Citation share is the new market share

TL;DR. Citation share is the percentage of relevant AI-engine prompts in your category that mention your brand. It comes in two forms: competitive (your share of all brand mentions in the same answers) and absolute (the share of all prompts where you appear). In the Zaraftis cohort it leads new-logo pipeline by roughly six to ten weeks at a 0.6 to 0.7 correlation, which makes it a leading brand indicator your funnel software cannot see. Build a 100-to-300 prompt set, run it weekly across the major engines, and track it like market share.

For: Marketing and growth leaders who want a measurable brand KPI for the era when buyers ask AI before they ask Google.

Not for: Readers who want the technical fix list first. Start with The AI-readability checklist, then come back for the metric that tells you whether the fixes worked.

Market share has been the most useful number in marketing for sixty years for one reason: it correlates with everything that matters and is hard to fake. If your category has ten dollars of revenue and you have three of them, you have 30%, and that 30% will keep showing up downstream in your pricing power, your hiring, your moat, and your acquisition multiples. The number is a summary statistic for “how much of the market thinks of you as the answer.”

The problem with market share in 2026 is that the market has stopped thinking. The market has started asking. And the entity it asks is an AI engine that gives a single confident answer to a single confident question. The market share of that one answer is the only market share that matters in the moments where buyers are forming their shortlist. There is a name for it. Call it citation share.

What is citation share?

Citation share is the percentage of relevant AI prompts in your category that mention your brand. That is the whole thing. There is a competitive version (your share among the brands an AI mentions when answering category prompts) and an absolute version (the percentage of all relevant prompts where you appear at all). Both are useful. The competitive version is more honest about your relative standing. The absolute version is more honest about your raw visibility.

Written as a formula:

citation_share_competitive(brand) =
   sum(brand_mentions in answers) /
   sum(all_brand_mentions in same answers)

citation_share_absolute(brand, prompt_set) =
   count(prompts where brand was mentioned) /
   count(prompt_set)

The thing that makes citation share different from a “share of voice” you might already track is the prompt set. The prompt set is the unit of analysis, and getting it right is most of the work. A bad prompt set produces a happy meaningless number. A good prompt set produces a number that actually predicts what your sales reps will hear next quarter. (For why the prompt, not the keyword, is now the unit of measurement, see The end of keyword rankings.)

How do you build a prompt set that means something?

The prompt set should look like the questions your buyers actually type into AI engines, weighted roughly by frequency and roughly by intent. There are three sources worth leaning on, in order of usefulness.

One: ask your sales team. They will tell you, in five minutes, the questions prospects open with. “We’re comparing you against [X] and [Y], can you tell us how you differ on [Z]?” That kind of question, generalized, is exactly the kind of prompt that buyers are now asking AI engines too, often before they get on a call.

Add the comparison and shortlist prompts that lead to demos. Add the “best [X] for [scenario]” prompts that show up in your category. Add the “alternative to [competitor]” prompts. You should be able to write thirty to sixty prompts in an afternoon for any reasonably defined B2B category.

Two: borrow from your search analytics. The questions that have always shown up in long-tail organic search are still the questions buyers ask. The trick is to translate them out of keyword form (“notion alternative product roadmap”) and into prompt form (“What are the best alternatives to Notion for managing a product roadmap?”). Phrasing matters more than people think. The same intent, asked two ways, can produce different answers.

Three: log the prompts you and your team actually use. Marketers are buyers too. Pay attention to how you research a B2B tool when you are about to evaluate one. The phrasing pattern you use is the phrasing pattern your buyers use, and it is freer of “what we wish people asked” bias than any survey would be.

Aim for a prompt set in the 100 to 300 range to start. That is enough to be statistically meaningful, small enough to run daily without burning credits, and varied enough to catch the gaps where you are quietly invisible.

What does good citation share look like?

Numbers without context are noise, so here are some. These are illustrative, drawn from patterns Zaraftis has seen across roughly 200 brands over the last six months.

For a category leader, a healthy competitive citation share is in the 35 to 50% range. You are not going to hit 100, because AI engines deliberately produce diverse answers, and there are always two or three competitors that will be co-mentioned. But you should be the first brand named on most prompts, and you should appear in nearly all of them.

For a strong challenger, a healthy citation share is 15 to 30%. You are clearly in the consideration set, you are mentioned alongside the leader more often than not, and your absolute coverage of buyer prompts is at least 60%.

For a brand that is “in the market but not in the conversation,” citation share is typically 2 to 8%. You are technically there, occasionally mentioned, but you are not part of the standard recommendation set. Most B2B brands fall into this bucket and do not realize it.

For a brand that is invisible, citation share is under 1%. Zaraftis sees this all the time, especially for brands that have decent SEO and a real product but have done none of the GEO work. They show up in a Google search and they do not show up in an AI conversation. The CEO is unaware, because the analytics dashboard does not measure the gap.

Why does citation share predict pipeline?

The reason a marketing leader should care about citation share is that, in the cohort Zaraftis has data on, it is one of the strongest leading indicators of new logo pipeline anyone has measured. Across the customers who have shared their CRM data, lifts in citation share lead lifts in new pipeline by roughly six to ten weeks, with a correlation coefficient that hovers in the 0.6 to 0.7 range. That is in roughly the same neighborhood as the relationship between brand search volume and pipeline, with the difference that citation share is much earlier in the funnel and much earlier to move.

What seems to be happening, mechanically, is this. A buyer asks an AI engine a category question. The engine names two or three brands. The buyer notes those names, looks them up directly a few days or weeks later, and the visit lands as direct or branded organic in your analytics. By the time the lead enters your CRM, the AI conversation that influenced it is invisible. The pipeline shows up. The cause does not.

