# GEO measurement has to evolve

<span class="byline">by John Morabito · April 19, 2026 · 9 min read</span>

**TL;DR**
- GEO breaks last-click attribution. Many AI answers produce no click. When they do, Google AI Overview traffic appears as "direct."
- Prompt visibility is the new leading indicator. Track presence, sentiment, and salience across engines.
- Branded search demand is the lagging evidence. 58% correlation with a seven-day lag is the documented benchmark.
- Correlative studies are making a comeback. AI tooling makes them accessible.
- The hardest part is mindset. Brand marketing has always worked this way.

## Why traditional attribution does not work

AI-generated answers often produce no click. The value is the brand impression, the consideration-set entry, the corroboration.

When clicks do happen, Google AI Overview traffic shows up as "direct" in analytics. No UTM, no referrer, no channel tag.

Require every impression to produce a trackable click and GEO always looks like zero return. That conclusion is wrong.

## Prompt visibility is the new rank tracking

In SEO, rank tracking was a leading indicator. Subordinate to traffic and conversion.

In GEO, prompt visibility is often the end product. When the AI lists your brand and the user reads the answer, the brand impression is made without a click.

Track three dimensions:
- **Presence**. Are you cited.
- **Sentiment**. What is being said.
- **Salience**. How prominently you feature vs. competitors in the answer.

## Brand demand is the true evidence

Winston documented a 58% correlation between TikTok views and Google branded search with a seven-day lag for a vacation rental client.

The same mechanism applies to GEO. Repeat exposure in AI answers drives familiarity. Familiarity drives branded search. Branded search drives direct traffic and conversions.

Track prompt visibility as leading. Track branded search and direct traffic as lagging. Use correlative analysis to connect them.

## Correlative studies are making a comeback

Before 20 years of last-click dominance, marketing measurement was built on correlation and incrementality. AI tools now make this analysis accessible to marketing teams without dedicated data science.

The strongest GEO measurement stacks are hybrid: prompt-level tracking for tactics, correlative brand-demand analysis for strategy.

## Practical steps

1. **Establish a prompt tracking baseline.** Audit current visibility across ChatGPT, Gemini, Perplexity, AI Overviews.
2. **Track branded search volume.** Set up monitoring in Google Search Console for all brand-related terms.
3. **Monitor referral traffic where available.** Perplexity citations produce measurable clicks. Track but do not over-weight.
4. **Run regular correlative analyses.** Monthly, compare prompt visibility vs. branded search, direct traffic, conversions.
5. **Benchmark share-of-voice.** Relative visibility vs. competitors consumers are comparing you against is more actionable than absolute.

## The mindset shift

Not every impression produces a click. That is fine. Apple does not track which billboard sold which iPhone. They know their consumer drives down that highway, and they show up 100% of the time.

## FAQ

**Why does traditional attribution fail for GEO?**

AI answers often produce no click. When they do, traffic shows up as direct with no referrer.

**What is the leading indicator?**

Prompt visibility across presence, sentiment, salience.

**What is the lagging indicator?**

Branded search and direct traffic. 58% correlation with a seven-day lag.

**What are the practical steps?**

Baseline, branded search tracking, referral monitoring, monthly correlative analysis, competitive share-of-voice.