Google does not use word count as a ranking factor at all, in either direction. John Mueller has stated this directly and repeatedly, including telling one site owner plainly that “word count is not a ranking factor, save yourself the trouble,” and separately noting on LinkedIn that nobody at Google is counting the words or links on a page. Depth, as Google’s systems and documentation describe it, is instead an assessment of whether the content fully satisfies what a searcher needs on that topic, whether it reflects information or insight the writer could only have from direct experience or genuine expertise, and whether someone leaves the page feeling they learned something they couldn’t have gotten from a dozen other pages saying the same thing. Length is a byproduct of covering a topic properly, not a proxy for having done so, and Google has been unambiguous that piling on more text to hit a number does not make a page more helpful.
Mechanism
There’s no disclosed “depth algorithm,” and it would be inaccurate to describe one, but Google has published something more useful than a black-box scoring model: a set of self-assessment questions in its “Creating helpful, reliable, people-first content” guidance, originally tied to the Helpful Content system and now folded into core ranking guidance. These questions are the closest thing to Google’s own articulation of what depth and comprehensiveness actually mean in practice, and they’re worth treating as the literal rubric rather than paraphrasing loosely.
On content and quality specifically, Google asks whether the content provides original information, reporting, research, or analysis; whether it provides a substantial, complete, or comprehensive description of the topic; whether it provides insightful analysis or interesting information that is beyond the obvious; and whether, if drawing on other sources, it avoids simply copying or rewriting them and instead adds substantial additional value and originality. It also asks whether the content provides a satisfying experience for people who just had an experience of visiting the page, and whether readers would feel they’d had a good experience.
Google’s “Who, How, Why” framing, part of the same documentation, adds another dimension. The “who” questions ask whether it’s self-evident who authored the content and whether the site or author has established expertise or firsthand experience on the topic. The “why” questions ask whether the content was made primarily for people rather than to attract search engine visits, and whether it exists because the site already has an audience that would find it useful on its own terms. This matters for the depth question specifically because comprehensiveness without genuine expertise tends to read as comprehensive-shaped rather than actually comprehensive: it hits the expected subtopics because a competitor analysis said to, not because the writer has enough command of the subject to know what a reader genuinely needs.
Padded long-form content fails these self-assessment questions in a specific, recognizable way. It restates the same point in different phrasing across several paragraphs, includes sections that summarize what a reader already inferred from the introduction, adds tangential background that doesn’t change what the reader does next, and covers breadth (many subtopics, shallowly) rather than depth (the subtopics that matter, thoroughly). None of that is a word-count problem in the sense Google means; it’s a comprehensiveness and originality problem that happens to produce a lot of words as a side effect.
Practical implication
Audit content against the self-assessment questions directly rather than against a length target. For a given page, ask concretely: does this contain at least one piece of information, framing, or analysis a reader couldn’t get from the top three competing results, verbatim or close to it. Could someone with genuine hands-on experience in this subject read this and find it credible, or does it read like a synthesis of other synthesis. Is every section doing work, meaning would removing it lose the reader something, or is it restating the intro in new words to look thorough.
Practically, this means the editing target shifts from “expand this section” to “cut what fails the substitution test” and “add what only someone with direct expertise would know.” A 900-word page that answers a narrow technical question completely, with a specific worked example or a genuine edge case the writer has actually encountered, will out-rank and outlast a 3,000-word page that restates common knowledge with more words per idea. If a topic is genuinely broad and the honest answer requires covering many subtopics, length follows naturally from that breadth, and that’s fine; the point isn’t that long content is bad, it’s that length is never the goal and is never assessed as one. The only reliable way to know if a page has genuine depth is to ask, section by section, whether a reader who already knows the basics would learn something specific and non-obvious from it, which is precisely the standard Google’s own guidance describes.
