Calculating ROI of GEO

Calculate GEO ROI as (attributed return − cost) ÷ cost, using two versions: a conservative ROI that counts only pipeline from tracked AI referrals and self-reported AI attribution, and an influenced ROI that adds a stated share of pipeline from segments where AI visibility rose. Our view: present both, with assumptions written next to the numbers, because AI influence is real but often invisible to click tracking.

Points marked "Our view" are interpretation; all figures below are illustrative placeholders.

On this page

• Why is GEO ROI hard?
• What costs should you include?
• What returns count?
• The template
• Worked example (fictional)
• How do you choose the influence share?
• Should you use revenue or pipeline?
• What about lifetime value?
• How should you present it?
• What leading indicators should sit beside ROI?
• How do you compare GEO with other channels?
• Who should review the calculation?
• Which pitfalls should you watch for?
• How do you improve the estimate over time?
• When should you not calculate ROI yet?
• Exercise
• Checklist
• From MentionX data: why visibility is a leading indicator
• Worked example: Ledgerline estimates GEO ROI for a quarter
• Where does the cost data come from?
• How do you count measured returns?
• How do you handle the invisible effect?
• What time frame should you use?
• How do you present uncertainty?
• What does a simple ROI table look like?
• How do you avoid double counting?
• Common questions
• Related
• Sources

Checked on 31 October 2026. "Our view" marks MentionX opinion. Ledgerline is a fictional company used for illustration.

Why is GEO ROI hard?

Buyers often read an AI answer, remember a name and arrive later via search or direct. Click-based attribution undercounts this. See attribution: connecting AI visibility to pipeline. Our view: that's a reason to show a range, not to avoid the calculation.

What costs should you include?

Our view: include only the share of time actually spent on GEO.

What returns count?

Our view: deduplicate: an opportunity counted in AI referral shouldn't also count in self-reported.

The template

Period: ____ (e.g. last 2 quarters)

COSTS
People time (€)            ____
Tools (€)                  ____
Content (€)                ____
Earned media (€)           ____
Agency/freelance (€)       ____
TOTAL COST (C)             ____

CONSERVATIVE RETURN
AI-referral pipeline (€)   ____
Self-reported AI pipeline (€, deduplicated) ____
Win rate (%)               ____
Gross margin (%)           ____
Conservative return (R1) = pipeline × win rate × margin = ____

INFLUENCED RETURN
Pipeline in segments with rising AI visibility (€) ____
Share attributed to GEO (assumption, %) ____
Influenced return (R2) = R1 + (segment pipeline × share × win rate × margin) = ____

ROI
Conservative ROI = (R1 − C) ÷ C = ____
Influenced ROI   = (R2 − C) ÷ C = ____

Assumptions and caveats:
- ____________________

Worked example (fictional)

Costs over two quarters: €60,000. AI-referral plus self-reported pipeline (deduplicated): €400,000. Win rate 20%, margin 80%. R1 = €400,000 × 0.2 × 0.8 = €64,000. Conservative ROI = (64,000 − 60,000) ÷ 60,000 ≈ 7%.

Segment pipeline with rising visibility: €1,000,000; assumed GEO share 10%. Extra = €1,000,000 × 0.1 × 0.2 × 0.8 = €16,000. R2 = €80,000. Influenced ROI ≈ 33%.

How do you choose the influence share?

Our view: start low and justify it. Evidence might include: self-reported AI mentions rising in that segment, visibility gains on that segment's prompts, and branded search growth without other campaigns. Show what ROI looks like at 5%, 10% and 20% so readers see the sensitivity.

Should you use revenue or pipeline?

Our view: use closed revenue when your sales cycle is shorter than the period; otherwise use pipeline adjusted by win rate, as above. State which you used.

What about lifetime value?

Our view: for subscription businesses, using first-year revenue is conservative. You can show a lifetime value version as a third line, clearly labelled.

How should you present it?

• Both ROIs side by side.
• Costs itemised.
• Assumptions listed, with the influence share sensitivity.
• What the calculation can't capture.

Our view: finance teams trust a careful conservative number more than an impressive one.

What leading indicators should sit beside ROI?

Our view: ROI lags. Visibility usually changes before pipeline does, and pipeline changes before revenue. Show leading indicators next to the ROI so leaders can judge direction: mention and recommendation rate on high-intent prompts, citation share, the number of incorrect facts in answers, AI referral sessions and self-reported AI leads. If indicators rise while ROI is still low, that's an early sign the investment is working. If they're flat after two quarters, it's a sign to change approach before spending more.

How do you compare GEO with other channels?

Our view: carefully. Paid channels have near-complete click attribution; GEO doesn't. Comparing a conservative GEO ROI with a paid search ROI will understate GEO, while comparing an influenced GEO ROI with last-click paid ROI may overstate it. State the attribution method for each channel in the same table. Where possible, use the same multi-touch or self-reported method across channels so the comparison is fair.

Who should review the calculation?

