There are millions of articles, ebooks and training courses out there that claim to tell you how to do SEO. As web people, we of course read many of these in an effort to keep on top of the latest techniques and to understand the most recent changes to search engine algorithms. But as naturally curious people, we couldn’t help feeling that we wanted to test these theories out ourselves a bit more scientifically, and see the outcomes with our own eyes.
So we pitted three of our best web experts against each other in a challenge to test their SEO skills and knowledge. The challenge was…
Who can get their website to rank highest in Google for a specific keyword.
But there was a catch. They weren’t allowed to create any links to the websites they created, so the only type of SEO they could use was on-page.
We decided that we needed keywords that didn’t give any competition from other websites, so that we could easily see who was winning. We chose an obscure phrase that still contained real words to make the experiment as realistic as possible. We chose the phrase Frictionless Owls, which (unsurprisingly!) had no competition when we began the challenge. The plan was to check the results after 30 days.
Starting out
So now we want to take you back to day one, and let’s first meet our challengers:
Ali
Experienced front-end developer
Website: www.frictionlessowls.co.uk
“I don’t want to use the ‘black hat’ approach of spamming my pages with vast quantities of keywords which can make websites unusable, nor do I want to use any complicated SEO techniques which may be used by SEO ‘experts’ in the industry. Instead I’m taking an approach that could easily be used by an ‘average’ person wanting to get a website online and listed highly in Google.”
Oliver
Experienced PHP developer
Website: frictionless-owls.com
“I’m going to generate huge volumes of semi-quality spam. A machine can produce more content than a human ever could, with less effort.”
David
Experienced front-end developer
Website: frictionlessowls.com
“I’m going to use as many SEO tricks that I’ve learned from the industry, without crossing the border into dodgy ‘black hat’ territory. I’m going to strategically place relevant keywords on each page.”
The rules
To clarify none of the competitors are allowed to purposefully create a link to their website, because this isn’t a challenge about linkbuilding or popularity, it’s a challenge purely about website quality and how this affects SEO.
How we’ll check the results
As we’re hosting these websites in the UK we’ve decided to check the results using Google.co.uk rather than Google.com. To avoid Google’s personalised search we’re checking the results on one of our impartial servers, rather than on one of our computers which would show varied results depending on the search terms you’ve used before.
And they’re off!
Our three experts get started on building their websites. With only a month to go there’s not much time to spare!
Day 1 – Ali chooses .co.uk
“Day one of the project and the first job is to decide on domain names. We decide to draw straws to see who gets to pick first, second and third and then all domains will be registered through the same account. I draw second and after Dave chooses www.frictionlessowls.com I’m trying to decide between going for the .co.uk address or whether a hyphenated version of the .com would be better. After some thought I decide to go for www.frictionlessowls.co.uk, mainly because we’d decided to use www.google.co.uk as the search engine to measure the results and I think having a domain name associated with the same country might work in my favour.”
Day 1 – David opts for .com and focuses on speed
“I register my domain first and go immediately for frictionlessowls.com, which includes both keywords in the domain. I chose the .com domain, hoping that the most popular extension might add some weight to the website’s rankings, even though we’re going to be testing with www.google.co.uk.
“I upload a basic holding page with a very basic XHTML structure. I want to make sure the code is lightweight so that the file size isn’t too large, because I know that download speed is factored into Google’s ranking algorithm now, so I think it’s important that my website loads quickly.
“I search online for the keywords ‘frictionless’ and ‘owls’ and in both cases the top result is a wikipedia page. Presumably the content on these pages contains words relevant to the keywords, so I take these and paste them into my new home page. I go through each paragraph and make the keywords bold using the <strong> tag as I know Google looks at these on the page and gives them more precedence when ranking the page.
“I put both keywords in the <title> tag of the page, because I know that Google considers this wording important when ranking the page.
I decide that having the www in the URL might water down the keywords, so remove it. The site now responds straight from the http:// address.”
Day 1 – Ali creates the perfect holding page
“Now the domains are registered it’s time to get something online. My approach will be to try and implement the generally recognised ‘rules’ of SEO such as using the keyword phrase often and emphasising it in headers and text.
