The Search Score is a mechanical measurement of one thing: whether the answer engines cite an operator's brand when its buyers ask commercial-intent questions in its category. That is the metric AI Distribution runs on, and Searchalong AI names it Share of Answer: how much of the answer, across a brand's tracked buyer prompts, names that brand. The Search Score is Share of Answer expressed as a bounded 0-100 number. The math is small enough to defend in a paragraph and the score itself is comparable and derived from the same instrumentation for every brand on the ranking. That is what makes it a useful reference number rather than another dashboard vanity metric.
Search Score = (citations across all tracked prompts and engines) / (total tracked prompt-engine pairs) × 100. Rounded to the nearest integer. Refreshed daily at 06:00 UTC.
The formula, expanded
Every morning the tracker queries each tracked buyer prompt against each of the 5 engines the plan covers. When the engine's answer names the brand (by brand name or by domain), that counts as a citation. Total citations divided by total prompt-engine pairs, times one hundred, is the composite.
A concrete example: three tracked prompts across 5 engines gives fifteen prompt-engine pairs. Cited on nine of them, the score is 60. Cited on three, the score is 20. Cited on all fifteen, the score is 100, which is rare and typically defensible only for a brand that owns the category vocabulary outright.
What each range means
- 85-100 · category leader. The engines default to you when buyers ask. Your only work is to hold the line.
- 60-84 · strong presence. You are cited on the majority of prompts. You are still competing for position #1 vs #3.
- 30-59 · emerging. The engines know you exist but do not always name you. This is where most B2B startups start. Every Citation-Lift ships here.
- 0-29 · invisible. The engines rarely or never cite you. Not because you are bad · because the engines have not seen enough co-mention with your competitors. This is fixable in 60-90 days.
Three inputs that move the score
Every point on the composite traces back to one of three underlying inputs. Working on any of them lifts the score; ignoring them for a quarter or more usually shows up as drift.
- Passage-level extractability on the operator's own pages. A product page with a fifty-word direct-answer block anchored on the buyer prompt gets extracted. A page built around a marketing hero image and no factual density does not.
- Third-party co-mention density in the reranker set. When ten listicles name the brand alongside its top category peers, it enters the candidate set the engine reranks. When zero do, it does not, no matter how well the on-site pages are written.
- Freshness on volatile pages. Pricing, feature comparisons, and integration lists reward recency in the reranker. A page updated within the past 30 days consistently beats a technically stronger page carrying a three-year-old datestamp.
Citation-Lift ships fixes for the first input. backlinks-from-cited-sites backlinks work on the second. The Content Studio republish cadence handles the third.
Reading a score honestly
The number is a snapshot, not a verdict. Two profiles worth understanding when a score comes back:
Ships a Citation-Lift on the biggest weekly gap, keeps JSON-LD FAQPage on every commercial-intent page, holds fresh datestamps on the top 10. The score compounds because the reranker keeps seeing the brand where it looks for authority.
Just added an audit to the workflow, has not yet shipped a fix, or has fixes shipped irregularly. The score often looks low because the engines have not yet had a reason to update their view of the brand.
The distance between those two profiles is not primarily team size or budget. It is discipline and cycle time. A brand that ships one Citation-Lift a week and republishes the top 10 commercial-intent pages on a rolling monthly cycle will out-cite a bigger competitor that runs a quarterly SEO campaign every time. That is the honest read of what the score is telling anyone who runs it.
What the score is NOT
Referral traffic from AI answers, because the engines do not always link.
Perceived thought leadership, which lives on LinkedIn analytics rather than in citation data.
Share of voice on non-commercial prompts, which inflates the score without lifting pipeline.
Attention on X or Reddit, which is adjacent to answer engine citation but not the same measurement.
A useful metric answers one clear business question. The Search Score answers exactly one: when the operator's buyers ask an answer engine a commercial-intent question in the operator's category, is the operator named. Everything else is a different measurement, useful for different decisions.
Run a live score at searchalong.xyz. 45 seconds, no signup, and the report will name the specific pages the engines picked over the operator's own on every tracked prompt.