Branded vs Non-Branded Questions and Discovery Visibility

In this lesson

Lesson 1 said SOV is measured only on non-branded questions. Here's why, plus the second key number that comes from those questions (Discovery Visibility) and what the free scorecard actually shows.

Branded and non-branded questions

A branded question contains your brand name. For example:

"Is Acme Helpdesk good for a five-person support team?"

A non-branded question describes a need without naming any brand. For example:

"What's a simple help-desk tool for a small e-commerce store?"

These questions were written for this lesson only.

Buyers often ask before they know you exist. That's the moment that matters: the AI either brings you up on its own, or it doesn't.

Why branded questions inflate visibility

In a branded question the AI will almost always mention you, because your name is already in the question. The answer containing your brand says nothing about whether the AI would find you by itself.

A fictional example: Acme Helpdesk tracks 8 questions, 3 of them branded. Acme Helpdesk appears in all 3 branded answers, but in only 1 of the 5 non-branded answers.

• Counting branded questions too: 4 ÷ 8 = 50%.
• Counting non-branded questions only: 1 ÷ 5 = 20%.

The second number is closer to reality. The first one offers false comfort about whether buyers can discover you.

That's why the SOV Score and Discovery Visibility are calculated on non-branded questions only. Questions whose branded status is unknown are left out too. SOV Tracker's question suggestions also filter out questions that contain the brand name.

Where branded questions belong

Branded questions aren't useless; they answer a different question. "What does the AI say about my brand, and is it right?" is about accuracy, not visibility. Module 5 covers checking whether AI answers get your brand facts right.

Short rule: visibility comes from non-branded questions; accuracy comes from what's said about you.

Discovery Visibility

Discovery Visibility is the share of answered responses to non-branded questions in which the target brand appears on its own.

Formula:

Discovery Visibility = non-branded answers that contain the target brand ÷ countable non-branded answers × 100

Note the word "countable". Only attempts that actually returned an answer go into the denominator. Lesson 3 covers the attempts that don't.

Worked example (representative)

Fictional brand, fictional numbers.

Five non-branded questions about Acme Helpdesk's category are asked on ChatGPT:

• 4 questions get an answer.
• 1 question times out and can't be measured.
• Acme Helpdesk appears in 2 of the 4 answers.

Discovery Visibility = 2 ÷ 4 = 50%

The timed-out question stays out of the denominator. Writing 2 ÷ 5 = 40% would treat a failed attempt as "your brand wasn't there".

SOV vs Discovery Visibility

Both numbers can come from the same answers, but they answer different questions:

• Discovery Visibility: "In how many answers do I show up?"
• SOV Score: "Of all mentions of me and my competitors, what share is mine?"

A fictional case: Acme Helpdesk appears in 3 of 6 answers, so Discovery Visibility is 50%. But competitors are named far more often in those answers, so SOV comes out at 20%. Both are correct. One measures presence; the other measures share against competitors.

What the free scorecard shows

The free scorecard is a smaller sample:

• It asks ChatGPT only.
• It asks 5 automatically generated questions that don't contain the brand name.
• It can run once every 30 days for the same business.

Its headline number is Discovery Visibility, not the SOV Score. The SOV Score needs a confirmed competitor set and the weekly five-engine measurement; the free scorecard has neither.

If the scorecard shows something like "2/4", read it as: your brand appeared in 2 of the 4 questions that were answered. The 'Answer coverage · 4/5' next to it shows how many of the 5 questions got an answer; it isn't visibility.

Details: free scan rules.

Site Readiness isn't SOV either

Site Readiness is a separate diagnostic that checks how open and readable your site is to AI crawlers. It's useful, but it doesn't measure how visible you are in AI answers. A strong Site Readiness result doesn't mean a strong SOV; the two are kept apart and never blended. Module 4 covers Site Readiness.

Summary

• A branded question names your brand; a non-branded question describes a need without naming one.
• Branded questions inflate visibility, so SOV and Discovery Visibility use non-branded questions only.
• Branded questions help with accuracy (Module 5), not visibility.
• Discovery Visibility = non-branded answers containing the brand ÷ countable non-branded answers. Representative example: 2 ÷ 4 = 50%.
• The free scorecard's headline number is Discovery Visibility, not SOV. Site Readiness is not SOV.