“I create a simple holding page using a few standard techniques which are generally believed to help SEO. For the content of the holding page I add a simple ‘Frictionless Owls Holding Page’ heading 1 (<h1>), a heading 2 (<h2>) of ‘About Frictionless Owls’ and two simple lines of text ensuring that the phrase ‘frictionless owls’ is highlighted in bold (<strong>) and italics (<em>). I also make the phrase link to a second page which is identical but has been saved as ‘frictionless-owls.html’.
“Behind the scenes of the holding page, I put in place a few other techniques. In the HTML I make sure that the title tag includes the phrase – I use it twice for good measure! (<title>Frictionless Owls – For all you Frictionless Owls fans</title>). I then include some very simple meta keywords and a description (<meta name=”keywords” content=”frictionless owls, frictionless, owls, owl” /> <meta name=”description” content=”A site all about frictionless owls” />). My final attempt at getting the phrase into the source code is to name my CSS files frictionless-owls.css and frictionless-owls-print.css.
“Now I have a holding page online at my chosen domain name I need to get Google to find it. To do this I make use of the Google Webmaster Tools. After verifying my site to asssure Google that I am the owner (by adding the google-site-verification meta tag) I submit a sitemap.xml file to show Google the sitemap of my site, listing both my index.html homepage and my linked frictionless-owls.html inner page. I verify both http://www.frictionlessowls.co.uk and http://frictionlessowls.co.uk to ensure that both options would be seen by Google.
“Now it’s just a case of wait and see!”
Day 1 – Oliver gets started slowly
“I create a plain holding page (a couple of sentences) and submit to Google via www.google.com/addurl”
So at the end of day 1, all three of our challengers have some content up on the web (even if it’s basic!), but whose will be the first website to appear on Google? So far, David is the only one who hasn’t submitted his site to Google.
Day 2 – David gets blogging
“Google loves fresh content, and I want to show that the site is being updated regularly. This means Google will send a bot to spider the site more often, and means that the changes I make will show up more quickly in search results. This is essential because I only have a month, and don’t want Google to only spider my site once. So I install a WordPress blog onto the subpage /owl-blog/ (once again getting one of the keywords into the URL).
“I add content to the blog, scheduling articles to go live once every hour. This should mean that Google notices the site being updated regularly,and checks back often to spider more pages. I paste the content straight from the wikipedia pages about owls, and from the page about frictionless planes. This content is not only littered with the keywords that I want to be ranked for, but also with related words that Google might realise are similar.
“The URL of each blog article will contain the category, which will be the keyword ‘owls’ or ‘frictionless’, just to put the keyword into the URL again. The URL will then contain the date the article was posted, followed by the slug (a shortened title). The slug will always contain the keywords I want to optimise for, so the addition of the categories into the URL might be redundant and might even be too much. I’m walking a dangerous line already between including too many keywords on the page, or just the right amount!”
Day 2 – Ali is number 1. But not for the page she expected.
“At around 5pm on day 2 a search for ‘frictionless owls’ on google.co.uk results in my site being the number 1 result – not bad for just 24 hours and a simple holding page! So far Dave and Oliver’s sites are nowhere to be seen!
“Interestingly it seems that the inner page with the frictionless-owls.html extension is the one that’s been picked up rather than the home page.”
So Google is only ranking Ali’s second page – the one with the keywords repeated in the URL. It looks like Google needs to be told twice how to rank this site!
Day 2 – David finally submits it to Google
“I want my site to load as quickly as possible, so minify) ) the CSS to remove unnecessary linebreak characters and to squeeze it down into the smallest filesize possible.
“As I feel I’ve got enough content to the homepage now, I add the website to Google via its ‘add page’ URL http://www.google.co.uk/addurl/. I’m not sure how long this process takes, but wanted to make sure there was enough content before I submitted it. Normally I don’t think this is necessary as Google will find your site via the incoming links, but in this experiment, I’m not allowed to create any incoming links!
“I also register the site with Google’s webmaster tools, which involves me adding a small meta tag into the home page to prove I’m the owner of the site. This lets me find out exactly when Google has crawled my site and if it finds any errors. Also I can provide Google with a link to a sitemap that better explains the content on my website. I create a sitemap on the website www.xml-sitemaps.com which builds one automatically for me, and upload this to my website so it appears at /sitemaps.xml.
