Attribution models

Users rarely convert from a single touchpoint. They might find you via a blog post, come back via branded search, see a retargeting ad, then finally convert on a phone call. Attribution models determine how each touchpoint gets "credit" for the conversion.

Why attribution matters for SEO

If your attribution model gives 100% credit to the last-click channel (often paid or direct), SEO looks worse than it is. Blog content that generates first-touch awareness gets no credit in last-click models.

The major attribution models

1. Last-click (last non-direct)

100% credit to the final touchpoint before conversion.

Pro: simple, measurable. Con: ignores everything before the final click. Disadvantages SEO, content marketing, social.

2. First-click (first touch)

100% credit to the first touchpoint.

Pro: rewards top-funnel awareness. Con: ignores everything after the first click.

3. Linear

Equal credit to every touchpoint.

Pro: simple, gives credit to everything. Con: treats a 5-second visit the same as a 10-minute research session.

4. Time decay

Credit weighted toward touchpoints closer to conversion. Older touchpoints get exponentially less credit.

Pro: reasonable model for short consideration cycles. Con: penalizes top-funnel content on long cycles.

5. Position-based (U-shaped)

40% to first, 40% to last, 20% distributed across middle.

Pro: values awareness + close equally. Con: arbitrary weighting.

6. Data-driven attribution (DDA)

Google Ads + GA4 offer ML-based attribution that assigns credit based on actual user journey data.

Pro: empirically-grounded. Con: requires sufficient data; opaque (you can't easily see why credit was allocated).

GA4 attribution

GA4 defaults to Data-Driven Attribution (DDA) where available. You can also see Last-click, First-click, Linear, Time decay, Position-based in the Attribution section.

Compare models side-by-side to see how SEO's contribution changes.

The multi-touch story for SEO

SEO's value often lies in first touches + middle touches. A typical journey:

1. First touch: Organic search, informational query finds your blog
2. Mid touch: Organic search, commercial query, considers your solution
3. Mid touch: Direct, user returns to your site by typing domain
4. Last touch: Paid branded search, finally clicks an ad with "buy now"

Last-click says: paid search. Multi-touch says: SEO contributed 75% of the path.

View-through vs click-through attribution

• Click-through: credit only for clicked interactions
• View-through: credit for impressions too (even without clicks)

For SEO, you usually only care about click-through. View-through is primarily a display-ad concept.

Cross-device tracking

Users start on mobile, finish on desktop. Without cross-device tracking, each device looks like a separate journey. GA4's user-ID feature + Google Signals help link devices.

Offline conversion attribution

If your conversions happen offline (phone calls, in-store visits, post-demo sales), marketing tools default to not seeing them. Options:

• Call tracking numbers that pass source data (CallRail, Invoca)
• Importing CRM conversions back into GA4 / Google Ads
• Server-side conversion API (Facebook, Google) for pass-through

Practical attribution strategy

1. Set a primary model for official reporting (often data-driven or linear for multi-channel businesses)
2. Show secondary views (last-click, first-click) to illustrate different stories
3. Track by channel group + funnel stage, not just last touch
4. Review attribution annually, don't change models frequently or comparisons break

Common misinterpretations

• Direct traffic is really what converts. Often direct = repeat customers whose first visit came from SEO months ago.
• Paid search is the most effective channel. Paid captures low-hanging fruit at the end of the funnel; SEO + content builds the funnel.
• Organic traffic isn't converting. Check attribution model. SEO's contribution might be invisible in last-click reporting.

Beyond attribution: incrementality

The real question: if you turned off this channel, would conversions drop? Attribution models estimate; incrementality testing measures. Tools like Haus, Measured.com, or in-market holdout tests measure real incrementality, often surprising.