Our view: someone in finance or revenue operations. Ask them to check costs, win rate, margin and deduplication. Their review makes the result more credible and often improves it, for example by finding costs you missed or a more accurate win rate by segment. Keep their comments with the calculation for the next quarter, and invite them to the quarterly review where the ROI is presented.

Which pitfalls should you watch for?

• Double counting opportunities.
• Using pipeline without win rate.
• Hiding the influence assumption.
• Comparing GEO ROI with channels measured differently without saying so.

How do you improve the estimate over time?

Our view: add the self-reported field to more forms, ask sales to record AI mentions in discovery calls, and keep visibility tracking stable. Each quarter, the conservative number gets more complete and the influence assumption matters less. Record changes to methods so readers can compare quarters fairly.

When should you not calculate ROI yet?

Our view: in the first quarter, when tracking isn't set up and costs are mostly setup. Report progress on measurement and leading indicators instead, and agree in advance when the first ROI calculation will be done.

Exercise

Fill in the template for the last two quarters using your real costs and conservative pipeline. Then calculate influenced ROI at 5%, 10% and 20% influence shares.

Checklist

• All costs itemised
• Conservative return deduplicated
• Win rate and margin applied
• Influence share justified and varied
• Both ROIs presented
• Assumptions written down

From MentionX data: why visibility is a leading indicator

Method. Read-only aggregate queries over MentionX scans from 22 August to 29 October 2026 (UTC): 2,217 engine runs (2,154 completed) across 18 anonymised projects and six engines.

ROI lags visibility. Our sample shows that how often brands are named depends heavily on the engine and the question type, which affects how much value can follow:

Our view: comparison and recommendation questions name brands most often and sit closest to purchase, so visibility gains there are the most likely to show up in pipeline. Weight your ROI story towards those segments rather than overall visibility.

Limits. A convenience sample of mostly B2B software projects, run through engine APIs, which can behave differently from the consumer apps. Perplexity, Copilot and DeepSeek figures come from a single day (26 October 2026), so treat them as a snapshot, not a trend.

Worked example: Ledgerline estimates GEO ROI for a quarter

Ledgerline is a fictional accounts-payable tool. Figures are illustrative.

Ledgerline's GEO lead totals quarterly costs: a part-time specialist, some writer and developer hours, a tracking tool and outreach time. For returns, she takes AI referral conversions from analytics and deals where buyers chose an AI assistant in the "how did you hear about us" field.

She applies Ledgerline's own average deal value and win rate, then shows three scenarios: referral-only (a lower bound), referral plus self-reported, and a higher estimate that applies a modest influence share to deals where AI was one of several sources. She presents the range, the method and the leading indicators, and says which figures are measured and which are estimated.

Where does the cost data come from?

Ask finance for fully loaded hourly costs by role, which include benefits and overheads. Multiply by hours spent on GEO work. Add tool subscriptions at their current price, noting the date. Include agency fees and any freelance content costs. Leave out work that would have happened anyway, such as routine SEO, or include it only with a clear share.

How do you count measured returns?

Measured returns come from two places. AI referral sessions in analytics, followed to conversions and pipeline, give you a direct count, though it misses people who read an answer and then searched for you. Self-reported attribution, from a "how did you hear about us" field, catches some of those. Count each deal once, even if both sources flag it.

How do you handle the invisible effect?

Many buyers see your name in an AI answer, then visit later by typing your URL or searching your brand. Analytics records that as direct or branded search. You can't measure it exactly. You can watch whether branded search and direct traffic rise alongside visibility gains, and you can ask in sales calls. Present this as context, not as counted return.

What time frame should you use?

B2B buying cycles can run for months, so a single quarter often understates returns. Show quarterly costs and returns, plus a rolling twelve-month view once you have the history. Note the typical sales cycle length so readers understand the lag.

How do you present uncertainty?

Show three numbers: a measured lower bound, a central estimate and a higher estimate, each with its assumptions listed. Avoid false precision; round to sensible figures. Our view: finance teams trust a range with clear assumptions far more than a single exact-looking number.

What does a simple ROI table look like?

Use one row per quarter with columns for total cost, measured return, central estimate, higher estimate and the ratio of each to cost. Add a notes column for anything unusual, such as a big launch or a tracking change. Keep the formulas visible so anyone can check them. A plain table that finance can audit is more persuasive than a polished chart.

How do you avoid double counting?

Deals can appear in analytics as an AI referral, in the self-reported field and in another channel's report. Agree with other channel owners how shared deals are credited.

Common questions

Can GEO ROI be measured exactly? No. Some effects are invisible to analytics. Use a measured lower bound and labelled estimates.

Which data source is most reliable? AI referral conversions in analytics, with self-reported attribution as a second source.

How soon should I calculate ROI? After at least a quarter, and longer for long sales cycles.

Related

• Setting GEO goals and OKRs
• Self-reported attribution

Sources

This template describes a calculation method and cites no platform rules. For tracking AI referrals, see Google Analytics Help, Custom channel groups. All accessed 30 Oct 2026.