# Multi-touch attribution

> Any attribution model that splits credit across more than one touch.

Source: https://datacenter.so/glossary/multi-touch-attribution
Also called: MTA.

Multi-touch models divide a conversion between the touches before it. The common ones:

- **Linear:** equal shares.
- **Position-based:** usually 40 percent to the first touch, 40 to the last, 20 split across the middle.
- **Time decay:** more credit to touches closer to the sale.
- **Data-driven:** shares learned from converting and non-converting paths.

All of them can only split credit between touches the tool recorded. Ad views, cross-device journeys and offline touches are usually missing. See [attribution models explained](https://datacenter.so/guides/attribution-models-explained) for one path worked through each model.

## Related terms

- https://datacenter.so/glossary/data-driven-attribution
- https://datacenter.so/glossary/first-touch-attribution
- https://datacenter.so/glossary/last-touch-attribution