In my opinion lean code is better than bloated code. I doubt that it's ever been a major ranking factor, but wonder why you would say that Google gives it even less weight now.

...why you would say that Google gives it even less weight now.

Modern site design has bloated code. Modern sites do more than deliver text and image content. People do things with modern sites, which requires more code. We can minify but it still remains that modern code is bloated. In any case, only scrapers and spammers create bare bones HTML so that's likely a negative flag now. There was a research paper many years ago about content analysis that noted that lean code was a hallmark of a spammy site and when plotted on a graph it was revealed to be a fact that sites with the leanest code tended to be spam sites.

Lean code is, imo, a deprecated ranking factor. Optimizing code for ranking purposes is a throwback to the caveman days of search marketing, circa 2003.

User experience metrics (all of them)

That's pretty broad right there.

Engaging content that provides an answer, teaches, informs, is useful, delights

Do you have any thoughts on how Google might distinguish between useful, interesting content and blah content (aside from user metrics)? Of the ranking factors on your list, this is the one that looks trickiest to measure.

Without dragging this conversation down to a discussion about bloated code, I think there might be some points on both sides. Techniques such as the bundling and minimizing of scripts do impact page speed, so in that case, addressing code bloat is addressing the user experience -- e.g. load time. Then there are practices like putting your css in a stylesheet versus inlining everything and whether that has any impact on ranking. I'm mean does Google care any more about crappy code vs WC3 compliant code for example? Whatever the answer, I guessing you could add that one to one of the two lists above.

Added: Might as well throw in page structure into this mix (headers, html5 semantic tags, etc).

(Actually I think it's a setup, ork ork)

I'll have a bash ;) I would add:

2015 Ranking
Responsive Design (Supposedly)
Microdata

Deprecated :
Validation
Static Content

only scrapers and spammers create bare bones HTML

Just how wide is that brush you've got there?

There was a research paper many years ago... it was revealed to be a fact that sites with the leanest code tended to be spam sites.

Now way out of date and especially so when using html5.

Lean and accurate code is good, it always has been.

Lean and accurate code is good, it always has been.

Maybe, but it has nothing to do with whether a page is useful to the searcher.

What do you mean by deprecated? Because some of those are still a factor with Google and visitors, such as lean code and long tail keywords. Unless lean code means something else, but if your site has bloated code and loads slow, this will affected your visitors. They dont want to sit there and wait for the page to load.

Do you have any thoughts on how Google might distinguish between useful, interesting content and blah content (aside from user metrics)? Of the ranking factors on your list, this is the one that looks trickiest to measure.

Repetitive content. When you have, let's say, a travel site and you repeat the same basic blurbs mixed in with site location and nothing truly unique to the location, the overall site is eventually going to get bombed. Those who continue to put lipstick on their pig can survive for some time, but the end result is pretty obvious.

Is shorter title tags a factor?

When reworking pages I've been doing that but have also been wondering why.

Google is aware of key words in long titles. And wouldn't they just cleave out the matching part of the title and show it in the serps? (You have less control over what shows but a better chance of showing matching keywords?)

Just asking. I'd like to know what best practices are.

One thing I do do now is leave my site's name out of the title (I'm not really trying to brand anything). I've noticed that with long titles google will still tend to include that (the site name), thus not showing as many key words in the serps title snippet.

My take on the ranking factors mixed with comments from the original post by MartiniBuster

2015 Ranking factors
User experience metrics (all of them) - I'd switch this with don't play cat and mouse with the algos (top heavy, panda, penguin, mobile friendly, pigeon, hummingbird, doorway)
Shorter title tags - Fully Agree
Original content - Agree but still hard to pull off
Engaging content that provides an answer, teaches, informs, is useful, delights - Agree. If you can get content worthy of the knowledge graph, Google will notice. At the very least they officially announced that they liked supplementary content, linking from or to other pages of the site. But this is again more related to the User Experience
Original images - Irrelevant. Sorry but can't agree on this unless you are talking about news/blog type of website. If it is business or e-com, so long as the images are relevant it is fine by Google and the users
Quality site design - define quality. Many people still consider flat design as cheap, disregarding the fact that almost all of the top 100 brands use this type of design. Nor I think this is a metric that Google gives a crap about. I have seen text only websites on the knowledge graph enough times to know that the algo does not care how modern or fancy your website looks like. I'd go with up-to-date code: Html 5, CSS 3, not JS heavy, no flash. Pretty much the basics.
Descriptive meta description - OK. I get the point, but does Google give a hoot about meta descriptions? - Nope. Not to mention many of the trendy niche sites use the meta description to bait more than explain or promote or w/e else is the norm. So while I agree in general, there is more to it than "descriptive".

