Show up in AI Answers Before Everyone Else

What each input and output means

Every number in the calculator is explicit—you choose the assumptions. Symbols below match the formulas in How this calculator works and the short hints under each field.

Inputs you enter

Monthly SEO spend (C)

Total cost of the SEO motion for the month: tools, contractors or agency, content production, and any in-house time you allocate. C is the base for ROI: modeled revenue is compared to this spend.

Organic sessions / month (T)

Count of organic sessions in your reporting window, usually from Google Analytics or equivalent. The embed does not read GA4—only the value you type counts. If you want Direct or other channels in the “organic story,” you must merge them yourself before entering T.

AI mention volume / month (M)

Number of mentions of your brand (or tracked entity) in AI-generated outputs you measure—e.g. exports or dashboards from Savannabay. This is not website traffic; it is visibility in answers.

Value per organic session (v_s, $)

Your average economic value for one organic visit: often “organic-attributed revenue (or pipeline) ÷ organic sessions” for a recent month, or “conversion rate × value per conversion” if you lack full attribution. It is a modeling input, not a GA4-built-in metric.

Value per AI mention (v_m, $)

Your average economic value for one monitored mention. There is no universal GA4 line for this—teams calibrate from experiments, branded lift, or conservative rules of thumb. Use a lower v_m when you want the AI line to reflect only the share of mention value you believe is incremental.

Outputs after you click Calculate

Classic (organic-only) monthly attributed revenue

T × v_s. Revenue (or value) implied by organic sessions alone, given your v_s assumption.

AI attributed revenue (monthly)

M × v_m using the full mention count M—all monitored mentions in the period, not a subset labeled “dark” in analytics. Lower M or v_m if you want to approximate only incremental or hard-to-see impact. The separate organic-only story drops this entire line if you ignore AI mentions in the model.

Blended monthly attributed revenue

(T × v_s) + (M × v_m). Organic line plus AI line. Your org can later decide whether to merge, cap, or de-duplicate for official reporting—the calculator keeps them visible side by side.

ROI classic

(organic revenue − C) ÷ C, expressed as a percentage. Return on spend if only the organic attributed revenue counts.

ROI blended

(blended revenue − C) ÷ C, as a percentage. Return on spend when the AI line is included in attributed revenue.

ROI lift from counting AI

Shown as a percentage that equals (ROI blended − ROI classic) in the same units as the two ROI rows. Algebraically, for a fixed C, this equals (M × v_m) ÷ C—the AI line’s contribution to the ROI ratio relative to monthly spend.

What is enterprise SEO ROI?

Return on investment (ROI) for enterprise SEO answers whether the business gets more value back than it spends on content, technical SEO, tools, and agency or in-house time. Teams often approximate “value” using organic traffic and a revenue proxy—for example, attributed revenue per organic session, or conversion rate × customer value. That works well for reporting, but it usually reflects what shows up in analytics—not every place buyers learn about you.

Why organic-only ROI can mislead

Organic sessions remain the backbone of many SEO dashboards. At enterprise scale, the gap is this: buyers increasingly get recommendations from AI surfaces—ChatGPT, Perplexity, Microsoft Copilot, Gemini, and embedded assistants. Some of those journeys produce sessions you can see in GA4; many others influence revenue through direct visits and branded search without a clean “AI” channel label.

A classic ROI model can be internally consistent and still understate total impact if you only count what analytics attributes neatly to organic search—while ignoring AI-mediated discovery and “dark” attribution.

When AI sends clickable traffic (it can show up in Analytics)

If someone uses an AI product and clicks a link to your site, that visit often appears in Google Analytics 4 like any other session. The session source / medium depends on how the product passes the referrer and UTM data—not on whether the user “came from AI” in a marketing sense.

Common channel groupings you may see include:

• Referral — when the browser sends a referrer from the AI or host domain.
• Organic search — when the click is routed through a search engine (e.g. Bing) and is classified as organic.
• Direct — when no referrer is passed or the session is stripped (very common in some flows).

Real-world examples of source / medium pairs teams look for (your property may vary):

• chat.openai.com / referral
• perplexity.ai / referral
• bing.com / organic
• (direct) / (none) — still possible even when the user originated from an AI-assisted journey.

Important: Classifications change as products update how links open (in-app browser, copy-paste, email handoff). Treat GA4 as a partial view of AI-influenced traffic—not the full picture.

When influence does not show up as “AI” in Analytics

Many high-value interactions never produce a trackable click from the AI surface to your site. Typical pattern:

• The user reads your brand in an AI answer (recommendation, comparison, citation).
• Later they open a browser and type your URL, click a bookmark, or search your brand name.

GA4 will often record that follow-up visit as Direct (or branded organic)—even though discovery started in the AI experience. That is a form of dark attribution: the channel that shaped demand is invisible in default reports.

For ROI storytelling, the consequence is blunt: if you attribute value only to sessions with a tidy “organic” or “paid” label, you systematically under-credit anything that lifts direct and branded demand after AI exposure. That is one reason enterprise teams pair analytics with AI mention tracking (e.g. Savannabay) alongside classic SEO metrics.

