On Page Search Engine Optimisation IS Dead

Is there any real need for on page optimisation anymore? All of my research on the SERPS seems to be pointing in 1 direction and that is, when it comes to high rankings for competitive terms only your link profile really matters.

After all Google is still ranking number 2 for the keyword “caffeine” even though all it does is slap it in the title tag and has a shed load of links, the result is only relevant to somebody searching for it from an internet marketing background, those wanting information on a mild drug are going to be disappointed.

Anyway, as well as examples like the above and ongoing comments in the SEO industry about ‘how headings are no longer of value’ I decided to carry out a little research myself.

The research was based on the heading tags h1 – h4 and on general keyword density. Like with so many of the other tests I perform this one was based on 25 search terms and the top 8 results were examined.

Headings Optimisation

I guess the first place to start is with the headings, it is an area I also believe has lost most of its value but let’s peak at the results before coming to a conclusion.

For those of you who aren’t sure how headings are supposed to work with SEO, basically it is taught that you should use your keyword in them to build relevance h1 holding the most weight and decreasing accordingly to h6.

The first test was simple, ranking pages 1-8 were tested to see if they had optimised their heading tags for the target keyword.

A score was given to each heading optimised on each site h1 = 8 h2 = 6 etc… The ranking position with the highest score had obviously better optimised headings.

heading optimisation graph

As you can see from the graph the blue line represents the true ranking order and the red line is how it would rank if rankings were based on optimised headings.

There are massive differences here and even though this is correlation and not causation I still feel it is a strong indication that headings are not quite as valuable as one may think.

If you look at the below chart;

headings optimisation chart

62.5% of pages on the first page have the h1 optimised for the target keyword, however 0% have the h4 optimised.

I can also confirm that 15% of the pages tested had not optimised any headings.

The only thing we can take from this is that if you are going to optimise your headings, stick to h1 and h2 the rest seem to be irrelevant but don’t expect it to work miracles for your rankings because it won’t.

Keyword Density

In this test the same 25 pages were checked for keyword density rates against the target phrase. The percentages were taken and the below chart shows the results.

keyword density chart

Even I was surprised at these results, when re ranked the sites there was almost a complete reverse with the lowest ranking sites having the highest keyword density.

So what does this mean??

Keep things to a natural level use the keyword naturally but don’t feel under pressure to use it in every paragraph because it clearly does not matter.

Where does this leave on page optimisation?

Well what can I say?

First of all there is a lot more to on page search engine optimisation other than keyword issues;

– Crawlability
– Internal Linking
– Sitemaps
– Click Paths
– Code to Content Ratio


However in my opinion the days of stressing over keyword density, keyword proximity or prominence are over along with heading optimisation, it just is not as powerful as it once was.

The only place to worry about your keywords, in my opinion, is in your title tag, outbound links and obviously try and get it in the content a couple of times. Apart from that the jobs a good one.

There are obviously a few areas I didn’t touch on;

– How does this apply to longtail keywords?
– Does keyword position matter? Top, bottom, sidebar, footer
– I also did not take into account image alt attributes

These are all posts for another time. So quit worrying about how many times you’ve stuffed your keyword into your content and start building some links 🙂


is the CEO of Branded3 a Search Marketing Agency in the UK. Tim has over a decade of experience in Search and regularly speaks at key events and conferences.

10 thoughts on “On Page Search Engine Optimisation IS Dead

  1. Almost boiling down to titles and link text. You can let the text flow naturally without trying to squeeze in that last phrase so I guess it’s for the better. Hasn’t everyone stopped stuffing for a while now? I recently re-wrote some content because even though it flowed natural I was worried the key phrase was too dense and feared a penalty. I’m not sure there is such a penalty but I made sure I’m not at risk. That might be a good test, take a top 10 SERP and stuff it to see if it goes down, any volunteers 🙂

  2. Hi David,

    lol, ye I think stuffing went out a while ago, however I have to admit to seeing quite a few sites bolding everything in the first paragraph and keyword stuffed title tags. I have a few sites I could test the stuffing on, the problem is they don’t really rank for anythign and the ones that do I don’t want to risk, I’ll have to see what I can do. I really think on page is all about natural content and making sure google can find it and index it, the rest is in the links, however I am going to look at some longtail serps and see what the differences are.

  3. So it’s

    ‘mould the title – write what you like – be happy’

    Sounds good.

    Now just to get those top ranking sites to link back to ours.


  4. David,

    Yep that is pretty much it. Let me be clear though, longtail less competitive terms still seemed to be weighed against on page factors so I would stillbe mindful of incorporating keyword phrases into deep pages and posts.

  5. Are you a web designer? Because you definitely sound like one:)

    Double-check your data, because SE’s (not only google) still have their ranking based on things such as POSH, SOC, keyword proximity and density algorithms. That is, semantics.

    Not to mention that there are rumours of SE’s taking the human factor from social networks, etc.

