A fall in search clicks does not tell you what broke. A page may have disappeared after a migration. Buyers may be asking a question the site never answered. A search result may still show the business but send fewer visits. Each problem calls for different work, yet all three can produce the same line in a report.
Search Equity is Caldrin's framework for the accumulated demand-capture value of a business's commercial pages, its coverage of buyer questions, its authority and independent sources, and its brand and entity clarity. Technical access is checked across all four. AI visibility is read separately, as a measurement of that work, not as a fifth asset. Where AI discovery matters, we count appearances in defined answers. The framework helps decide what to inspect and build. It is not a balance-sheet asset, a universal SEO score, or a promise that returns will compound.
Caldrin did not coin the phrase. Bill Hunt's Search Equity Gap compares organic visibility a business once held with what it holds now and diagnoses the loss. That is especially useful after a migration or a known decline. Our framework starts with a different decision: what can a buyer find and verify today, and which missing or damaged asset deserves investment next? The two approaches can be used together.
Start with the buyer's question
Imagine a software vendor that ranks first for its own name. Its site has no page explaining the workflow buyers research before they know the vendor exists. The brand ranking is accurate. It does not answer whether the company can enter an unbranded shortlist.
Now imagine a service company with a useful page for that workflow, but the page became unindexable after a template change. More copy would not repair the immediate fault. First find out whether the page resolves, can be crawled and can appear in the relevant search surface.
These are illustrations, not client outcomes. They force a useful distinction: the assets and conditions that make discovery possible are different from the readings used to observe it.
| Inspect the asset or condition | Ask a concrete question | Check a separate reading |
|---|---|---|
| Commercial pages and useful assets | Does a real page answer the buying question and give the visitor a next step? | Relevant landing-page visits and qualified enquiries |
| Query and intent coverage | Which category, problem, comparison and service questions have no useful answer? | Query-level visibility within a stated market and period |
| Authority and independent sources | Can buyers find credible outside material that accurately supports important claims? | Relevant links, mentions and cited sources, assessed for quality |
| Brand recognition and entity clarity | Are the company, people, products and relationships represented consistently? | Branded discovery and observed identification in defined answers |
| Technical access, checked across the above | Can the relevant systems fetch and use the material? | Response, crawl, index and eligibility checks |
Technical access is a prerequisite, not a substitute for useful material. Google says a page must be indexed and eligible for a snippet to qualify as a supporting link in AI Overviews or AI Mode; eligibility does not guarantee that it will appear. OpenAI describes OAI-SearchBot as the crawler used for websites in ChatGPT search features. A Google indexing check does not answer whether that separate crawler is allowed. Neither check proves that an AI system will recommend the business.
Inspect the gap before commissioning work
Take one buyer problem and write down the phrases a buyer might use for it. Separate category searches from comparisons, urgent problems and named-brand searches. For each group, inspect the current result pages and the sources that appear in relevant AI answers. Record dates, locations and instruments. Then ask:
- Is there a genuinely useful page or asset for the question, and can it be found?
- Does the page explain the product or service at the level a buyer needs to evaluate it?
- Where important outside sources appear, do they describe the business accurately? Is the source independent, and is the evidence credible?
- Does the visitor have a clear route to a relevant enquiry, signup or sales conversation?
- Which of those observations would change the next investment decision?
The order matters only when evidence makes it matter. A blocked commercial page should be fixed before writing another version of it. If access is sound but an important comparison has no page, the missing asset deserves attention. If the right buyers arrive but few make a qualified enquiry, inspect the promise, proof and contact path before buying more traffic.
This is an inspection method, not a claim that every business needs the same pages or that independent mentions can be purchased on demand. A useful source may be an industry publication, a real review, a partner's technical documentation or a research paper. The question is whether it helps a buyer verify a claim, not whether it adds another item to a link count.
Google and AI answers require separate readings
Search Console's performance reports show how often links to a property appeared and were clicked in Google Search under the report's scope. They can be broken down by query and page, subject to the product's reporting rules. They do not tell you on their own why a buyer chose or rejected the business.
An AI visibility test counts a different thing: what a named system returned for a defined prompt set, on particular dates and conditions. A mention, a citation and a recommendation are not interchangeable. If the prompt list, model label, collection method or definition changes, the series may move without a comparable change in the market. Record the count and denominator for each outcome; do not convert it into an all-purpose Search Equity number.
The distinction also applies to source selection. In Ahrefs' study of 15,000 long-tail prompts, the researchers compared URLs cited by four AI assistants with Google and Bing results for the original queries. The overlap was limited in that sample. The data came from Ahrefs' own Brand Radar tool and was collected at one point in time; the authors note that results may have changed. It does not show what any particular brand will receive, and an original query is not necessarily the only query an assistant may use. It does give a practical reason to inspect actual cited URLs rather than assume a Google ranking explains an AI citation.
Qualified enquiries have a third denominator. A sales team needs a written definition of qualification and a record of where conversations began before it can connect discovery to commercial outcomes. Even then, assisted journeys and missing attribution make causation hard to assign. Impressions, scraper runs and qualified enquiries should sit next to one another in a report, not be blended into a single score.
Make the next decision explicit
If clicks fell after a migration, compare the old and new page inventory, redirects, index eligibility and query-to-page coverage. Hunt's loss comparison is useful here because there is a before period and an event to investigate. If access and coverage hold, then look at changes in demand, competitors and result formats. Do not label the whole drop a content problem before those checks.
If product pages still rank but sampled AI answers cite comparison sites and trade publications, freeze the prompt set and inspect those sources. Are they current? Do they name the right product and use case? Would a credible contribution or a clearer commercial page help a buyer verify the company? Treat each proposed change as a hypothesis until repeated measurements and commercial evidence support it.
If visibility grows while qualified enquiries do not, inspect the query mix and conversion path. A business can earn more impressions for questions that do not lead to a purchase. The correct response may be better pages, clearer proof or a different target question. The report should say which, and why.
Search Equity is useful when it changes the work you would commission. Its parts have not been validated as a predictive scale, and no figure in this article measures a company's total Search Equity. Some assets may retain value for years; others can lose visibility after a technical change or a shift in buyer behaviour. Measure the relevant readings, state their limits, and invest in the asset or repair the evidence actually points to.
Caldrin publishes methods and limitations in Research. Our Search + AI work applies this inspection to commercial discovery.