Previously, we had been using Free Keywords which is a free piece of software that you can download to your desktop. This tool takes its data from the Overture database.

Both keyword selection tools are different in terms of the where and over what time period the data is collected from. I carried out a few test searches and compared data from both tools to see what kind of difference there is between the number of searches reported in Keyword Discovery and Overture for the same search phrases.

Test One - Niche:

Due to the fact that the data from Overture is for the last 30 days and the data from Keyword Discovery is from the last 365 days, the OV figures have been multiplied by 12 to allow a fairer comparison.

Test Two: Competitive:

As you can see, there is a huge difference in the results returned between the two tools. How can this be explained and which tool is more accurate?

Three of the main differences in the data used for these two keyword research tools are as follows:

1).        Time period. Although I have tried to take this into account by rationalising the data, a simple multiplication by 12 does not consider other factors such as seasonal changes. Exam papers for instance are likely to popular searches during mock exam period but not over the summer.

2).        Both tools have different “data cleansing” techniques. For instance, Overture de-pluralises, de punctuates, etc.

3).        Both KD and OV gather their data from very different sources – OV is based on searches made across the Yahoo network, whereas Keyword Discovery is made up of searches carried out on over 200 search engines worldwide.
Neither of these is right or wrong, but both have inherent flaws to their model. Overture for example, is prone to inflated results due to skewing by automated rank checkers. Overture would also count, for example, the following terms as the same term:

Online website marketing
Website marketing online
Marketing website online

The best solution to nailing a reasonably accurate and relevant keyword list to work from is to review a range of keyword research sources and then devise a set of algorithms to apply when combining the data.

Another option might be to run a PPC (pay per click) campaign for a week or two to gather data about the number of impressions you get for keywords you are testing.

First, lets just get a few points clear about SEO and where it differs from Pay Per Click (PPC) advertising.

1).  SEO is about making your website rank highly for keywords in the “natural” (i.e. free) listings on the search engines.
2). There is no charge for natural (sometimes called organic) SEO listings.
3). Natural search engine optimisation is a long term initiative, results do not appear overnight.
4). Search engines do not tell anyone how they rank websites (i.e. what their algorithm is), so SEO involves a certain amount of experimentation to find out what does and does not work.
5). Time and time again, statistics prove that people are more likely to click on natural listings than the PPC adverts
 

So what exactly does all this mean to my business and the bottom line?

1). Want more customers, more sales and more leads? The internet is the biggest growing market place full of potential buyers waiting to hook up with your products and services. People are no longer scared of buying online.
2). It is pretty much guaranteed that some, if not most, of your competitors have invested in SEO, don’t let them get ahead!
3). Because organic clicks are essentially free, in the long run your CPC (Cost Per Click) will be greatly reduced when compared with PPC spend
4). Paid search is constantly increasing in price as more and more people bid to rank for certain keywords. This trend is unlikely to change, so now is the time to bite the bullet and get started on your natural SEO.
5). Paid search is getting more like search engine optimisation, therefore you need good SEO in order to perform well at PPC. Check out my earlier blog on how PPC is like SEO for more information.
6). You don’t need to be number one - people get good traffic in the top five and even the top ten spots
7). Good SEO will help improve your website in terms of accessibility and usability, so not only will you get more customers, you will also have a better website!

1.       Great: Your keyword is very relevant and your Quality Score needs no improvement.
2.       OK: Your keyword is relevant, but you can still benefit from a higher Quality Score.
3.       Poor: This keyword is not very relevant to users and we suggest that you replace it with a more specific keyword. If you would like to keep advertising with this keyword, we recommend optimising by using more targeted ad text and improving your landing page content.

For me, this transparency improves my experience of Adwords. In the past, the quality score has been an unknown factor which could, potentially, have quite an impact on the amount you are bidding on certain keywords and your subsequent advert positions. This is certainly an improvement.
 

A few years ago, I would not have seen this as any kind of issue, but now paid search has changed. Gone are the days where you would bid for a keyword get your adverts listed and start seeing traffic. Now the top paid search engines (Yahoo, Google and MSN) all use something called “quality scoring”. Quality scoring can basically be defined as a focus on advert and landing page relevancy and quality, bid price and performance. 

What this means is that you need to make sure that you write adverts that include the keywords you are targeting and that are set-up to land on pages that are also focussed around those same keywords. Keep keywords, adverts and landing pages relevant and consistent. It is because PPC customers now need to think about optimising their landing pages that paid search is moving closer to SEO. 

I actually think that this is a very positive move. I for one was very mistrusting of paid advertising - clicking on a listing would only be followed by disappointment that I had been tricked into clicking on a website totally unrelated to my search query. Needless to say, people must have had very high bounce rates. So this improved quality system must be better for users as surely they must be getting more relevant paid results. It must be better for advertisers as although it might take a bit longer to set-up, the final outcome should be reduced bounce rate and, ultimately, a higher conversion rate. It must be great for the search engines as happy customers spend more!     

Both the Google Mini and the IBM-Yahoo search engines are aimed at the entry level market and can be utilised for internal and external search across intranets, company databases, public websites and network files. To interact with OmniFind, users have a special browser based search page or the search box can be integrated within an existing webpage. The search page can also be customised by changing fonts, colours, logos and text to match your own branding.

I downloaded a copy of IBM OmniFind Yahoo! Edition to find out if it’s as good as they say it is! Initially, I have just installed it locally on my Windows PC (there is also a Linux version). It installed very quickly and easily with only the usual changes required to my firewall software. Now I have a smart looking blue OmniFind Administration Console to play with!

