h1 h2 and h3 tags In Google etc.

header tag question

[javelin]

Google has stated that multiple h1 tags in a page does not impact ranking. Yet historically it once upon a time held some weight. For example you can look at Bing and they still will give you a warning if multiple h1s are on the page.

My question is how many here still are strict about header uses ie 1 H1 and following subsequent historical uses? I for example will use typically no more than 2 h1 tags.

Do you believe it still will impact page ranking in the same way as it has in the past, or how Google and others will see keywords perhaps placing prominence on different focus words?

[Brett_Tabke]

Googles expanded usage of a rendering engine to analyze content, tends to make me think that <h> tag value is becoming a thing of the past. How Google uses that engine is anybody's guess. They have only said it exists and what version of Chrome they are using for it.

I asked Gary Illyes about this very thing when he keynoted Pubcon Florida for us a couple years ago. There was no cut-n-dried answer from him, or from the Twitter spokes people.

Lets be clear, (as is everything with Google SEO these days) these are totally guesses: CSS constructs make it easy to devalue H tag weight, or increase stock font weight without any apparent difference between the two. However, a good portion of the web still uses H tags appropriately and are good signals for Google to use.

The only issue I see to valuing H tags is the question of how is Google using rendering data? We could make quite a few guesses, but I think they would all be wrong. Between machine learning, and visual indicators, it is a blank slate of possibilities, that only Google's vast trove of user data could point in the proper direction.

I say - when you have a marketing lack of direction then spaghetti spray-n-pray is the rule of the day. eg: use it where appropriate for the user and hope for the best with the engines.

[javelin]

use it where appropriate for the user and hope for the best with the engines.

This has kind of been my philosophy with their use. Thanks for the input on that. I hate "writing" for an engine when it should be for the reader/user of the site. This is why Bing irritates me with their "warning you have multiple h1 tags on xyz pages". I do enjoy the fact that some things we cannot see so we write and hope that in the end they conform to what a reader would expect to see.

[not2easy]

There still is no one size fits all for rules on headings so you should be fine so long as it isn't a case of using <h tags instead of <b> or using header tags for styling. They do have a purpose, but you are not restrained from using them to fit the content of your page.

I'm accustomed to using <h tags for headings as intended but I do not think it would hurt ranking to have more than one <h1 tag. Most of this is common sense, not to abuse markup for gimmicks. I think WordPress caused some confusion years ago because it auto added <h1 tags so if you added your own, there were two. Google got over it, apparently Bing didn't. I'd say if doing better in Bing is the goal then use one <h1 tag.

I don't think if I suddenly noticed that I had 200 pages with more than one <h1 tag that I would stop everything to go and fix that.

[tangor]

According to the existing spec (which few ever follow) there's only one H1 for the page OR section, both of which are different. H1 can also appear in captions.

Sadly, even with all the CSS wonderful we have out there, h tags are still being used to "style" fonts instead of their intended purpose to OUTLINE a document.

[JesterMagic]

@tangor one per section... that doesn't seem right for the thinking now days in regards to accessibility. Most pages have many sections (not doubting your word it says that though)

I use 1 H1 per page and use the headings to outline documents to make it easier for people who use screen readers.

Ideally Google should rely heavily on this as well to help categorize the content on the page for the search results but from my experience fixing the heading structure of the page doesn't seem to cause any notable increase in ranking.

[RedBar]

I have six H sizes in my style sheet, I only ever use H1 the once, then again I do follow this:

According to the existing spec (which few ever follow) there's only one H1 for the page OR section

FWIW I do not use an H1 for a Section.

[javelin]

At some point as I have time I may go back and review my pages for this. I have never used them for styling and I am not worried about it impacting my seo strategy. Yet for posterity or others that may come after me at least I will have created a standard for how things "should" be run in my wordpress stuff.

As ive stated I only ever use 2 per page. For documentation or research sites I want to change the flow a bit.

[aristotle]

impact ranking

Do you mean the possible impact on the ranking for a specific keyword?

As I recall, it was generally thought at one time that including a keyword in a header, or several headers, might give a small extra boost to a page's google ranking in searches that involved that keyword. In my opinion this might be the case even now.

But I wouldn't recommend trying to "stuff" particular keywords into your headers. It should happen "naturally" as a way to convey the content contained in various sections of the page.

[lucy24]

For arcane historical reasons, most of my pages start at h2 with no h1 at all, and the usual header is h3. I still get visitors sent by G###. Does this mean that if I bumped up to h1 I would get more traffic?

Nah. Don't think so. It's just another of the 1001 factors that get thrown into the bin.

[Marshall]

One thing to keep in mind, Google aside, is how screen readers see <h> tags. Not following a "traditional" tree layout or using them for styling can be confusing for a screen reader. [w3.org...]

[RedBar]

@Marshall - Thanks for reminding me of W3. It was their guidlines I followed 10 years ago when creating my first template site creation for html5 along with w3schools.

[lucy24]

Heh. To this day I don't put anything online before running it through the w3 validator trio: html, css, link checker. I remember having to complain to The People In Charge when the html and css validators gave opposite and mutually exclusive responses to one element.

[javelin]

Not following a "traditional" tree layout or using them for styling can be confusing for a screen reader.

Very very true.... good reason to implement changes putting seo reasons aside.

[robzilla]

Modern search engines aim to judge pages based on what real users are seeing on their screens, not on what webmasters tell them is most relevant or important. So you can throw your book of old-school SEO tricks out the window, that's all been largely irrelevant for years now. They'll render your pages just like a browser would and then interpret them visually and textually. Then they'll probably assign labels, like "this page has lots of photos", which they can later combine with query intent. People searching something like "beautiful beaches in europe" will probably expect results rich in photos, not text. In interpreting what a page is about, this would suggest that a <p> element styled to 30px would be assigned more "weight" than an <h1> styled to 14px. Another good reason to do it like that is because with everyone and their dog creating content these days, not everyone can be expected to know that elements like <h1> even exist, and that needn't be used against them; if the content is the same visually, does it matter what elements you use?

As noted, for screen readers, of course, it certainly does. So, SEO aside, I'd suggest you just stick to the "rulebook" and don't use H1 more than once.

[JesterMagic]

@robzilla schema, Semantics and structured data all helps with SEO and I wouldn't call them tricks. A page put together properly ranks better (I am not saying this is a huge ranking factor) than a page which is not as search engines can understand what the content is (in a very general way) better. This also means your page has a better chance to be a featured snippet etc. The downside to this though (since Google understands your page better) is Google more likely can display the answer the searcher is looking for in the search results which means the searcher has no reason to actually visit your page.

Yes Google renders pages like a browser for a while now which also helps in understanding content (to a point because IMO the algorithm still is pretty dumb). The main reasons for doing this was so CSS and javascript could be processed and the results rendered so the entire content of the page would be available and not just what was on the initial HTML page.

[robzilla]

@robzilla schema, Semantics and structured data all helps with SEO and I wouldn't call them tricks.

Used properly they're hints, used improperly they're tricks ;-) I was mostly referring to the idea of using H1 multiple times, which is definitely a trick (that may actually have worked 20 years ago; now I'd say it's more likely to trigger a spam algo).

The main reasons for doing this was so CSS and javascript could be processed and the results rendered so the entire content of the page would be available and not just what was on the initial HTML page.

Initially, sure, but I'd be very surprised if they haven't trained an AI system to interpret pages visually, fed by tons of data, including the quality raters probably.

[martinibuster]

Headings matter but not in the traditional way of "You Puts the Keywords in the Baskets" way that SEOs commonly use it for.

Heading weight is no longer a thing and it hasn't been a thing for many years.