Share of Model: The New Metric That Replaces Page Rank

For two decades, we measured digital authority with Page Rank. In the age of AI search, the metric that matters is how often you're recommended — not how high you rank.

---

In 1998, Google introduced Page Rank — a way to score how much the internet trusted your website by counting and weighting the links pointing to it. Page Rank reshaped an entire industry. SEO agencies built their businesses around it. Real estate agents who understood it early gained a lasting edge.

Page Rank still exists in some form inside Google's algorithm, but its relevance is fading fast. When a buyer asks ChatGPT "Who are the best real estate agents in Austin?" they don't see ranked links. They see names. Two or three of them, presented as direct recommendations.

The agents whose names appear in those answers aren't necessarily the ones with the highest Page Rank. They're the ones with the highest Share of Model.

What Is Share of Model?

Share of Model (SoM) measures how frequently and prominently an AI engine recommends your brand in response to relevant queries.

Think of it as the AI equivalent of "share of voice" in advertising, but applied to generative search. Instead of measuring how much shelf space your ads occupy, SoM measures how much recommendation space your brand occupies in AI-generated answers.

The formal definition:

Share of Model = (Your weighted mentions / Total relevant queries) x 100

Where "weighted mentions" accounts for citation position. Being the first agent named in a recommendation carries more weight than being listed fourth. And being recommended without caveats ("I highly recommend...") outweighs a qualified mention ("You might also consider...").

An illustrative example:

Say there are 12 standardized buyer-intent queries for your market. You run them across four AI engines (48 total queries). Your name appears in 8 responses, and you're listed first in 3 of them.

Using a simple weighting system — first position = 1.5x, second = 1.0x, third or later = 0.5x — your weighted mention count might be:

• 3 first-position mentions x 1.5 = 4.5
• 3 second-position mentions x 1.0 = 3.0
• 2 third-position mentions x 0.5 = 1.0
• Total weighted mentions: 8.5

SoM = (8.5 / 48) x 100 = 17.7%

That number becomes your baseline. Every month, you run the same queries and track whether it's growing, shrinking, or holding steady. It replaces the old SEO question — "What page am I on?" — with a question that actually predicts leads: "How often am I recommended?"

Why Share of Model Matters More Than Rankings

Rankings are binary. Share of Model is proportional.

In traditional search, you're either on page one or you're not. Position 11 is functionally invisible. This binary structure means small ranking changes can have dramatic, unpredictable effects on traffic.

Share of Model is a continuous metric. You can be mentioned in 2% of queries or 20%. Every improvement is measurable and every gain is incremental. There's no cliff between page one and page two — there's a gradient of visibility that you can steadily climb.

Rankings compete with 10. Recommendations compete with 3.

A Google search results page shows ten organic links. Even position one shares attention with nine competitors, plus ads, maps, and featured snippets.

An AI recommendation typically names two to four options. Being in that group means you're sharing the stage with two or three competitors, not nine. The attention density per mention is dramatically higher. And because AI presents its recommendations with explanatory context — "I recommend Agent X because they specialize in..." — each mention carries an implicit endorsement that a search ranking never does.

Rankings decay gradually. Share of Model can shift quickly.

SEO rankings are famously slow to build and slow to lose. Moving from position 15 to position 5 takes months of link building and content. Dropping from 5 to 15 usually requires significant neglect.

Share of Model is more responsive. Engines that search the live web, like Perplexity, can pick up a new review, a well-structured neighborhood guide, or new schema markup as soon as those pages are crawled — there's no months-long link-building cycle standing between the change and the answer. This means agents who act now can see movement sooner than traditional SEO usually allows — but it also means that complacency costs you visibility faster.

How to Measure Your Share of Model

You don't need special tools to start measuring SoM. Here's the manual method:

Step 1: Build Your Query Set

Create 12 standardized queries that represent how a buyer or seller in your market would ask AI for recommendations. Mix broad and specific:

• "Who are the best real estate agents in [city]?"
• "I'm relocating to [city]. Which agents should I talk to?"
• "Best luxury real estate agents in [city]?"
• "Which [city] agents specialize in investment properties?"

Keep the queries identical month to month so your data is comparable.

Step 2: Run Across All 4 Engines

Test each query on ChatGPT, Perplexity, Gemini, and Claude. Record:

• Whether your name appeared (yes/no)
• Your position in the recommendation list (1st, 2nd, 3rd, or later)
• The tone of the recommendation (strong endorsement, neutral mention, qualified/caveated)
• Which competitors appeared alongside you

Step 3: Calculate and Track

Apply the weighted formula above. Record the number monthly. A spreadsheet works fine at this stage.

What Good Looks Like

There's no published industry standard for SoM yet, so treat these as working targets for an individual agent rather than survey data:

SoM Score	Rating	What It Means
20%+	Excellent	You're a dominant presence in AI recommendations for your market
10–19%	Good	You're showing up regularly but have room to grow
5–9%	Emerging	You're on the radar but inconsistently recommended
1–4%	Marginal	Rare mentions, usually on only one engine
0%	Invisible	AI engines don't know you exist

Starting at 0% is normal — AI answers name only a few agents per question, and every market has far more agents than slots. The agents who climb tend to have the same fundamentals in place: schema markup on their websites, a steady flow of recent Google reviews, and substantive local content.

Growing Your Share of Model: The Flywheel

SoM has a compounding property that makes early investment disproportionately valuable.

Here's how the flywheel works:

Higher AI visibility leads to more people searching your name (because AI recommended you), which leads to more content mentioning you (search activity signals relevance to content creators and aggregators), which leads to more AI training data about you, which leads to even higher AI visibility.

This flywheel is the core reason early movers in AEO have a structural advantage. Each cycle reinforces the next. The agent who starts building SoM today isn't just gaining today's mentions — they're training tomorrow's AI models to recommend them by default.

The Compounding Risk of Waiting

The flip side of the flywheel is equally important: while your SoM stays at zero, your competitors' SoM is growing. Share of Model is inherently relative. In a market where AI typically recommends three agents, those three slots aren't expanding. Every month that a competitor builds SoM, your path to breaking into the recommendation set gets harder.

The logic is straightforward: the agents already being recommended keep accumulating reviews, mentions, and citations, which gives AI engines more data about them and more confidence in recommending them. Displacing an agent with that kind of head start takes more effort than claiming a slot no one has locked in yet.

The best time to start building SoM was a year ago. The second best time is right now.

Your SoM Starts With a Baseline

You can't grow what you can't measure.

REMG's free AI visibility audit runs 12 queries across all four AI engines — the same methodology behind Share of Model — and gives you your starting numbers. You'll see your mention rate, your grade, which engines recommend you, and which recommend your competitors instead.

That baseline is your SoM starting point. From there, every action you take — adding schema markup, earning reviews, publishing local content — moves the number. And you can measure the movement.

Get Your Free AI Visibility Audit →