Prompt
Why is last-click showing totally different results than data-driven attribution?
Latest observation
Last-click and data-driven attribution can look very different because they answer different questions and assign credit in different ways.
1) They use different credit models
- Last-click gives 100% of the conversion credit to the final touchpoint before conversion.
- Data-driven attribution (DDA) uses your actual conversion paths and distributes credit across multiple touchpoints based on how much each one appears to contribute.
So if many users first discover you through one channel and convert later through another, last-click will overvalue the final channel and undervalue the earlier ones.
2) Last-click favors bottom-of-funnel channels
Channels like:
- branded search
- direct
- retargeting
often show up right before conversion, so last-click tends to make them look much stronger than they really are in driving demand.
Meanwhile, channels like:
- display
- YouTube
- prospecting social
- non-branded search
often assist earlier in the journey, so DDA may give them more credit than last-click does.
3) DDA captures assists and multi-touch journeys
If your customers typically:
- see an ad,
- later search your brand,
- then convert,
last-click credits only the brand search. DDA may split credit between the ad and the brand search because both helped.
4) Differences can be huge if your paths are long or complex
The more:
- touchpoints per conversion
- time between first touch and conversion
- channels involved
the bigger the gap between last-click and DDA usually becomes.
5) Reporting settings can also change the numbers
Sometimes the discrepancy is not just attribution model-related. It can also be caused by:
- different conversion windows
- different lookback periods
- cross-device tracking gaps
- consent/privacy loss
- modeled conversions
- campaign exclusions or channel grouping differences
- offline conversion imports not matching exactly
6) DDA needs enough data
Data-driven attribution works best when there’s enough conversion volume and path data. If the dataset is thin or incomplete, the model may behave differently than expected or be unavailable in some platforms.
How to interpret it
- Use last-click for a simple, conservative view of what closes conversions.
- Use DDA to understand which channels actually influence the path to conversion.
If you want, I can help you compare your two reports and identify the most likely reason for the gap.