Meta Descriptions, Domain Authority and Four Other Ranking Factor Myths
Founder, OrganicRank. Builds the crawler, the scoring and the fix pipeline described in these posts. Every claim here is one you can reproduce on your own domain.
Most SEO myths survive because the thing they describe does matter - just not in the way the myth says. A meta description will not rank your page and will absolutely change how many people click it. Here are six, each with what Google has actually said and what the honest version is.
The short version
- Meta descriptions are not a ranking factor. Google confirmed this in 2009 and has restated it since; it also rewrites them for the majority of results.
- Domain Authority is a Moz metric, not a Google one. Google has said plainly it has 'no notion of overall domain authority' as a ranking score.
- Keyword density has never been a factor. There is no threshold to hit and no mechanism by which repetition would signal quality.
- Bounce rate and time on page are not read from Analytics. Google does not use your Analytics data to rank your site.
- Each of these still matters - for click-through, for competitive benchmarking, for readability. The error is calling them ranking factors.
Are meta descriptions a Google ranking factor?
No. Google confirmed in 2009 that meta description content is not used in its ranking algorithms, and has repeatedly restated it since. It affects whether people click your result, which is a real and separate benefit.
The original announcement is unambiguous and still the clearest statement of it: Google's Webmaster Central blog said that the meta description tag is not used in ranking, while the meta keywords tag is not used at all. Nothing in the intervening years has reversed it.
What complicates the picture is that Google frequently rewrites the description it shows. Studies of large result sets consistently find that Google generates its own snippet for the majority of queries, choosing a passage from the page that matches the specific search. So even the click-through benefit is partial - you are writing a suggestion, not a guarantee.
The practical conclusion is not to skip them. It is to write them as ad copy for the query the page targets, accept that Google will sometimes prefer its own, and stop treating a missing description as a ranking fault. It is a conversion fault, on the results page.
Is Domain Authority a Google ranking factor?
No. Domain Authority is a third-party score created by Moz to predict ranking strength from the link graph. Google does not read it, and has said it has no equivalent overall domain-level score in its ranking systems.
Google's Search Advocates have addressed this repeatedly and directly, stating that Google has no notion of overall domain authority as a single score applied across a site. The clarification that usually follows is important: Google does use some site-level signals, so 'domain authority is not a thing' overstates it in the other direction. What does not exist is the number.
This matters practically because of how DA is used. As a comparison metric between you and the sites outranking you, it is genuinely useful - it is built from the link graph, and the link graph does drive rankings. As a target to optimise, it is meaningless, because you would be optimising for a third-party model rather than for the links the model is estimating.
The tell is when a report recommends 'increase Domain Authority to 40'. That is not an action. Earning fifteen relevant referring domains is an action, and DA moving is a side effect worth measuring, not the goal.
Does keyword density affect rankings?
No. There is no target percentage, there never was, and Google has explicitly said there is no ideal keyword density. Repeating a phrase does not make a page more relevant; it makes it worse to read.
The useful test is mechanism. Ask how a signal would help a search engine return a better result. A page that says a phrase eleven times is not more useful than one that says it twice and answers the question - so there is no reason for a ranking system to reward it, and modern retrieval works on meaning rather than term counts anyway.
What is true is that a page should contain the language of the query, once, in the places that establish what the page is about: the title, the H1, the first paragraph. That is intent match, not density, and the difference is that intent match is satisfied by saying the thing clearly one time.
Does Google use bounce rate or time on page?
Not from your Analytics. Google has stated it does not use Google Analytics data in ranking. The broader claim that it uses no user-interaction signals at all is not supportable either - but bounce rate as measured in your dashboard is not the signal.
This one deserves care, because both extremes are wrong. Google's representatives have consistently denied using Analytics metrics, and the reason is straightforward: the data is trivially manipulable and most sites do not have it installed. A ranking system cannot depend on a signal that half the web does not emit.
At the same time, court documents released during the US antitrust proceedings described systems that use click and interaction data at scale. Those are aggregate behavioural signals derived from the results page, not your bounce rate. The honest summary is that user behaviour matters in some aggregate form and your Analytics numbers are not how Google sees it.
The practical upshot is unchanged: a page that satisfies the searcher is the goal, and the bounce rate in your dashboard is a diagnostic for you rather than a lever on Google.
Do exact-match domains and domain age rank better?
No to both, in the sense implied. Google released an update targeting low-quality exact-match domains in 2012, and has said domain age is not a ranking factor. Older domains rank better because they have accumulated links, not because they are old.
Exact-match domains are the clearer case: buying bestwidgets.com does not rank you for 'best widgets', and Google shipped an update specifically to reduce low-quality EMDs in results. A keyword-matching domain neither helps nor harms a good site; it just is not the shortcut it was fifteen years ago.
Domain age is the more instructive myth, because the correlation is strong and the mechanism is wrong. An eight-year-old domain typically has eight years of links, mentions, and content. A domain registered eight years ago and parked has none of that and ranks accordingly. Age is a proxy for accumulation, and the accumulation is the factor.
| Commonly called a ranking factor | Ranking factor? | What it genuinely affects |
|---|---|---|
| Meta description | No | Click-through rate, when Google uses yours |
| Domain Authority | No | Useful third-party benchmark against competitors |
| Keyword density | No | Readability - negatively, past a point |
| Bounce rate from Analytics | No | Your own diagnosis of which pages disappoint |
| Exact-match domain | No | Brand recall, sometimes; nothing algorithmic |
| Domain age | No | Proxy for accumulated links and reputation |
How do you check a ranking factor claim yourself?
Look for a primary source, prefer a dated statement from Google over a correlation study, and ask what mechanism would make the signal work. Most myths fail the mechanism test before you even reach the sourcing.
The reason this discipline pays is that the myths are not randomly distributed. They cluster around things that are easy to measure and easy to sell - a density percentage, a score out of 100, a number in a dashboard. The factors that actually carry weight are harder to reduce to a number, which is precisely why the proxies became popular.
- 1Find the original claim. A great deal of SEO advice cites another blog post that cites another, and the chain frequently terminates in nothing.
- 2Prefer Google's own documentation and named representatives, with a date attached - guidance genuinely does change.
- 3Distinguish correlation studies from causal claims. 'Pages ranking on page one tend to have X' does not establish that X causes ranking.
- 4Apply the mechanism test: how would this signal help a search engine return a better result? If there is no plausible answer, be sceptical.
- 5Check whether the claim is falsifiable in your own data. Many are, cheaply, on a handful of pages.
Sources
Every figure above comes from the vendor's own published material. Check them.
- Google Webmaster Central Blog - Google does not use the keywords meta tag in web ranking (2009)
- Google Search Central - Control your snippets in search results
- Google Search Central - Creating helpful, reliable, people-first content
- Google Search Central Blog - Domain name and ranking (EMD guidance)
- Moz - Domain Authority, and what it is (third-party metric)
Frequently asked questions
Should I still write meta descriptions?
Yes - as ad copy for the query the page targets, not as a ranking tactic. Google will often substitute its own, and when it uses yours a better-written one earns more clicks.
If Domain Authority is not a Google factor, why do tools report it?
Because it is a reasonable estimate of link-graph strength and it is comparable between sites. That makes it useful for benchmarking against competitors. It becomes misleading only when treated as a score Google reads.
Is there an official list of Google ranking factors?
No. Google documents categories of signals and confirms or denies specific claims, but has never published a ranked list. Any list presenting itself as official is inferred, and usually inflated by splitting one factor into many.
Do these answers change over time?
Some do, which is why every claim here carries a date and a source. Meta descriptions have been confirmed as non-ranking since 2009 and are unlikely to change. Anything involving user-interaction signals is the area where public understanding has moved most.