Head-to-headTOFU
Last-click vs data-driven attribution
The short answer
Last-click attribution gives all the credit to the final touchpoint before conversion. Data-driven attribution uses conversion path data to share credit across touchpoints. Last-click systematically over-credits bottom-of-funnel channels like brand search and retargeting. Data-driven is less biased, but it still describes correlation within the tracked journey — not what would have happened without a channel.
Side by side
| Last-click | Data-driven | |
|---|---|---|
| Credit logic | Final touch only | Modelled across touches |
| Bias | Toward bottom-funnel channels | Less, but model-dependent |
| Data needs | Minimal | Enough conversions and paths |
| Transparency | Easy to understand | Harder to explain |
| Incrementality | No | No |
A worked example
A customer sees a YouTube ad, later clicks a Meta ad, then searches your brand on Google and buys. Last-click gives all the credit to brand search. Data-driven attribution in GA4 may share credit across the Meta click and the Google click based on patterns across many journeys — but it may miss the YouTube view entirely. Neither can tell you whether the sale would have happened without any of the ads.
What to use each for
- Last-click — quick sanity checks, simple single-channel accounts
- Data-driven — allocating budget between campaigns within a platform
- Blended MER — judging whether marketing overall is working
- Holdout tests — deciding whether a channel is truly incremental
How each model treats a typical journey
| Touchpoint sequence | Last click credit | Data-driven credit (illustrative) |
|---|---|---|
| Instagram ad view → YouTube video → Google brand search → purchase | 100% Google brand search | Shared across Instagram, YouTube and search based on observed patterns |
| Google generic search → email → direct → purchase | 100% direct or email | Shared, with more to generic search if paths show influence |
Strengths and weaknesses
| Last click | Data-driven | |
|---|---|---|
| Simplicity | Very simple | Complex, less transparent |
| Upper-funnel credit | Very little | More |
| Data requirements | Low | Needs volume |
| Bias | Toward closing channels | Toward tracked touchpoints |
| Best use | Comparison, diagnostics | Default optimisation |
Going beyond attribution
- Track blended MER and new-customer CAC
- Run geo holdouts for major channels
- Use platform lift studies where available
- Compare attribution models to understand ranges
- Consider marketing mix modelling at scale
Pros and cons of each
Last-click
- Pros: simple and stable
- Cons: ignores earlier touchpoints, undervalues awareness
Data-driven
- Pros: credits multiple touchpoints using data
- Cons: a black box, platform-specific, still not incrementality
Example: rebalancing upper-funnel spend
Last-click reports undervalued YouTube. Data-driven attribution and a geo-holdout test showed its real contribution, justifying a higher budget.
What Indian brands should do
With users moving between Instagram, YouTube, Google, WhatsApp and marketplaces, last-click reports often over-credit search and retargeting. Use data-driven attribution for day-to-day optimisation, blended MER for business health and periodic holdout tests for big decisions.
Decision checklist
- Enough conversion volume? Data-driven attribution works better
- Big upper-funnel spend? Validate with experiments
- Multiple platforms claiming credit? Use blended metrics
Frequently asked questions
Which does GA4 use?
Data-driven attribution is the default for key events in GA4 where eligible.
How do we measure real impact?
With incrementality tests and blended metrics like MER.
Why do Google Ads and GA4 show different numbers?
They use different attribution scopes, windows and data. Reconcile each against your back-end sales.
Should we switch GA4 back to last click?
Data-driven is usually a better default for optimisation. Use last click as a comparison lens, not the only truth.
Why do Meta and GA4 disagree on attribution?
They use different data, windows and models. Meta includes view-through conversions; GA4 relies on tracked sessions.
Is data-driven attribution accurate?
It's more nuanced than last click, but it can't see untracked influence. Validate with incrementality tests.
Does data-driven attribution prove incrementality?
No — it distributes credit; only experiments show causal impact.
Is GA4's data-driven attribution reliable?
It's useful for relative comparisons but not proof of causation.
More comparisons
All comparisonsServices mentioned
Terms used on this page
Talk to a strategist
Thirty minutes. Your numbers. A straight answer.
Book a strategy call with the people who'd actually run your account. We'll tell you what we'd do — or that you don't need us yet.