“Before uploading I go through and remove any pages from the XML sitemap that it was listing unnecessarily. It had found my wordpress blog login page, so I remove this as it’s not a page I want to be included in search results. The sitemap allows you to specify how often each page is updated. I set daily as the option for most pages, apart from the home page which I set to hourly because I might frequently make changes to the site and want Google to know this so that it can come back and check often. I want to be careful that I don’t make false claims here otherwise Google might realise and ignore the information I’ve given them on the sitemap.”
Day 3 – Ali is still the queen of Google
“A quick check early on day 3 shows that my site is still at the top with the others still not in sight, and it’s now the main domain showing in the listings rather than the inner page.”
Day 3 – David cheats when it comes to content
“I publish some more content on the blog. I’m not exactly proud of where I’ve been getting the content, because I’ve been copying and pasting sections from websites about owls! I know this is extremely dishonest, but I want to see how much tolerance Google has before they mark my website down for having duplicate content. It’s likely that my site will not even be listed, but on the off-chance it will, I should be rewarded for having many pages of content.
“I notice this morning that Ali’s website appears in Google already. But I’m not too worried as she got a head start on me of almost a day, because I wanted to provide some content before Google spidered the website.”
Day 4 – Oliver sets up his spam machine!
“I’ve code a basic sentence generator using CFGs (Context Free Grammars). The basic generator can produce virtually infinite sentences which – for the most part – make grammatical sense, but are largely nonsensical. At this stage I’m limited to past tense singulars with no agreement. As I’m generating narratives I’ve thrown in some basic dialog forms as well, e.g. “Oliver exclaimed ‘<sentence>’”.
“I consider using Markov chains for generation but rule it out as it’s likely to be far too spammy for Google. I figure that CFGs will create content that will be definitely unique and possibly even genuinely entertaining for real people!
“At this stage the generator produces static files which are uploaded manually; I begin with 10 pages. The aim is to make this auto-generate virtually infinite content automatically in the coming weeks.”
So both Oliver and David are using dubious content to try and ‘trick’ Google. David is copying his from other websites – unethical because it’s stealing content – and Oliver is generating content that’s useless to people via a script. Currently Ali only has a holding page with very basic information, but what will she do to beat the others?
Day 4 – David gets his first taste of success
“Success! My website is now being listed in Google for the keywords frictionless owls. Strangely though, it’s chosen to list an unusual page, instead of listing the home page or even the home page of the blog. It’s listed the URL http://frictionlessowls.com/owls-blog/owl-category/frictionless/
This is a list of article summaries. It’s a page that doesn’t even appear in the sitemap I uploaded yesterday!
“I imagine the possible reasons for it listing this page are:
- This page has the right combination of keywords and content to appear in search listings.
- This is the only page that doesn’t look too spammy or contain duplicate content.
- This is the only page that Google has found so far. This is unlikely as I’ve already given Google a full sitemap so it should be aware of every page on the site so far.”
Day 5 – David usurps Ali’s throne
“I check Google for our search term and my site is now appearing above Ali’s! In actual fact I’ve got 3 of my pages listed: the home page, the blog home page, and the random page that was listed yesterday.
“This isn’t a victory yet though; all this proves is that Google has ranked the website. But I want to make sure that I keep this spot by providing regular content. I publish some more blog articles over the course of the day so that Google knows content is being added regularly.
“For the home page I want to do something a bit more special than just copying and pasting content, so I rewrite the content so that it’s totally fresh and original. This should be great for search engines.”
Finally David is starting to behave in a slightly more ethical way! Writing fresh original content is great for search engines, but will it make up for all those pages of copied and pasted content? Google hates duplicate content, especially in their latest Panda update.
Day 6 – Oliver submits a sitemap
“I upload some more newly generated pages.
“I submit a sitemap and register with Google Webmaster Tools.”
Oliver’s content doesn’t make much sense to a user, but Oliver is hoping that it looks like fresh and original content to Google. As well as picking random words, his content is filled with owl related words like ‘barn’ and ‘nests’.
Day 7 – David is worried about bots
“Today I amend some of the options in Google Webmaster tools. It’s telling me it’s not finding a robots.txt file, even though I don’t need to disallow any bots, so I add a near-empty file to make sure the Google spiders know they can access my site without any problem.
