Email Multi-Touch Attribution

Definition

Email multi-touch attribution (MTA) is a family of attribution methods that distribute credit for a conversion across all the marketing touchpoints that preceded it, rather than assigning the entire conversion to a single touchpoint. It recognises that a purchase often follows several interactions — an ad, a search, and multiple emails — and attempts to quantify each one's contribution. Multi-touch attribution gives email a more realistic share of revenue credit.

How It Works

Multi-touch attribution maps a customer's journey before a conversion and divides credit among the touchpoints along that path. The key difference from single-touch models is that multiple interactions receive credit.

• Journey tracking — the model records each touchpoint, including email opens, clicks, ad views, and site visits, before conversion.
• Credit distribution rules — different MTA models divide credit differently, such as equal linear credit, decaying credit, or position-based weighting.
• Email's share — email receives credit proportional to its role in the journey, which is often more accurate than last-click alone.

Multi-touch attribution is a core tool within email-attribution and is especially valuable for journeys with longer sales cycles, where a single final email did not do all the work.

How It Works in Practice

• Linear attribution — every touchpoint receives equal credit, so a conversion with four touches gives email 25% for each email touch.
• Time-decay attribution — touchpoints closer to the conversion receive more credit, favouring the emails sent late in the journey.
• Position-based attribution — the first and last touches receive the largest shares, with the remainder split among the middle touches.

The choice of rule changes email's credited revenue, so the model should be selected to match how the business actually converts.

Best Practices

• Match the model to the sales cycle — long B2B cycles often suit time-decay or position-based models, while short B2C cycles may suit simpler rules.
• Validate against experiments — where possible, compare MTA results with holdout-based lift to calibrate the model.
• Use consistent tracking — MTA is only as good as the touchpoint data feeding it, including click-through-rate signals.
• Review credit distribution regularly — as the marketing mix changes, the model should be re-examined.

Example

A customer sees a display ad, opens a newsletter, and clicks a promotional email before purchasing £200. Under last-click, the promotional email gets all £200. Under a linear multi-touch model across three touchpoints, email receives two-thirds of the credit, or roughly £133, a more realistic reflection of email's contribution.