PART 1 OF 7
Why autocomplete becomes evidence
In trademark and competition disputes, parties often point to what Google shows: a rival's name in the suggestion box, a competitor ranking for a branded query, or results that seem to blend two organisations together. The argument is that someone engineered that outcome.
Sometimes that is right. Often it is not. Autocomplete and organic rankings respond to aggregate searcher behaviour, page content and links across the web. A suggestion pairing two names may simply reflect that people who searched for one also searched for the other.
PART 2 OF 7
What the analysis actually tests
A defensible opinion does not start from the screenshot. It starts from the mechanisms that produce what appears on screen and tests each against the record:
- How autocomplete predictions are generated and what typically influences them
- Whether the accused site's pages, metadata or links show targeting of the other mark
- How rankings for branded queries moved over time, and what coincided with the moves
- Whether the observed pattern is consistent with organic behaviour across comparable queries
This is the core of what an expert witness in seo brings to a trademark matter: separating ordinary algorithmic behaviour from deliberate conduct, with data rather than assertion.
PART 3 OF 7
A published example
One of Dan Stratford's published engagements turned on exactly this question:
International Metaphysical Ministry, Inc. v. Wisdom of the Heart Church — U.S. District Court, Northern District of California (2025). Trademark and digital competition. Testimony refuted allegations of SEO manipulation by demonstrating how Google search and autocomplete algorithms work, showing with data that the observed results were consistent with organic behaviour rather than intentional deception. The matter ended in a dismissal with prejudice following resolution.
The value in that kind of testimony is explanatory. Judges and juries rarely have a working model of how suggestions and rankings are produced, and without one, a coincidental pattern can look like intent.
PART 4 OF 7
Common mistakes in autocomplete claims
- Relying on one person's logged-in search, which may be personalised
- Treating a single snapshot as a stable result rather than one moment in a moving system
- Ignoring how the other party's own publicity or news coverage drives searches
- Assuming that ranking for a competitor's name requires manipulation
Each of these can be tested. Each also tends to be raised on cross-examination if it was not addressed in the report. Anticipating them - by running searches logged out, from several locations and over a period of time - turns a vulnerable exhibit into a stable finding that both sides can reproduce.
PART 5 OF 7
Building a repeatable record
Because suggestions and rankings change, a strong report does not rest on a handful of captures. It documents how searches were run - logged out, from stated locations, on stated dates - and repeats them so that the pattern can be checked by the other side. Where historical data is available from rank-tracking tools or archives, it shows whether a suggestion or ranking pattern predated the conduct alleged to have caused it.
The same discipline applies to the accused site. Archived versions of its pages, title tags and metadata show whether it ever targeted the other party's mark. If the site never used the name, and the pattern appears across comparable queries that the site could not have influenced, the organic explanation becomes much stronger. If the site did target the name, that is a finding too, and the report should say so plainly.
Comparable queries are the control group. If similar organisations with no dispute between them show the same kind of paired suggestions, that is strong evidence the pattern is a product of how people search rather than of anything either party did. Choosing those comparisons carefully, and explaining why they were chosen, is what makes the analysis hold up when the other side tests it with its own examples.
Where search evidence meets the legal question
Search evidence answers technical questions: how a suggestion was generated, why a page ranked, whether conduct is visible in the record. It does not answer whether a mark was infringed or whether competition was unfair. Those conclusions belong to the court. Keeping the two separate protects the testimony. An opinion that strays from how search works into what the law requires is the kind most often challenged, while a clear account of the mechanics gives counsel something to build their own arguments on.
That separation is also why preparation matters. Counsel and the expert should agree early on which questions the report will answer, which data will be requested in discovery, and what will be done if the opposing party's own search records are not produced.
PART 6 OF 7
Confirming credentials independently
Counsel evaluating an expert for search-related testimony can review an independent directory profile rather than relying only on a practice site. Dan Stratford's profile on the SEAK expert witness directory, under the heading seo expert witness, lists his credentials, years in practice and deposition history. His education is BS Biology, Rockhurst University; MBA, Regis University, and he has 29 years in digital marketing, beginning in the mid-1990s.
PART 7 OF 7
A note for law firms watching their own search results
Law firms are frequent participants in branded-search competition themselves, and how a firm appears in search is shaped largely by what it publishes. For firms reviewing that side of their practice, a write-up on legal content strategy offers further reading on planning legal content.