. It's expensive to do well (usually needs a data science team or an agency), but for businesses with real spend it's worth it.