A link from Als weekly round up on Self Made Minds to an article on improving organic click through rates got me thinking. Sadly the article fails to indicate how they quantitatively measured their results which means the claim of 3x increase of click through is to be taken with a pinch of salt, make that a healthy dose of salt. While getting people to actually click links is normally outside of an SEOs remit it still can be interesting to attempt.
Given my penchant for experimentation, for those who are mildly curious my attempts at SERP artwork saw a drop in clicks, you will not be surprised I plan on taking this little experiment on.
The theory
People are attracted to things that stand out, so if you capitalize every letter in your title people are more likely to click it. I have taken the idea further with my second experiment and taken some advice from PPC will include a direct call to action i.e click here. The theory being people actually click when they are told to!
How to test?
Well I have had to coax a pair of clients (thanks guys!) into helping, both clients have the advantage of keywords with high volume of visits per week and 2 sites within the first page. This means we have two normals as well as two test site. We can use the normal to see if the amount of traffic is fluctuating heavily (a mention on the news for example). As we already have 4 weeks data for all 4 sites, 2 norms and 2 experiments we can see that the sites are fluctuating a little but not much, so on Friday I started of two experiments, in 2 weeks time I will have some basic findings but it won't be for 5 weeks before we can actually give real results.
Experiment 1
The first subject will have it's entire title Capitalized, no words changed or added in any way.
Experiment 2
Additional first word "CLICK HERE for" will be added to the title, note the upper case for click here. This could potentially influence the pages ranking and cause it to lose position given the title is keyword heavy.
So with the mixture ready it's just a case of bake (I can tell you both have been indexed and are showing their new titles on regional results)
But it looks spammy
Al made the point that it looks totally spammy to have capitalized titles and I agree more importantly so did the clients so I have whipped up a piece of javascript to help with the problem.
//Grab info from <title>
var old = document.title;
//Split lower on whitespace RegEx
parts = old.split(/\s+/)
// New Empty Variable
newtitle = "";
// create loop fo each word in array
for (i = 0; i < parts.length; i++) {
// Uppercase the first letter, and ensure the rest is lowercase add single white space.
newtitle += parts[i].substr(0, 1).toUpperCase() + parts[i].substr(1).toLowerCase() + " ";
};
//send back to <title>
document.title = newtitle;
});
This simply splits the title into individual words and capitalizes the first letter of the string and lower case the rest.
For the second experiment I simply did not show the first 3 words by changing the starting point of the array.
for (i = 3; i < parts.length; i++) {
Problem solved (yes there is a slight blip on start up as you can see the title swapping)
So come back in a few weeks and I will tell you the results...






I hate to be slow - but was there any reason why you did not visually fix this with CSS? text-transform:capitalize would make the first letter only capital for the user and this would work as soon as the CSS took effect. If you add it inline to the H element then it will be transformed as it loads.
Or did I totally miss something?
Except we are changing the title tag and as far as I know no browser supports css within a title element
sounds like a great experimnet, and as always I look forward to the results. I am thinking that the CLICK HERE for might really effect people, although i don’t know if the impact on rankings would be worth it.
And kudos to your brave clients.
I am just writing up the first set of results, they might surprise you!
The first set of results are live - http://www.timnash.co.uk/05/2008/clickthrough-experiment-update/
Tim, your name popped up in a conversation with Jeffrey Smith, who I’m due to be meeting up with in Birmingham next month.
So here’s me saying hi!
With the background out of the way, looking at your idea, it strikes me as being a mixture of small-scale linkbait combined with SEO and usability, too.
I’m toying with the idea of doing something similar, just to see what happens, for the fun of it, if nothing else.
So if there’s anything you’d like to see from my little experiment, let me know.
Speak soon…
Hi Wayne
we have crossed paths a few times before but cool on the Birmingham meet.
I’m not really sure how the experiment could be described as linkbait? the purpose is to attract clicks not links.
As you can see from the update even that is not really working though the test subjects are willing to carry on so maybe it was just a bad first couple of weeks.
Always interested in experiments of this sort and any data so let me know when your done and if you need a hand.
Hi Tim, I was going to describe it as clickbait, but I thought it was a contrivance too far! At least you figured out what I meant in the end.
I’ll probably spin the article into a headline writing thing, which I’ve only touched on once or twice.
BTW, I live just above you in South Yorkshire, so we’re actually pretty close…
Responses to this post: