Attribution Models
Last updated: April 9, 2026
Definition: What are attribution models?
An attribution model defines which touchpoint in the customer journey a conversion – meaning a purchase, a sign-up, or another desired action – is credited to. In most cases, a shopper interacts with several ads before buying: they see a display ad, click a Sponsored Products ad, return through a Google search, and finally make the purchase. The attribution model determines who gets the “credit” for that purchase.
The choice of attribution model has a direct impact on how campaigns are evaluated: the same budget can deliver completely different performance figures depending on the model. Anyone who wants to assess their advertising channels fairly and allocate budgets sensibly needs to understand which model their platform uses – and where its limits lie.
Table of contents
Last click attribution
With last click attribution (also called last touch), the final touchpoint before the purchase receives 100% of the credit. If a user first sees a display ad, then clicks a Sponsored Brands ad, and finally buys through a Sponsored Products ad, the Sponsored Products campaign alone is credited with the sale.
Last click: 100% of the credit → last click before the purchase
Advantage: Simple, clear, easy to follow. Well suited when the customer journey is short (e.g., impulse buys).
Disadvantage: Systematically overvalues lower-funnel channels and undervalues awareness and consideration touchpoints. In a last-click model, display and video campaigns often appear “ineffective” even though they were what sparked the purchase in the first place.
First click attribution
With first click attribution, the first touchpoint receives 100% of the credit. In the example above, the display ad would be credited with the entire sale because it established the first contact.
First click: 100% of the credit → first contact with the brand
Advantage: Shows which channels bring new customers into the funnel. Valuable for brands that prioritize new-customer acquisition.
Disadvantage: Ignores everything that happens after the first contact. Retargeting and conversion campaigns are systematically undervalued.
Other attribution models at a glance
Linear attribution
Every touchpoint receives an equal share of the credit. With three touchpoints, each gets 33.3%. Advantage: all channels are accounted for. Disadvantage: no differentiation by impact.
Time decay attribution
Touchpoints that are closer in time to the purchase receive more credit. The last contact gets the most, the first gets the least. Well suited for products with long decision cycles.
Position-based (U-shaped) attribution
The first and last touchpoints each receive 40% of the credit, while the touchpoints in between share the remaining 20%. A good compromise between first click and last click.
Data-driven attribution
Algorithms analyze the actual conversion paths and distribute the credit based on each touchpoint's measured influence. The most accurate model, but it requires large volumes of data. Multi-Touch Attribution builds on this approach.
How Amazon assigns conversions
For Sponsored Ads (Products, Brands, Display), Amazon uses a last click attribution model with a 14-day lookback window for clicks and a 14-day window for View-Through Conversions. A purchase is credited to the last clicked ad, provided the click occurred within the past 14 days.
Amazon's attribution in detail
- • Sponsored Products / Brands: Last click, 14-day window, click-through only
- • Sponsored Display: Last click + view-through, 14-day window
- • Amazon DSP: Last touch (click or view), configurable windows (1–90 days)
- • Amazon Attribution: Measures external traffic sources (Google, Meta, etc.) and their influence on Amazon sales
One important note: because Amazon only considers the last interaction, double counting can occur when a user clicks ads from different sources. DSP, Sponsored Products, and Amazon Attribution each have their own attribution – so a single sale can theoretically show up in multiple reports at the same time. Tools like Amazon Marketing Cloud help resolve these overlaps.
Last click vs. first click: which model when?
Last click is a good fit when…
- • … you have short customer journeys (impulse buys, low-priced products)
- • … you primarily manage lower-funnel campaigns (Sponsored Products, retargeting)
- • … you need a simple, easy-to-follow attribution
First click is a good fit when…
- • … you want to evaluate the value of awareness campaigns
- • … new-customer acquisition matters more than short-term ROAS
- • … you have long decision cycles (high-priced products)
In practice, neither of the two single-touch models is perfect. The most honest answer comes from multi-touch models or data-driven attribution – but these require more data, more tools, and more analytical expertise. For day-to-day Amazon Advertising, the most important thing is to know that Amazon uses last click, and to factor that knowledge into how you manage your campaigns.
Frequently asked questions (FAQ)
Can I change the attribution model on Amazon?
Not in the standard reporting of the Advertising Console – Amazon fixes it to last click with a 14-day window. In Amazon DSP, the lookback windows can be adjusted. Anyone who wants to use a different model can apply their own attribution logic to raw data via Amazon Marketing Cloud.
Why can a single sale appear in multiple reports?
Because Amazon Sponsored Ads, DSP, and Amazon Attribution each have their own attribution. If a user clicks a Sponsored Products ad and then a DSP ad, the sale is counted in both reports. The total number of attributed sales is then higher than the actual number. AMC solves this problem with deduplicated attribution.
What does the 14-day window mean in practice?
If a user clicks your ad and buys within 14 days, the purchase is credited to that ad. If they only buy on day 15, the purchase no longer appears in the advertising reporting – even if the ad provided the initial impulse. For products with long decision cycles, the 14-day window underestimates the true advertising value.
Related terms
Understand your advertising impact better?
With HORAiZON ONE you analyze your Amazon campaigns based on real data and make well-founded budget decisions.
Try it for free now