Citation share, then, is the metric that puts the cause back into the picture. It is the leading indicator of the branded interest that becomes pipeline. If you are a CMO trying to explain why your branded search is up 22% even though your paid spend is flat, you are likely looking at the downstream effect of a citation share movement that started two months ago.

The objection: “Mentions are not conversions”

Here is the pushback that comes up almost every time Zaraftis presents this metric. “Sure, AI mentions us. So what? They didn’t click. They didn’t fill out a form. We can’t attribute pipeline to a citation.”

This is the same argument people made against measuring “share of voice” in offline advertising in 1985, and it was wrong then for the same reason it is wrong now. The fact that an effect is not directly attributable in your funnel software does not mean the effect is not happening. It means your funnel software is not built for it. Citation share is, structurally, a brand metric. You measure it because the brands that win on it win downstream, even when you cannot draw a clean line between the mention and the demo request.

The teams that internalize this are going to look like geniuses in 2027. The teams that wait for a clean attribution model before they will spend on it are going to spend the year explaining why their pipeline got worse and they cannot quite say why.

How do you operationalize citation share without overcomplicating it?

You do not need a fifty-slide framework to get started. The minimum viable version is:

1. Define your prompt set, 100 to 300 prompts.
2. Run it across the major AI engines weekly. Daily is better, but weekly is enough to start.
3. Track three numbers: absolute citation share, competitive citation share, and sentiment of mentions.
4. Put the trend lines in your monthly marketing review.
5. When the numbers move, dig into which prompts moved and why.

Almost any tool that monitors AI surfaces can give you the raw data, Zaraftis included. The discipline is the part that is non-trivial. You need a prompt set you actually trust, you need to run it on a real schedule, and you need to make the resulting number a thing your team is judged on. Until citation share is on someone’s quarterly goals, it will not get the attention it needs.

The strategic implication

If citation share behaves like market share, then the strategic question for any brand becomes: what would it take to grow our citation share by 10 percentage points over the next two quarters? That is a question with operational answers. Better third-party reference content. Cleaner structured data. A canonical comparison page for every meaningful competitor. A stronger entity description across the open web. The work is real, the levers are concrete, and the result is measurable in a way that, frankly, market share itself never was. (The concrete levers are the AI-readability checklist; the why-they-work is How AI engines decide which brands to cite.)

This is the part that catches strategy teams off guard the most. Market share, in most categories, is a number you measure once a year, in a private report from a research firm whose methodology you cannot quite verify. Citation share is a number you can measure every day, with a methodology you control, on a dashboard your team owns. It is the more measurable version of the same idea. That is not a downgrade. That is an upgrade with a different name.

If your CEO has not asked you about your citation share yet, your CEO is about to. The question is whether you have an answer.

Frequently asked questions about citation share

Q: What is the difference between citation share and share of voice? A: Share of voice usually counts how often your brand appears across a channel. Citation share is specific to AI-engine answers and is anchored to a defined prompt set: the percentage of relevant category prompts where your brand is named. The prompt set is what makes citation share predictive rather than vanity. Citation share also splits into competitive (your share of all brand mentions in the same answers) and absolute (your share of all prompts) forms.

Q: How do I calculate citation share? A: Competitive citation share is your brand’s mentions divided by all brand mentions across the same set of answers. Absolute citation share is the count of prompts where your brand appeared divided by the total prompt set. Run a fixed prompt set of 100 to 300 buyer-intent questions across the major AI engines on a weekly cadence and compute both.

Q: What is a good citation share? A: It depends on your position. Illustrative bands from the Zaraftis cohort: a category leader sits at 35 to 50% competitive citation share; a strong challenger at 15 to 30%; a brand “in the market but not in the conversation” at 2 to 8%; an invisible brand under 1%. Most B2B brands are in the 2-to-8% bucket and do not know it.

Q: Does citation share actually drive pipeline? A: In the Zaraftis cohort that shared CRM data, lifts in citation share led lifts in new-logo pipeline by roughly six to ten weeks at a 0.6 to 0.7 correlation. The mechanism is indirect: a buyer asks an AI engine, notes the brands named, and looks one up directly later, so the visit lands as direct or branded organic and the AI conversation is invisible to attribution. Citation share is a leading brand indicator, not a last-click conversion metric.

Q: How many prompts do I need to track citation share? A: Start with 100 to 300 buyer-intent prompts. That is enough to be statistically meaningful, small enough to run daily without burning credits, and varied enough to surface the prompts where you are quietly invisible. Source them from your sales team, your long-tail search analytics (translated into prompt form), and the prompts your own team uses when evaluating tools.

Q: How often should I measure it? A: Weekly at minimum, daily if you can. AI answers shift as engines re-ground and re-rank, and a regression can show up overnight. Put the trend lines in your monthly marketing review and tie them to someone’s quarterly goals, or the number will not get the attention it needs.

How we know

The figures in this article come from the Zaraftis platform: a prompt-tracking system that runs buyer-intent prompts against ChatGPT, Gemini, Perplexity, Claude, Google AI Overviews, AI Mode, Copilot, and Grok on a weekly cadence. The citation-share bands (35 to 50% leader, 15 to 30% challenger, 2 to 8% in-market, under 1% invisible) are illustrative patterns drawn from roughly 200 brands tracked over the six months ending May 2026. The pipeline figures (six-to-ten-week lead, 0.6 to 0.7 correlation) come from the narrower subset of those customers who shared CRM data; the correlation is real and directional, not a controlled study.