About the study
Cyrus Shepard and Dawn Shepard of Zyppy surveyed 131 SEO professionals and collected 13 665 ratings across 103 ranking factors. Each factor was rated on a seven-point scale, from “strongly positive impact” to “strongly negative,” and respondents were also asked to name the three most important factors. The study was published in September 2026 and continues the tradition of Rand Fishkin’s 2005 survey.
An important caveat: these are practitioners’ opinions, not confirmation from Google. The results show what the industry believes, and should be read that way.
Top 10 results of the survey
| # | Factor | Percentage selecting it for the “top 3” |
|---|---|---|
| 1 | Relevance | 57,1% |
| 2 | Links | 54,8% |
| 3 | Content quality | 47,6% |
| 4 | Authority and trust | 36,5% |
| 5 | Behavioral and click signals | 29,4% |
| 6 | Brand signals | 27,0% |
| 7 | User satisfaction | 19,8% |
| 8 | Technical site health | 17,5% |
| 9 | Topical authority | 14,3% |
| 10 | Internal links | 11,1% |
Relevance and intent come first
The highest-rated individual factor was neither a keyword nor a link, but alignment with search intent: how well a page solves the task that brought someone there. Simply matching words won’t achieve that. One respondent puts it plainly: a technically flawless page won’t help if someone was looking for one thing and got something else.
When several pages address the same query, information gain is what matters: does the page add something the others don’t? The finding on meta descriptions is also telling—they scored so low that most respondents don’t believe they directly affect rankings.
Links: reports of their death are greatly exaggerated
Links came in second, even though Google has publicly stated that they are not among its top three signals. The highest ratings went to links from trustworthy domains, topically relevant pages, and sites with real audiences. Spammy links, by contrast, were among the strongest negative factors.
One nuance is anchor text. Exact-match anchors work up to a certain share of a link profile, but beyond that they start to hurt: the proportion matters, not the count. There’s also an AI-related effect: after a brand was mentioned in a major publication, its visibility in chatbot answers increased several times over in a week. Links are no longer just about rankings—they also affect how often a brand is cited in AI-generated answers.
Content: original data matters more than volume
Within the content quality category, original research and first-party data received the highest ratings. The logic is simple: things that can’t be copied from others and are hard to generate become scarce. Original photos, reviews, and data from a company’s own catalog are all signs of effort.
Opinions on AI-generated content are mixed. It isn’t inherently harmful; it becomes harmful when published at scale without adding value. If generated content undergoes rigorous editorial review and draws on unique data, experts see positive results. There’s also a fine line: overly consensus-driven text goes unnoticed, while overly controversial content risks being filtered out.
Brand and behavioral signals
Brand signals are rising in the ratings. Branded search volume, reputation, and trust all score highly, with some describing branded search as the new link. Advertising spend received only a slightly positive rating: it has no direct impact, but may improve other signals by increasing visibility.
Behavioral signals gained ground following data from antitrust proceedings and an API leak. User satisfaction and task completion received the highest ratings. The opposite signal is a quick return to search results after clicking through. Experts consider bounce rate a poor metric, but believe Google notices “pogo-sticking” back to another result. One open question for the future is how to measure behavior as more and more views become agentic and happen without a person.
Technical SEO and internal links
Experts call a sound technical foundation a requirement, not a source of growth. Good technical work won’t rescue weak content, but poor technical work can tank rankings. One real-world example: fixing incorrect canonical URLs led to an almost immediate increase in impressions and clicks.
Many consider internal links an underrated lever that a site has complete control over: they show which pages are important and connect topics to one another. In terms of impact, a quality internal link is compared to a good external one.
User experience and speed
User experience factors again received positive ratings. Mobile usability carries the most weight. Speed and Core Web Vitals, however, were among the most divisive: some experts consider them overrated, while others insist they are critical for large stores. They agree that context matters: a few seconds of load time are tolerable for a simple local site, but not for a large e-commerce site.
What to do in practice
- Start with intent. Check what task the page serves and whether its format matches what the user expects.
- Add what others don’t have. First-party data, tests, photos, reviews, and specific figures set a page apart from similar ones.
- Build links on trust. One piece in a respected industry publication is more valuable than hundreds of links with no audience.
- Build your brand. Branded searches, reputation, and mentions affect both rankings and citations in AI.
- Don’t ignore technical SEO. Indexing and canonicalization errors quietly limit everything else.
- Pay attention to behavior. A quick return to search results is a signal that the page didn’t solve the user’s problem.
Limitations of the study
This is a snapshot of opinions, not a measurement of the algorithm. The respondents are experienced practitioners, but their ratings are subjective and influenced by trends and personal experience. Google doesn’t confirm its signals, and some company statements differ from the survey’s findings. It’s more sensible to use the results as a map of priorities, not as a precise list of weights.