How to spot AI-related sessions in Google Analytics 4

GA4 paths differ slightly by workspace layout, but a practical starting point is:

• Open Reports → Acquisition → traffic overview or Traffic acquisition (or an exploration built on the same dimensions).
• Set the primary dimension to Session source, Session medium, or the combined Session source / medium (depending on your report).
• Use the search or table filter and try tokens such as: chat, openai, perplexity, copilot, bing, gemini, ai.

For a broader catch-all in explorations or segments, teams sometimes use a regular expression on session source (adjust to your data and false-positive tolerance), for example:

(chat|openai|perplexity|copilot|gemini)

Caution: bing.com traffic is not exclusively “AI”—it is mixed with classic Bing organic and other Microsoft surfaces—so interpret those rows in context. Build segments or Looker Studio filters you can defend in a monthly review, and revisit them quarterly as referral patterns shift.

What we are seeing in 2025–2026

Across many categories, AI assistants are becoming a new discovery layer—sometimes described as a complement to traditional SEO rather than a replacement. Practical patterns we hear from practitioners:

• More “dark” attribution — influence without a clear multi-touch path in GA4.
• Rising Direct — especially when answers build brand recall but users navigate manually.
• Branded search growth without a proportional lift in generic organic—suggesting demand is forming off-SERP before it shows up as classic SEO wins.

None of this replaces disciplined technical SEO or content. It means enterprise ROI models should ask: are we measuring only what GA4 labels cleanly, or also the demand AI visibility creates? This calculator is a simple sandbox for the second question—using explicit assumptions finance can debate.

What are AI mentions?

Here, AI mentions are instances where your brand (or tracked entity) appears in AI-generated answers or outputs you monitor—for example with Savannabay. Volume can move with prompts, models, citations, and competitive narratives.

This calculator does not prove causality from a mention to revenue. It gives a transparent way to stress-test how sensitive your ROI narrative is when you add a separate AI-attributed revenue line, using assumptions you can explain in a deck.

How this calculator works

Definitions match the numbers we show above:

• Monthly SEO spend (C) — total program cost for the month.
• Organic sessions (T) × value per organic session (v_s) → organic attributed revenue (≈ T × v_s). Here v_s is the average dollar value you assign to one organic visit for modeling—it is not a built-in GA4 metric. It compresses attribution into one number so you can stress-test ROI; your finance team may use a different definition.
• AI mention volume (M) × value per mention (v_m) → AI attributed revenue (≈ M × v_m). Use a conservative v_m if you want to bake in incrementality and overlap with organic in one number.
• Blended revenue — sum of organic and AI lines (two explicit lines so you can discuss overlap or double-counting in your own methodology).
• ROI — (revenue − cost) ÷ cost for the classic view (organic only) and for the blended view.

AI attributed revenue is M × v_m for all mentions in M. Discussion of dark attribution elsewhere explains why teams add this line next to organic—it does not mean this row isolates only unseen or Direct traffic. ROI lift from counting AI is the difference between blended ROI and classic ROI.

Dark traffic: The tool does not read your GA4 “Direct” bucket or auto-detect dark sessions. Whatever you type as organic sessions stays on that line only. The AI mention line is M × v_m for all mentions in M—not a dark-only subset—while the copy on dark attribution explains why that full mention line still matters beside organic reporting.

Why this matters for enterprise teams

Benchmarking — Compare “ROI story A” (organic only) vs “ROI story B” (organic + AI visibility) on the same spend.
Planning — If AI mention volume is a strategic metric, set targets and show how ROI moves under a conservative value per mention.
Alignment — Shift the conversation from whether AI visibility “counts” to which assumptions the org accepts.

FAQ

Is this financial advice?

No. Illustrative model for discussion and planning only.

Will this match my GA4 numbers?

Organic revenue here is a model from your inputs, not a GA4 export. GA4 will capture some AI-referred clicks, but not the full influence when users arrive as Direct or branded search after seeing your brand in an answer.

What is “dark attribution”?

Marketing impact you believe happened (e.g. discovery via an AI answer) that does not appear as a labeled AI session in analytics—often because the next visit is recorded as Direct or organic brand.

Does the calculator count “dark” visits from GA4?

No. It never pulls GA4 automatically. Organic sessions are only the number you enter—usually organic from GA4—so Direct (including many AI-influenced paths) is not included in that field unless you deliberately merge numbers yourself. The AI mention line is a separate modeled estimate meant to stand in for visibility and downstream demand that analytics under-attribute.

Does blended revenue double-count?

Organic and AI are separate lines so your team can decide what to merge for reporting.

How should I pick “value per organic session”?

It is your modeling assumption for “how much one organic session is worth on average.” Strongest approach: revenue (or pipeline) attributed to organic for a recent month ÷ organic sessions in GA4 for the same period—e.g. $175k ÷ 500k sessions ≈ $0.35 per session.

If you do not have revenue attribution yet, approximate with organic conversion rate × value per conversion (order value, lead value, or discounted LTV)—e.g. 2% × $80 ≈ $1.60 per session. If you are early, use a deliberately low default and treat the output as directional only.

How should I pick “value per AI mention”?

It is a single modeling input: the dollar value you assign to one monitored mention on average. If you would have used a separate “attribution %” before, fold that judgment into a lower v_m instead. Many teams start conservative and scenario-test upward.