Quick Answer: Attribute revenue to AEO by combining two signals: UTM-tracked AI referral traffic in GA4 (a hard floor that misses most zero-click influence) and a "How did you first discover us?" survey option for AI assistants, blended through a multiplier model and reported as a range of plus or minus 25%. Divide estimated AEO revenue by total citations to get cost and revenue per citation: mature B2B SaaS programs reach under $75 cost per citation and above $150 revenue per citation, at least a 2x ratio, by months 6-12.
Every AEO engagement hits the same wall. You're three months into optimizing for AI visibility. Share of Voice is climbing. Citations are up. Sentiment has shifted from neutral to positive. Then the CFO walks in and asks: "What is this actually producing?"
If you can't answer with a number, the budget gets pulled. We've seen it happen. Great AEO programs killed at the quarterly review because nobody could connect the visibility work to revenue.
We built an attribution methodology to solve this. It's not perfect (no attribution model is), but it produces a defensible estimate that has survived CFO scrutiny at every client engagement we've run it on. Here's how it works.
Why Last-Click Attribution Fails for AEO
The core problem: AEO influence is mostly invisible to traditional analytics.
When ChatGPT recommends your product, the buyer doesn't click a tracked link. They hear your brand name, remember it, and either type your URL directly or Google your brand name a week later. Your analytics sees "Direct" traffic or "Branded Organic." The AI influence never shows up.
Even when users do click links in AI responses, mobile apps strip referrer headers. Desktop web clicks from ChatGPT carry referrer data, but that's maybe 30% of the total. The rest disappears.
This creates a measurement gap that kills budgets. The spend is visible; the return is not.
Our Two-Signal Approach
We use two imperfect data sources and combine them into one estimate. Neither signal is sufficient alone. Together, they triangulate something defensible.
Signal 1: UTM-tagged AI traffic. GA4 can capture sessions where the referrer header comes from an AI engine (chat.openai.com, perplexity.ai, etc.). We set up a custom "AI Search" channel group in GA4 for every client. This gives us a hard floor: tracked sessions, tracked conversions, tracked revenue. Real numbers, but a significant undercount.
Signal 2: HDYHAU survey responses. We add a "How did you first discover us?" question to the client's signup or post-purchase flow. The key option: "ChatGPT / AI assistant." This catches what UTMs miss, including zero-click recommendations, mobile app referrals, and delayed visits that happen days later.
The gap between these two signals is the insight. UTMs show what tracking can see. HDYHAU shows what buyers actually remember. The difference is the invisible AI influence.
The Math: Multiplier Model
The formula is straightforward:
tracked_ai_revenue = chatgpt_utm_revenue + (hdyhau_ai_pct x total_revenue)
multiplier = hdyhau_ai_responses / chatgpt_utm_sessions
estimated_aeo_revenue = tracked_ai_revenue x multiplier
What this means in practice: Take a B2B SaaS client doing $500K in quarterly revenue. GA4 tracked 45 sessions from AI sources, producing $3,200 in direct revenue. Their HDYHAU survey (400 respondents) showed 8% selecting "AI / ChatGPT." That 8% applied to total revenue gives us $40,000 in HDYHAU-attributed revenue. Combined with the tracked $3,200, total tracked AI revenue is $43,200.
The multiplier (32 HDYHAU responses / 45 UTM sessions = 0.71) tells us whether our tracking is catching more or less than the survey. When it falls below 1, we floor it at 1, meaning the estimate is at least the tracked number. In most engagements, the HDYHAU signal is doing the heavy lifting early on, and the multiplier only kicks in meaningfully once you have enough UTM volume to produce a ratio above 1.
We present results as a range (typically +/- 25%) rather than a point estimate. For the example above: $32,400 to $54,000 in AI-attributed revenue. This communicates appropriate uncertainty while still giving leadership a number to work with.
From Estimate to Cost Per Citation
Here is where the model gets operationally useful. Once you have estimated AEO revenue, you can calculate cost per citation:
cost_per_citation = total_aeo_spend / total_citations
revenue_per_citation = estimated_aeo_revenue / total_citations
This is the metric that unlocks budgets. When you can show a CFO that each citation costs $62 but generates $180 in attributed revenue (2.9x return), AEO stops being "that AI visibility thing we're trying" and becomes a channel with measurable economics.
We track cost per citation monthly alongside Share of Voice and the other AEO KPIs. As programs mature, cost per citation should decrease while revenue per citation holds steady. If it doesn't, something in the strategy needs adjustment.
Validation: The Correlation Model
After 12 weeks of data, we layer in a second model to validate the Multiplier Model. This one uses regression analysis:
Revenue ~ SOV + Paid_Spend + Seasonality
OLS regression tells us: "For every 1% increase in Share of Voice, revenue changed by $X, after controlling for other factors." When this model agrees with the Multiplier Model, confidence goes up. When they diverge, we investigate.
We blend the two models 50/50 when both are available. If the regression has an R-squared above 0.6, we weight it more heavily. The blended estimate is what goes into the quarterly review.
What "Good" Looks Like
After running this framework across multiple B2B engagements, here are the patterns we see:
- Month 1-3: Cost per citation is high (often $100-200+). You're investing in content updates, outreach, and monitoring setup. Revenue attribution is mostly HDYHAU-driven because UTM volume is low.
- Month 4-6: Citations compound. Cost per citation drops as existing placements keep generating mentions. The Correlation Model becomes available and starts validating the Multiplier Model.
- Month 6-12: Mature programs typically see cost per citation below $75 and revenue per citation above $150 for B2B SaaS with $10K+ ACV. The ratio should be at least 2x.
The leading indicators that predict future revenue: SOV trajectory (is it climbing?), citation quality (are high-authority domains citing you?), and sentiment shift (are AI engines recommending you, or just mentioning you?).
How to Present This to Your CFO
We've learned a few things about selling AEO attribution internally:
Lead with the framework, not the number. Walk through the methodology first. CFOs respect the logic of "we track what we can see, survey for what we can't, and triangulate an estimate with a confidence range." It's honest and auditable.
Show the range. A single number invites scrutiny. A range with a stated methodology invites discussion. "AEO contributed between $32K and $54K this quarter" is stronger than "AEO contributed $43K."
Benchmark against other channels. If your paid search CAC is $400 and your AEO cost per citation implies a $180 CAC, that comparison sells itself. Put AEO in the same KPI dashboard as your other channels.
Show the trend. Early quarters will look expensive. That's fine. Show cost per citation declining quarter over quarter. AEO is an investment that compounds, not a campaign that resets each month.
We run this attribution framework for every AEO engagement at Jolly Consulting. If you're investing in AI visibility and can't answer the CFO's question yet, let's fix that.
Related reading: