A poor 404 page and user interaction with it, can only lead to a ‘poor user experience’ signal at Google’s end, for a number of reasons. I will highlight a poor 404 page in my audits and actually programmatically look for signs of this issue when I scan a site. I don’t know if Google looks at your site that way to rate it e.g. algorithmically determines if you have a good 404 page – or if it is a UX factor, something to be taken into consideration further down the line – or purely to get you thinking about 404 pages (in general) to help prevent Google wasting resources indexing crud pages and presenting poor results to searchers. I think rather that any rating would be a second order scoring including data from user activity on the SERPs – stuff we as SEO can’t see.

Sometimes, Google turns up the dial on demands on ‘quality’, and if your site falls short, a website traffic crunch is assured. Some sites invite problems ignoring Google’s ‘rules’ and some sites inadvertently introduce technical problems to their site after the date of a major algorithm update and are then impacted negatively by later refreshes of the algorithm.
QUOTE: “Medium pages achieve their purpose and have neither high nor low expertise, authoritativeness, and trustworthiness. However, Medium pages lack the characteristics that would support a higher quality rating. Occasionally, you will find a page with a mix of high and low quality characteristics. In those cases, the best page quality rating may be Medium.” Google Quality Evaluator Guidelines, 2017
I do not obsess about site architecture as much as I used to…. but I always ensure my pages I want to be indexed are all available from a crawl from the home page – and I still emphasise important pages by linking to them where relevant. I always aim to get THE most important exact match anchor text pointing to the page from internal links – but I avoid abusing internals and avoid overtly manipulative internal links that are not grammatically correct, for instance..
This broken-link checker makes it easy for a publisher or editor to make corrections before a page is live. Think about a site like Wikipedia, for example. The Wikipedia page for the term "marketing" contains a whopping 711 links. Not only was Check My Links able to detect this number in a matter of seconds, but it also found (and highlighted) seven broken links.

So you have a new site. You fill your home page meta tags with the 20 keywords you want to rank for – hey, that’s what optimisation is all about, isn’t it? You’ve just told Google by the third line of text what to filter you for. The meta name=”Keywords” was actually originally for words that weren’t actually on the page that would help classify the document.
QUOTE: “Returning a code other than 404 or 410 for a non-existent page (or redirecting users to another page, such as the homepage, instead of returning a 404) can be problematic. Firstly, it tells search engines that there’s a real page at that URL. As a result, that URL may be crawled and its content indexed. Because of the time Googlebot spends on non-existent pages, your unique URLs may not be discovered as quickly or visited as frequently and your site’s crawl coverage may be impacted (also, you probably don’t want your site to rank well for the search query” GOOGLE
Hi Noya, all the info suggests that dwell time IS taken into account in search ranking, and we know that Google measures time on page and bounce rate in Analytics, too. Plus the search engine gets smarter all the time. With the machine learning component of RankBrain, we wouldn’t be surprised if Google can tell the difference between sites where visitors stick around, bounces where the visitor gets an answer immediately, and bounces where the visitor keeps searching.