“I’ve check the search rankings and Ali’s ahead of me today. I noticed that she’s mentioned the keywords in the page title twice. I’m going to try the same to see if I can get ahead of her.
“I add a dynamic news feed to the home page which pulls through a summary paragraph for the latest 2 blog articles, so my site will always be showing brand new content on the home page every time I publish a new article.”
Day 8 – Ali launches a counter-attack
“Over the past few days my number one listing for the website has been overtaken by Dave’s www.frictionlessowls.com and I’ve been pushed down to 5th place, although this does seem to depend which browser I use when I conduct the search.
Day 7 search results in Chrome
Day 7 search results in Internet Exporer 8
“When searching using my default browser (Google Chrome) I’m still the number 1 result, even when not logged in to my Google Account with access to my Google web history. However if I search in a browser which I haven’t used before for this search I am in the number 5 spot. This just goes to show that your ranking may not always be as clear cut as it seems!
“Oliver’s site www.frictionless-owls.com is also starting to make an appearance at the number 6 spot although interestingly this is a link via a ‘whois’ third party website and his actual site is still not appearing even though it’s online and is featuring plenty of content.
“As my site is starting to slip and a week has gone by, I decide to make some changes to try and push my site back to the top.
“Rather than just the simple holding page which got me my initial success, I’m going to create a full content managed website to see if the increase in content and activity has a positive result on the rankings.
“My approach to this is again going to be one of the ‘average’ web user. I don’t want to fill the website with duplicated content about ‘friction’ or ‘owls’ just to get up the rankings. After all, if this was a real website and I was searching for a band called ‘Green Umbrella’ and the top search result was full of information about the colour green and the history of umbrellas I’d leave and find another website immediately. Instead I’m going to concentrate on the phrase ‘frictionless owls’ and see what effect this has.”
Day 8 – Oliver starts to get worried
“I add 5 additional pages and update my title tags to include ‘frictionless owls’. Google shows no sign of the site yet.”
It’s hard to tell if Google has ignored Oliver’s pages of auto generated content, dismissing them as spam, or if they just haven’t been archived yet. Or it’s possible that the keywords aren’t prominent enough on the page to get this site showing in our very specific search.
Day 8 – Ali has started her own rock band
“My simple CMS site is now online at www.frictionlessowls.co.uk. To give the site a purpose and to try and recreate what an ‘average’ user might try to find on the web, I decided Frictionless Owls was an internationally successful music group!
“I got a basic install of the Concrete5 CMS in place on the server and used a pre-built skin/theme for this so there was no particular SEO techniques built into the HTML or CSS, instead all of the SEO will be done ‘in page’.
“I firstly created a set of pages to form the main outline of the website and made sure each page had a meta title, keywords and description as well as a suitable web address.
“Once the pages were in place and linked to each other using consistent side navigation, it was time to add content. To create content for this site I took inspiration from sites about Take That, Alice Cooper and the Beatles! By using and amending the copy from these sites I felt that I created a realistic band website which had been changed enough to avoid Google finding duplicate content.
“Textual content was added to the home page, About page and Music page, along with a couple of images which used the phrase frictionless owls in the alternative text. For the time being I have placed a small amount of ‘holding’ text on the remaining pages saying that content is coming soon. This will allow me to regularly update the site and add fresh content as the next few weeks pass. I also added a link on the Photo Gallery page to a Google image search for frictionless owls just in case the reciprocal link improved the chance of Google finding the site.
“By having all these pages present on the website now I am able to create a sitemap of 6 pages which I resubmit to Google via Google Webmaster Tools in the hope that it will quickly recognise and spider my new content and improve my search ranking.”
So far Ali’s the only challenger to create sensible, well written content that isn’t either copied and pasted or automatically generated.
Her site will have a natural amount of content, including an image gallery. But with the site being a simple holding page for over a week, will Google come back and spider the site in time for it to be ranked highest?
Meanwhile Oliver’s site still isn’t having any luck in Google. Could it be that the content just looks too spammy? Will Google even bother spidering his site at all?
To be continued…
Find out the results in part 2
Pingback: Spam, plagiarism or human effort. Which worked best in our SEO challenge? Part 2
Pingback: links for 2011-06-06 | James A. Arconati
Pingback: How one tiny hyphen destroyed our SEO efforts