The common types
Last-click gives all credit to the final touch before the sale. First-click gives it all to the first. Linear splits it evenly. Time decay gives more credit to recent touches, and position-based favors the first and last. Data-driven models use your own conversion data to estimate each channel's real contribution.
Where you see it
Google Ads and Google Analytics 4 now use data-driven attribution as the default, and Google retired most of the older rule-based models in 2023. Ad platforms like Meta also report conversions using their own attribution settings.
Why platforms disagree
Each platform usually gives itself credit for conversions it touched. Add up the numbers from every platform and you often get more sales than you actually made. That's a sign of overlapping credit, not extra revenue.
Why Attribution Model matters
Your attribution model shapes where you spend money. A last-click view can make brand search and retargeting look like heroes while starving the ads that first introduced customers to you. Understanding the model helps you avoid cutting what actually drives growth.
The channel that looked useless
An ecommerce brand nearly paused its video ads because they showed few last-click sales. A data-driven view and a short holdout test showed those ads started many of the journeys that later ended in branded searches. The brand kept the video budget and adjusted how it reported results.
Common mistakes
Trusting every platform's numbers at once
Platforms each claim credit. Use one source of truth for total revenue and compare channels within it.
Changing models mid-report
Switching attribution settings changes how numbers look overnight. Note the change so you don't mistake it for a real shift in performance.
Treating attribution as proof
Attribution shows correlation along a path, not true cause. Holdout and incrementality tests are the way to confirm what a channel really adds.
Where we work on this