Deprecated
Keywords - Disagree. Google is still keyword heavy. Believe me when I tell you that if you have less than 1 keyword or synonym mention per paragraph, you will not rank as high. Take it as you wish.
Focus on longtail phrases - Agreed
Focus on ranking for specific keyword phrases - agreed
Lean code - I'd say don't use static code unless you can't avoid it.

My list :

2015 Ranking factors :
Create pages and content for human beings. (broad concept I know but you get the point)
Create pages with user experience in mind.
Implement the keywords you would like to use in a natural manner. Stuffing is a big no-no.
Unique content > spinned content > scraped content ; or in other words, do not be cheap or lazy with your text and image creation
Never forget your bare-bone On-Page SEO - H1, H2 and title tags
Use your Meta description. It does not affect rankings but it does affect your CTR.
If possible create complementary content for your pages on other pages and vice versa.

Deprecated
Keyword stuffing in any way, shape or form
Content creation for the sake of content creation aka content spinning (the plague of 2014 SEO and SEM, in my personal opinion)
Thin pages created dynamically via your CMS solution.
backlink baiting content pages
Long titles (unless you are Techcrunch)
Keyword heavy title descriptions!

Engaging content that provides an answer, teaches, informs, is useful, delights 

Does anybody really think a machine can read, analyze and rate content? If a machine could do this then nothing else would matter as the content is what makes a site. Just the fact that there are so many factors affecting a sites ranking proves the machines can not rate content.

If a machine could do this then nothing else would matter as the content is what makes a site

Many people with over the top good quality content have thought this as well and can't get their sites higher than page ten. Content is just *one thing*. Of many.

Does anybody really think a machine can read, analyze and rate content?

I think it cannot. What it can do is use:
- google raters, then extrapolate to find common points
- use user behaviour
- apply statistics and maths with regards to words / phrase usages etc
- etc.
And then look for patterns. If they can identify a seeded set of a great content and also a set of medium / bad / spam content, they could look for the common denominations that appear only in that set and use these when scoring the content.

But it is still heuristic. This is why they do need to know who wrote it, where does it come from, who links to it etc. This is where the brand comes in - the trust signal.

Maybe, but it has nothing to do with whether a page is useful to the searcher.

That's not what was referred to!

[edited by: aakk9999 at 5:09 pm (utc) on May 7, 2015]

I think meta description is depreciated (just like meta keywords). It's impossible to write a good meta description for dynamically-generated websites and Google does better job creating page description.

WTF? That's not what was referred to!

I think you missed my point, which was that lean code may contribute to the user experience (if it makes a page load noticeably faster) but users aren't searching for pages with lean code: they're searching for pages that meet their needs. As the saying goes, "the proof of the pudding is in the eating," so it makes sense for Google to focus on a page's value to the user, not on the elegance of the page's underlying code.

Someone else mentioned "trust signals." I wonder how much of a role trust plays in the algorithm, and if that role has changed much in the last few years? How do sites earn trust? How do "trust" and "authority" intersect? And is a loss of trust permanent, or can lost trust be regained?

I don't understand why we are debating whether lean code is a ranking factor or not. I don't care if lean code is a ranking factor or not I am going to try to have the best code that is reasonable to create with my limited time and budget. I don't think anyone wakes up and says "Today I am going to intentionally build a website with bloated code that is slow loading".

We all want fast websites and good lean code is one way to improve speed. We often have to be reasonable and use less than perfect code to provide a better user experience with code that is simply GOOD ENOUGH, not totally bloated or perfectly lean but good enough to be fast and provide a nice user experience across different devices and browsers.

I am not a fan of this list of ranking factors and deprecated factors. For example shorter title tags is very vague idea, I think title tags written for user benefit is a better concept than blindly cutting title tags to be too short.

Meta description is a low priority for my own sites since Google tends to dynamically generate them for the serps and I can use that time on more important ranking factors.

