An attribution model is the rule that decides which ads, channels or touchpoints get credit for a conversion. In Meta ads, the model you pick for an ad set shapes the results you see and the people Meta's delivery looks for.
Why the model matters
A customer rarely buys after one ad. They may see a Reel, click a search ad and buy from an email. The model answers one question: who gets the sale? The same $200 order can be credited to one channel or split across three, and each answer changes your cost per result and your view of what works.
A model sets how credit is split. A separate setting, the attribution window, sets how long after an ad interaction a conversion can still count.
Let Madgicx's AI do the work
Its AI marketer finds what's working and prepares the next move. You approve.
Try for Free ($0 Trial)Common models
| Model | How credit is given | Best for |
|---|---|---|
| Last-click attribution | All credit to the final click before the conversion | Short buying cycles, simple reporting |
| First-click | All credit to the first click | Seeing what starts journeys |
| Linear | Equal credit to every touchpoint | A rough view of the whole path |
| Time decay | More credit to touchpoints closer to the sale | Longer buying cycles |
| Position-based | A larger share to the first and last touchpoints, the rest split in the middle | Journeys with a clear start and finish |
| Data-driven | A model estimates each touchpoint's share from past journeys | Accounts with enough conversions to learn from |
The last four split credit across several touchpoints, which is the idea behind multi-touch attribution. Credit that follows a click only is click-through attribution.
Which models Meta offers
Meta's attribution settings page describes three models you can use in an ad set:
- Standard credits conversions after impressions, clicks or engagements, within the windows you choose. Meta says it optimizes delivery for those windows and behaviors.
- Incremental optimizes for conversions a model predicts the ad caused, and reports on them.
- Custom lets you share attribution data from your own analytics tool, so delivery follows your logic.
Meta also says results can't be compared across ad sets that use different models in the overview table. Keep the model the same across ad sets you plan to compare, and use the compare attribution settings feature in Ads Manager to see one ad set under other windows. For the default setting Meta's Results page states, see the attribution window entry linked above.
How to choose a model
- Look at the buying cycle. One-day impulse buys can live with last-click. A month of research needs credit shared across the path.
- Count your channels. With only Meta ads, the platform's own model is usually enough. With search, email and social together, compare models side by side.
- Pick one source of truth for budget decisions and write down which model and window it uses.
- Check the result against revenue. A model is a rule for sharing credit, not proof of cause. Incrementality tests are what show whether sales would have happened without the ads.
Example. An online course creator sells a $200 course. A buyer clicks a Facebook ad on Monday, a Google search ad on Wednesday and an email on Friday, then buys from the email.
| Model | |||
|---|---|---|---|
| Last-click | $0 | $0 | $200 |
| Linear | $66.67 | $66.67 | $66.67 |
| Position-based (40/20/40) | $80 | $40 | $80 |
Each row adds to $200 (the linear row to $200.01 after rounding). Under last-click, Facebook looks useless for this sale. Under the other two it gets credit for starting it, which is why a creator who cuts Facebook on last-click numbers alone may see email and search results fall later.
In Madgicx. The One-Click Report and dashboards bring Meta, Google Ads, TikTok, Shopify and Klaviyo data into one place, so you can set each platform's credited results next to your store revenue.
FAQ
What is an attribution model in marketing?
It is the rule a platform or analytics tool uses to share credit for a sale or sign-up between the ads and channels a customer touched. Last-click, linear and data-driven are common examples.
Which attribution models does Meta use?
Meta offers standard and incremental attribution models, plus custom attribution that uses data from your own analytics tool. You choose them at the ad set level.
What is the best attribution model?
None fits every business. A short, one-channel funnel can use last-click. A longer path across several channels is better served by a model that shares credit. Whichever you choose, keep it consistent so trends stay comparable.
What is the difference between an attribution model and an attribution window?
The model decides how credit is split. The window decides how long after an ad click, view or engagement a conversion can be credited. You need both to read a result.
Does the attribution model change my ad delivery?
On Meta, it can. Meta says standard attribution optimizes delivery for the windows and behaviors you select, and incremental attribution optimizes for predicted extra conversions.
Related
- Related terms: Attribution window, Multi-touch attribution, Marketing mix modeling
- In depth: Meta ads revenue attribution
- Answer: What is ad attribution
Sources
- About attribution models and attribution settings, Meta Business Help Center
