Multi-Channel Attribution: Common Models and Best Practices
Multi-channel attribution credits every touchpoint on the path to a sale. The common models, how they differ, and five practices for using them well.
Every conversion has a history. A customer might see a social ad, read a review, get an email and finally search for your brand before buying. Understanding what each of those touchpoints contributed is central to spending a marketing budget well.
Multi-channel attribution is the method for doing that. It assigns value to each interaction that led to a conversion, instead of crediting a single one. The idea is simple, but applying it well takes some care. Here are the fundamentals, the common models and the practices that make attribution useful.
What multi-channel attribution does
Multi-channel attribution estimates how much each marketing channel and touchpoint contributed to an outcome, such as a sale or a lead.
Single-touch models credit only one interaction, usually the first or the last. Multi-channel attribution acknowledges that several touchpoints were involved and looks at the whole journey. That helps marketers see which channels contribute most to conversions and make better decisions about budget and strategy.
The common models
- First-touch attribution credits the first touchpoint.
- Last-touch attribution credits the last touchpoint before conversion.
- Linear attribution splits credit equally among all touchpoints.
- Time-decay attribution gives more weight to the touchpoints closest to the conversion.
- U-shaped attribution gives the most credit to the first and last touchpoints and spreads the rest across the ones in between.
Each model answers a slightly different question, which is why the choice of model can change which channels look strongest. For a fuller explanation, see What is multi-touch attribution?
Best practices
- Define clear goals. Before choosing a model, decide which outcomes and key performance indicators matter to the business.
- Use more than one model. Comparing several models gives a more complete understanding of how channels contribute than relying on any one.
- Account for complex journeys. Customer journeys can involve many touchpoints across many channels. Your approach should reflect that rather than flatten it.
- Test and iterate. Keep checking whether your model reflects the real impact of each channel, and adjust it when it doesn't.
- Integrate data sources. Bring together data from every channel and platform you use to get a complete view of the journey.
Where click-based attribution falls short
Attribution models built on tracking individual clicks have a blind spot: they can only credit what they can track. Channels people don't click, like TV and podcasts, and walled gardens that keep their user data to themselves, are hard for them to see. Privacy changes such as Apple's App Tracking Transparency have made those gaps larger.
Provalytics approaches the problem differently. Our methodology measures all media, including walled gardens, using aggregated, non-personal marketing and conversion data, without website tags or personal data.
The principles in this guide still apply: clear goals, more than one view, every channel included. Measuring that way, without depending on tracking individuals, is what makes the answer complete.
Frequently asked questions
What is multi-channel attribution?
It is a way of measuring how much each marketing channel and touchpoint contributed to a conversion. Unlike single-touch models, it recognizes that several interactions usually lead to a sale and assigns value to each one.
Which attribution model is best?
No single model is right for every business. First-touch shows what starts journeys, last-touch shows what closes them, and linear, time-decay and U-shaped models spread credit in between. Comparing several gives a fuller picture than relying on one.
What is U-shaped attribution?
U-shaped, or position-based, attribution gives the largest shares of credit to the first and last touchpoints and splits the rest among the interactions in between.
Can multi-channel attribution measure TV or walled gardens?
Click-based attribution struggles with both, because it can only credit touchpoints it can track. Methods that use aggregated spend and sales data, such as marketing mix modeling, can measure them without tags or personal data.