Research
We measure how AI assistants, search engines, and buyers actually evaluate brands — then publish the datasets, methodology, and rankings so journalists and analysts can cite them directly.
Which research finding is strong enough to inform this decision?
Use the research library to decide whether a study can meaningfully inform a specific SEO question, not to collect conclusions in isolation. Start with the decision in front of you, then inspect the study question and the evidence boundary around the answer. Check whether the observed population, collection period, search surface, model or platform version, exclusions, classification rules, and measured outcome resemble the situation you need to evaluate. A finding may be well supported inside its own scope and still be weak evidence for a different query set, audience, market, site condition, or period.
Before relying on a headline, read the method and identify exactly what the study observed. Confirm what entered the sample, what did not, how responses or search results were collected, how observations were coded, and what integrity notes or limitations accompany the analysis. Keep direct observations separate from interpretation. If an edition, dataset, explorer, table, or other supporting artifact is available on the research page, inspect that material before carrying the conclusion into a brief or report. When you cite a finding, preserve the exact research page, edition context, and the scope language needed to understand what the finding does and does not support.
Turn a relevant finding into a bounded hypothesis for your own site or market. Compare it with current first-party search performance, analytics, technical conditions, customer evidence, and the live search experience you are evaluating. If your local evidence points in another direction, examine why the study context differs rather than treating the published result as a rule. Recheck the citation when the research edition, model, search surface, source set, or measurement approach changes. The most useful research record keeps the original question, evidence boundary, uncertainty, date context, and source attached to the decision so another reader can inspect the reasoning later.
Research areas
What we measure.
Each published edition is dated and documented end to end. More verticals and editions can be added after they pass the same integrity checks.
Questions to resolve before using a finding
- What should I verify before I cite a research finding?
- Verify the study question, edition context, observed population, collection method, search surface or model version, exclusions, classification rules, and stated limitations. Cite the research page that supports the claim and keep enough edition or access context in your record for another reader to understand which evidence you used.
- Can a measured pattern tell me what will happen on my site?
- Not by itself. The pattern describes the sample and method that were actually studied. Use it to form a bounded hypothesis, then compare that hypothesis with current first-party performance, technical conditions, audience behavior, market context, and the live search environment before choosing an SEO action.
- Why might a result differ across research editions?
- The observed search environment can change, including models, search systems, source coverage, brands, pages, and recorded responses. An edition can also use a revised collection or classification method. Compare the method, scope, and date context before treating a difference as evidence of a broader market change.
- How should I communicate uncertainty when I reference a study?
- Describe what the study measured, who or what was observed, the relevant period, and the limits of the design. Label direct observations as observations and interpretation as interpretation. Avoid causal wording unless the method can support it, and point readers to the research method or evidence page so they can inspect the basis for the statement.
- How do I turn research into a practical SEO decision?
- Define the finding as a testable hypothesis tied to the routes or search questions you are evaluating, establish the current baseline, choose the evidence you will watch, and record what would make you continue, revise, or stop the action. Keep the source citation with that decision so later review can distinguish your local result from the original study finding.