GLOSSARY

Multi-touch Attribution (MTA)

MTA distributes conversion credit across touchpoints via algorithms — Shapley, Markov, logistic — at user level; best paired with MMM and incrementality.

Last updated:

Quick answer
Multi-touch attribution (MTA) assigns credit for a conversion across multiple marketing touchpoints along the buyer journey, using user-level data and either rule-based or data-driven models (Shapley value, Markov chains, logistic regression). Apple's privacy changes and cookie deprecation have degraded the identifiers MTA depends on, pushing mature programs to triangulate MTA with MMM and incrementality experiments.

WHAT IT IS

Algorithmic MTA commonly uses logistic regression, Shapley value, Markov chains, or survival analysis trained on individual-level journey data — paid media, email, site events, organic — stitched via deterministic or probabilistic identity. Outputs are channel-, campaign-, and creative-level credit that can feed bidding, budget, and journey decisions within days.

HOW IT WORKS

MTA's value has narrowed as privacy changes (iOS ATT, cookie deprecation, Chrome Privacy Sandbox, walled-garden opacity) degrade journey completeness. Modern practice uses MTA for in-platform tactical optimization, pairs it with incrementality tests on major channels, and reserves Marketing Mix Modeling (MMM) for board-grade strategic allocation.

WHEN TO USE

Use MTA when user-level data is available and complete, when tactical optimization cycles demand fast feedback, and when the team is equipped to validate signals against incrementality holdouts.

RELATED

SOURCES

Related questions.

What is multi-touch attribution?
Multi-touch attribution (MTA) is the practice of assigning credit for a conversion across multiple marketing touchpoints along the buyer journey. It uses user-level data — ideally resolved identity across devices — to model the contribution of each touch rather than crediting only the first or last interaction.
How does MTA work?
Rule-based MTA (linear, time-decay, position-based) assigns credit via a predefined weighting. Data-driven MTA uses statistical or machine-learning models — Shapley value, Markov chains, logistic regression — to estimate each touchpoint's contribution from the data, typically the most accurate option but also the most demanding on data.
Why has MTA become harder?
Apple's iOS 14+ privacy changes (App Tracking Transparency, Mail privacy protection), third-party cookie deprecation in browsers, and tightened consent requirements (GDPR, Quebec Law 25, PIPL, PDPL) have all degraded the user-level identifiers MTA relies on. Coverage of the customer journey has eroded since 2021.
What should replace or supplement MTA?
Marketing mix modeling (MMM) for aggregate, privacy-safe, channel-inclusive measurement; incrementality experiments (geo-holdouts, ghost ads, causal inference) to validate both MTA and MMM; and a consent-first CDP that restores first-party identity where possible. Triangulation beats any single model.
How does NUUN Digital implement MTA?
We implement MTA only alongside MMM and incrementality testing, never as a single source of truth. We set decision rules for when MTA, MMM, and experimentation disagree — so measurement serves spend decisions, not methodology debates.

Need this term in action?

Book a call