Conversion attribution

Conversion attribution in marketing is a process of finding which channel led the user to perform a desired action (a Conversion ) on a website or in an app. It allows marketers and analysts to correctly assign value to a particular event, research how different touchpoints and channels impact each individual desired action.

There are several conversion attribution models:

  • Last-click
  • First-click
  • Position-based
  • Last-non-direct-click
  • Time-decay
  • Linear
  • Custom

It’s also possible to use custom models. They come in handy in case the standard ones won’t meet your specific business needs. These models allow you to adjust baseline models to fit how you approach attribution for different campaigns, channels and projects.

To measure conversion it’s useful to use model comparison tool. It’s a report that improves your conversion analysis by comparing different attribution models.

But you can also improve your analysis with an attribution report. This report makes it possible to analyze conversions and sources of conversions on your website. It lets you to better understand the user’s paths that led them to the conversion.

Find more information about the conversion attribution on the Piwik PRO blog:


  • Five things every marketer should know about web analytics in 2026

    Web analytics is changing fast. AI is moving from buzzword to actual business impact, privacy rules keep shifting on both sides of the Atlantic, and marketing teams are rethinking their tool stacks. What does this mean for analytics strategy in 2026? We asked industry experts to share their predictions.

  • first party data

    First-party analytics without consent: Your Digital Omnibus compliance guide

    The Digital Omnibus is the European Commission’s simplification initiative to modernize the EU’s digital rulebook and reduce consent fatigue. The framework would enable first-party analytics without consent when specific criteria are met, ending years of uncertainty about the use of legitimate interest for web statistics.