AI Mode vs Calculator Publishers: 10,000 Traffic-Mix Scenarios

Google's May 2026 AI Mode rollout has finance publishers panicking. The pattern in the industry data is sharper than people think: the protection isn't long-form content quality — it's whether your page does something the AI can't. We modeled 10,000 synthetic finance publishers to map exactly which traffic mixes survive.

Finding 1: The category multiplier dominates the publisher

Per-query-type loss multipliers (AI exposure × CTR drop when shown):

The tool-required category loses 6%. Definition queries lose 53%. A publisher's traffic loss is dominated almost entirely by how much of their traffic sits in each bucket — not by content quality, age, domain authority, or schema.

Finding 2: Publishers cluster bimodally — survivors and casualties

The 10,000-publisher loss distribution doesn't form a smooth bell curve. It bunches around two centers — one near 24.7% (tool-dominant publishers) and one near 39.4% (definition-dominant publishers). Most finance publishers can predict their fate by looking at their top 20 URLs and asking: "does this page require a tool to answer the query, or just text?"

Finding 3: The cheapest defense is restructuring, not new content

A publisher with a 60% definition-content / 40% tool-content mix loses ~34% of traffic. The same publisher with a 40% definition / 60% tool mix loses ~24% — roughly half. That re-mix doesn't require new content; it requires moving the calculator higher on definition pages and tagging them with `SoftwareApplication` schema so they get re-classified as tool-required intent.

Finding 4: AI Overviews still cite — and that citation behavior favors specific patterns

AI Mode doesn't eliminate publisher visibility — it changes the visibility mechanism. The pages that do get cited as sources in AI Overviews share patterns industry studies repeatedly identify:

• Specific numerical claims with primary-source citations — not "compound interest is powerful" but "at 7% real return, $500/mo grows to $1.22M in 40 years."
• Dataset / Original-Research schema — pages that publish their own data and tag it as `Dataset` get pulled in disproportionately.
• Tight, structured answer paragraphs at the top of the page — what we call "speakable summary" — formatted so the AI can lift one block as a citation.
• Author + Organization entity coverage — `sameAs` linking the publisher to external entities (Wikidata, Crunchbase, Twitter/X) raises the citation rate substantially.

The defensive playbook is therefore three moves: (1) add tools to your text-dominant pages, (2) ship original research and tag it as `Dataset`, (3) wire up author + organization entity schema with real external `sameAs` references.

Sample publishers (8 across the loss spectrum)

Methodology

• 10,000 publishers, deterministic via Mulberry32 PRNG seeded 20260524.
• Traffic mix per publisher: Dirichlet-style random share across 5 query types, normalized to sum to 1.
• Baseline traffic: uniform 1,000–1,000,000 monthly visits.
• 5 query types with these modeled parameters (AI exposure × CTR drop when shown):
• Definition / explainer ('what is compound interest'): 85% × 62% = 53% net loss.
• Tutorial / how-to ('how to calculate APR'): 78% × 55% = 43% net loss.
• Comparison ('snowball vs avalanche'): 70% × 48% = 34% net loss.
• Best / review ('best HYSA 2026'): 68% × 45% = 31% net loss.
• Tool-required ('compound interest calculator'): 22% × 25% = 6% net loss.
• Post-AI traffic per publisher = Σᵢ (mix[i] × baseline × (1 − exposureᵢ × dropᵢ)).

Where the parameters come from

AI exposure rates by query type are modeled from publicly reported SISTRIX, Authoritas, and Semrush studies (2024–2026) showing AI Overviews appear most on informational queries and least on transactional/tool queries. CTR-drop-when-shown ranges of 30–70% are from publisher case studies (Search Engine Land, Detailed.com, multiple in-the-wild measurements). We picked mid-range values for each category to avoid cherry-picking either alarmist or optimistic ends.

These are modeled estimates, not measurements of any specific publisher's traffic. The simulation's value is the relative spread between query-type winners and losers — that spread is robust to ±10 percentage points of parameter uncertainty in any single cell.

Limitations

• 5 query types is a simplification. Real publishers have dozens of subcategories; some (commercial-investigation, navigational) we omitted because their AI Overview behavior is still settling.
• Revenue ≠ traffic. CPM differs sharply by query type — high-intent tool queries monetize 3–5× better than definition queries. Tool publishers' revenue resilience is therefore stronger than the traffic numbers alone suggest.
• The model is a one-step projection. Real CTR will continue to drift as users habituate to AI Overviews. Year-2 numbers will likely be worse than year-1.
• Citation-rate-back-to-publisher is not modeled. A page cited as a source in an AI Overview gets some traffic back; our simulation treats CTR drop as net.
• Direct + referral + email traffic are unaffected by AI Overviews and not modeled in the simulation.