SEO myths, debunked.

Almost every myth here shares one shape: a proxy gets mistaken for the real thing. A number that was once a decent stand-in becomes the target, a tool grades you against a rule it invented, or a genuine signal gets inflated into a lever it was never strong enough to be. Learn to spot that shape and most of these stop being surprising. Below they are grouped by where they do the most damage.

Signals people overrate

These are real things that Google either uses lightly or does not use at all, inflated in the retelling into the reason a page lives or dies. Core Web Vitals are a tiebreaker, a small signal that settles close calls rather than rescuing a weak page. Bounce rate is not a ranking factor at all, because it lives in your analytics and Google Search cannot see it. Meta descriptions do not affect rankings and get rewritten for most results anyway. And keyword density is not a thing, just a fossil of an era that ended with keyword stuffing. Its close cousin, word count, is a correlation with the causal part removed: thorough pages are long, which is not the same as long pages being thorough. Schema markup is not a ranking factor either, though it is sold as one hardest of all: it buys eligibility for a richer listing, not a position. And LSI keywords are not a thing, being a 1980s technique for small document collections that Google has said outright it does not use. Social signals are not a ranking factor: shares are not counted and the links are nofollow, so the value runs through distribution rather than the mechanism being sold. And you can have more than one H1, a rule that survives mostly because counting elements is easy, as does the habit of treating alt text as a keyword slot when it exists to replace the image for someone who cannot see it. Titles are worth a separate mention: Google rewrites your title tags too, not only your descriptions, so the tag is a strong suggestion rather than the headline.

Penalties that are not penalties

The word penalty gets attached to any drop, because it puts the blame outside the site. Usually there is no penalty at all: you are not being penalised, you are being outranked, and a real manual action announces itself in Search Console. There is no duplicate content penalty either, only Google picking one URL and setting the others aside. And the tool that promises to clean up your links is a loaded gun: routine disavowing usually demotes links that were helping.

Getting found and indexed

A cluster of myths surrounds how Google discovers and indexes pages. You do not need to submit your site to Google, which finds sites by following links. The Google sandbox is a myth, though new sites really do wait, for the plain reason that they have earned no signals yet. Robots.txt does not stop indexing, it controls crawling, which is a different thing people reliably confuse. And in Search Console, page with redirect is not an error, it is a description of a redirect doing its job. And the one people reach for when a page will not index: an XML sitemap does not get pages indexed, it only says the URLs exist, which is why resubmitting it for the tenth time changes nothing. In the same family, 404 errors do not hurt your rankings, since a 404 is the correct answer for something that is gone, and a 301 does not lose 15% of your link equity, a figure misremembered from a damping factor that never described redirects. And the most satisfying wrong answer of the lot, crawl budget is probably not your problem, which relocates a quality question into an infrastructure one.

Links, authority and third-party metrics

Authority is where vendor numbers get mistaken for Google's judgement. Domain Rating is not a Google metric, it is an Ahrefs estimate, useful as a reading and dangerous as a target. And on the other side, nofollow links still have value, because nofollow is a hint now rather than a firewall, and a link was never only a pipe for authority in the first place. The same confusion sits behind domain age, which confers nothing on its own and merely stands in for everything a domain accumulates while it sits there. Nor is the name itself a lever: exact-match domains are not a shortcut, having been targeted in 2012 and reduced to a very small factor since. And the number you report deserves the same scepticism, because your ranking is not a single number: it varies by location, device and history, so a tracker reports one sample rather than your position.

Content, strategy and the newer myths

The last group is about effort spent in the wrong place. You do not need a blog post for every keyword, because a page should map to an intent, not a phrasing. Traffic is not the goal, and more of the wrong traffic actively hurts. Most SEO audits are a crawl export with a logo on it, long precisely because nobody did the hard part of deciding what matters. We explain why we stopped adding FAQPage schema once its rich result was taken away from most sites. And the newest scare, AI Overviews are not the traffic apocalypse you were sold, just a shift in which queries send clicks. Alongside it, Google does not penalise AI content: there is no authorship penalty and no detector accurate enough to build one on, only the old rule that content with nothing in it does not deserve to rank, whoever or whatever typed it. The same applies to how often you publish: publishing frequency is not a ranking factor, and a cadence target is a reliable way to manufacture the thin pages the rest of this list warns about.

The pattern worth keeping

If there is one thing to take from the whole series, it is the habit of asking, of any SEO rule, where it came from and whether it is the real thing or a stand-in for it. Proxies are useful right up until they become the goal, at which point they get gamed and stop measuring anything. Read the numbers, all of them, and never let one become the brief. The brief is the business outcome, and almost every myth here is a proxy that forgot it was only ever standing in for that.