As for deprecated factors, I don't understand why keywords is listed. Maybe I don't understand but that sounds like it is ok to create a page for "red widgets" and never mention "red widgets" once on the page. You can debate if it is less critical to mention keyword but I think most SEOs would agree that ideally the keyword should be on the page. Theoretically you can exclude the keyword and rank but why handicap yourself? Excluding keywords is also likely in conflict with the concept of improving the user experience since users tend to engage better with content that mention the keyword and its relevant synonyms. I still come across users that use ctrl+f during usability tests.

Deprecated
Lean code

How do you feel about this:

Deprecated
Code to text ratio with text on the high side

As for deprecated factors, I don't understand why keywords is listed. Maybe I don't understand but that sounds like it is ok to create a page for "red widgets" and never mention "red widgets" once on the page.

I think it's more about "keyword density" being deprecated as a ranking factor.

In answering a query about "red widgets," Google probably looks for pages that mention red widgets, but the fact that a page has "red widgets" in every paragraph or mentions "red widgets" once every x00 words isn't likely to have as much impact as it did years ago.

To borrow a phrase from Matt Cutts, Google is focusing more on "things, not strings."

As for deprecated factors, I don't understand why keywords is listed.

Well, if it's referring to the keywords metatag, it makes perfect sense. Otherwise, it doesn't make sense that Google would match a page with a search term that doesn't even contain the term (or dance around it in some significant way). Perhaps "keyword stuffing" is a more accurate term here.

Added: I'd also like to point out that there may be a BIG difference between "deprecated" and something that can still earn you a penalty.

Keywords are nothing without understanding user intent. Here's the example commonly cited in research papers:

Jaguar

Is it a car? Is it an animal? Is it a band? Is it a song?

The number one result in Google for the query How to tie a prince nymph doesn't feature the word "How" anywhere on the web page. From my computer it's Charlie's Flybox, the site of noted fly tyer, fly shop owner and author, Charlie Craven. That's a darn good result.

What is a keyword phrase?
Is the word "how" a keyword or is it an indicator of user intent? If it can be used as an indicator of user intent and discarded as a required keyword, then what does that mean to you and I about formulating web pages about "How to tie a prince nymph?"

Rephrase or redefine
Perhaps not all parts of a search query are treated as keywords to be matched on a page? Maybe some parts represent entities and some parts indicate user intent? And if so, does that mean you don't have to optimize for How do I, How to, and Where is queries? Should we rephrase the concept of what a keyword phrase is? Redefine it?

Ok, maybe I'm messing with your head a little. The component "how" may exist on an anchor text somewhere. But it still remains that the phrase "How to tie a prince nymph" does not exist on that page. And it has been acknowledged that user intent and a move away from "strings to things" is where the algorithm is going. In fact, much of the future type research seems to me to be moving away from things like matching simple keyword phrases.

Is it premature to mark Keywords as deprecated? Or should it be rephrased or redefined?

Original content
Engaging content that provides an answer, teaches, informs, is useful, delights
Original images
Quality site design

Don't all of these boil down to "a good page"? The question then is: other than making a good page, what on-page factors can be counted separately?

They can't possibly give much weight-- or anti-weight-- to code bloat, considering that I gave up counting stylesheets when I hit a page that listed over 30 (I am not making this up). And that was a page that I'd landed on via a Google search.

Edit:

Is the word "how" a keyword or is it an indicator of user intent?

Isn't that a question of fact, whose answer is knowable? Certain words like "a" and "the" and, er, "and" * are definitely not keywords
:: insert obligatory reference to "The The" -- or, for that matter, The Who -- here ::
I remember back when I looked at keywords in GWT-- which I've long since given up doing-- I was bemused to learn that "it's" counts as a keyword. Even if you're not in the business of selling San Francisco-based ice cream treats, which I'm not.


* Unless you've got content in a language Google doesn't know. (Again, IANMTU).

The list didn't say "keyword density" or "keyword stuffing". IMHO listing "keywords" under the deprecated column is confusing and a potential dangerous misconception. We don't want people to start thinking that they shouldn't be including keywords & relevant synonyms in their content.

I am pretty sure almost every SEO person would agree that it is generally a wise idea to mention the keyword at least once on your page.

Certain words like "a" and "the" and, er, "and" * are definitely not keywords

Those are stop words. Should how be considered a stop word?

I am pretty sure almost every SEO person would agree that it is generally a wise idea to mention the keyword at least once on your page.

Is it wise to list the entire keyword phrase? What about a longtail phrases? Is it wise to use keyword phrases associated with search queries in virtually every title tag across an entire site?