    Your research points exactly the other way. Get 100 people to point their fingers at a trash can and it will be a hit?!

    If you don’t believe me it’s ok, but take a look at your comment count. If your data had something true, it would be a major discussion. Not to mention I was referred by a SEO’s profile in Facebook.

    So, think again. 🙂


  6. @yMladenov,

    Ok, let me start by saying I respect any opinion when it comes to SEO, however the info in this article does not relate to opinion, it is factual correlation data. Given this let me try and address your points.

    First, no I’m not a designer, in fact my design skills are probably pretty amateur. I am an SEO and have ranked many sites for competitive terms not to mention ranking this site “google seo”, “seo techniques” etc…. after 6 months of work.

    Second I am not saying don’t have a density, proximity, prominence edge to their algorithm. What I am saying is that when it comes to ranking for competitive terms the power of your link profile counts with on page keywords taking a clear back seat.

    I take your point about semantics and word relations however I remain skeptical about it, not saying Google don’t have the technology but I see no real evidence of them using it. (Search ‘car’ and then ‘cars’ 6 of the results on the first page differ between keywords – is that really a semantic algorithm??)

    Third – My research is what it is, a low percentage of sites are making use of their headings and the sites with the higher keyword density aren’t the sites ranking the highest for their target keyword. If you don’t trust my opinion on the matter maybe one of the biggest SEO sites may help you see that on page and heading optimisation is losing value. (Rand Fishkin links back to some more of my rubbish unfounded research from the post, you’ll have to tell them to check their sources, as they are linking back to a designer talking rubbish 🙂 )

    Fourth – The post is not intended to be linkbait, it is not meant as an opinion, these are facts, the sites were structured as the results show. Dust bin??? Thats harsh but I’ll let you.

    Fifth – There is 2 reasons why comments are low on the post 1) My blog is small and still making a name for itself (but growing fast) 2) Most experienced SEO’s know that onpage factors are losing value, especially when it comes to high value keywords.

    I am sorry if this comment sounds a little bitchy, but I think your comment was a little aggressive.

    I tell you what so were both happy, you keep laughing at my silly research and mocking it and go back to concentrating on density, proximity and that other stuff, I”l build a solid, optimised link profile and issuing crappy research posts that get picked up and linked to my seomoz.org.

    I thought again I hope the response is OK.

    (This comment is a serious waste of my time, so I hope anyone who reads it appreciates it 🙂 )

  7. Hey, Tim!

    I just got back here wandering through my inlinks reports from yahoo..

    I’m sorry for the late responce, but I don’t usually post a comment with the intention of starting a discussion.

    You’re not bitching at all, in fact, I’m delighted to see an author replying to such harsh and groundless criticism in a nice way. I don’t think your comment was a waste of time. No written content is a waste of time on the internet. 😉

    So, back to our topic.
    Well, I’m pretty much going to disappoint you with my response – I don’t really have any data to compare with yours to give you a good point because I optimize for a local market, and with that said I can only tell you that we’re on the opposite sides of the line.

    There are SEO’s who don’t give crap about semantics and there are semanticists who don’t really go too deep with link building. No need to define who is who, I hope.

    About Rand linking to your stats – I’m not questioning your credibility here. I’m pretty sure you’re doing way better than me, due to everyday life-related factors or simply to commitment to SEO. I greet you for that.
    However, one thing I learned by dealing with business persons is that no matter how credible the guy is, mistakes are human, and as a partner said once “you shit, just like everybody else”.

    Also, I still think your comment count would’ve gone up if your stats were “the thing”. Or maybe they are, but not many people appreciate that? 😉

    And so, without saying something too meaningful, I think we’re just going for a different approaches.

    I started learning semantics about 2 years ago and here’s something that really boosted my studies in the right way – http://bit.ly/3CJFGh
    You might have already read that, but if not – I hope you find it a solid ground for research.

    And don’t mind comments like my previous, I’m sure you’re about to see many like this in your web experience. Everyone has the right to have a bad day, once in a while, is it not so? 🙂

    Have a happy new year celebration.


  8. p.s. if you haven’t already I think you gotta get O’reilly’s SEO Warrior. Pretty useful resource by my opinion!

  9. Hi yMladenov,

    Wow I almost forgot abou0t this discussion.

    I do fully take your points on board. I do believe Google have the technology to do semantics or LSI, however I simply see no evidence. I am not narrow minded enough to think I can build a sites link profile and nothing else, it simply astonishes me how much weight Google put on links.

    The last comment back was a knee jerk reaction done on my iPhone, so I probably could have thought it through a little better. The research I did was simply a view based on facts, I wish on page optimisation was assessed a little different but thats just not the way at the minute.

    If you have any data on semantics that show the benefits or some results I would live to read it, thanks for the link I’ll check it out as well.

    As you say two different sides of the coin, I know what works for me but respect your opinion.

    Thanks for the resources and happy new year!

Comments are closed.