The graphical interface is very user friendly, with options down the left hand side that allow you to crawl websites, crawl files, manage the search experience and the system.

I set the crawlers off on Commercetuned and, even though it is not a massive website, they found all of the pages within a few seconds. Once you have crawled and indexed your documents, you can begin searching using the search page. This looks like Yahoo’s usual search box, except very simplified without all the other “portal” information that I think clutters their home page. You can personalise this page anyway, so if you don’t like it or want to put your own logo on - no problem.

The results for a few simple searches, like “Search Engine Optimisation” and “Social Bookmarking” display the pages we specifically created to be themed around those keywords in position number one - so it seems quite relevant. As it is, there is a utility that allows you to manage the search experience by tweaking the following:

1. Define synonyms - here you can link different words with the same meaning. This helps if you discover that users are not finding the most relevant documents as those documents use different words to those words used in their queries.
2. Define featured links - this is very similar to having sponsored listings at the top of your search results. By defining featured links you can help users find a specific page or document more quickly by adding a featured link that sits at the top of the search results and looks different to the other results.
3. Tune Results by adjusting ranking (I wish I could do this in Google!) - You can do a number of useful tweaks in this area, namely adjusting the results by date modified, URL path (i.e. pages closer to the root are more relevant) and also by analysing weblinks which seems to work quite like Page Rank.

OmniFind also offers data and statistics on all queries carried out; this includes keywords, number of results returned, response time and so on. This will be useful when you are tuning search results to find out what language your customer’s are using and also if they are seeing the results you want them to.

I am going to play around with OmniFind over the course of this week and see what else it can do, but having given it this brief review, I can see that IBM-Yahoo have created a really useful tool. It is free, easy to install and use and can be customised. Try it out and see how it can help you!

Unfortunately, I missed the opportunity to take the Beta version of Deepfish for a swim as there was only limited availability, if I am called up in the seond round, check back here for a full report on how useful it really is! Check out the MSN Live Labs website for further information.

In a move that stunned the media industry, Google recently expanded into online video by purchasing YouTube - thus offering a free video film service online. Old media is now striking back, with the filing of a $1 billion lawsuit by Viacom (owner of brands such as Paramount Pictures and MTV) against Google/YouTube for copywrite infringement. This could be very costly for Google if lost.

There is yet more trouble on the horizon as Rupert Murdoch’s News Corp are joining with Google’s Internet rivals Microsoft and MySpace (owned by Murdoch) to offer entertainment and videos to their massive audience. This retaliation from competitor’s could stunt Google’s growth. In holding between a 40 and 70 per cent share of online advertising revenue, Google has become incredibly successful in a very short space of time, but by diversifying into other areas outside of search, such as mobile, new and video, Google has worried investors which has been reflected in it’s falling share costs since the beginning of this year.

This wrangling of power between old and new media looks set to continue as the entertainment world is redefined in the digital age. Google needs to take the bull by the horns and risk loosing popularity (as Microsoft did in the past) in order to cut itself a bigger slice of the new media market.

1). Identify test page

I decided to keep my first multivariate test as simple as possible, by making just a few small changes to this page:

social bookmarking

The first step is to decide which areas of the page will be changed. I am splitting the page into three sections – headline, image and text. I then tell Google Website Optimizer the page URLs of the test page and also one conversion page URL. In this case, I am going to use the contact us page.

2). Tagging the test page

The next step involves adding tags to the specific sections of the test page for which I am going to test the content variations. In English, this means I insert special tags around the heading, image and content that I have changed.

Although this site already uses Google analytics to track visitors and conversions, I also need to add some extra tagging to the conversion (contact us) page so that I will be able to tell whether people came from the test page or other areas of the site. To be honest this is the hardest part so far and a bit of hassle.

Also, adding the Google script to this page had made it non XHTML 1.0 compliant, which the rest of the site is. This is something that will need to be sorted out by the Commerce Tuned CSS specialist. I would have thought that since Google seems to love W3C compliance and well formed code; it would give it users something that was also compliant?

Once I have done this, Google Optimizer goes off to check everything is set-up correctly. After a few more tweaks, I manage to get everything set up to its satisfaction. I now have pages very heavy with script that are non compliant, but Google Website Optimizer seems to like them so that must be OK!

3). Create Variations

For each section that I have tagged (headline, image and content), I now create a variation. This is done within the Google software and is very straight forward. I literally click the variation I wish to change and then a little box pops up into which I type the name of my variation (for example heading 2). I then use a CMS (Content Management System) style interface that allows to me to add text and HTML in order to create the actual variation. It is even possible to preview the changes on the live test page to make sure everything is OK before saving them.

4). Review test settings and launch

After checking a summary of settings and variations, the test is launched. I have set-up some Pay Per Click (PPC) adverts that will land on the test page, so that should be a great way to direct traffic to the page and to monitor what happens.

After I have gathered some data, I will review the Google Optimizer reports and see how the experiment went. It will be interesting to find out if this tool is actually likely to help me as an SEO person, help my client towards a higher conversion rate.

For those of you who aren’t familiar with the concept of A/B split (or multivariate) testing it basically involves the scientific testing of various parts of a web page such as images, headings, layout, buttons, navigation to work out which versions your users prefer, or better still, which one is most likely to result in a conversion. A/B testing is the simple test of changing just one element of a page against another to see which works best, whereas multivariate is the equivalent of thousands of A/B split tests. It’s like post click marketing, helping you make the most ($$$) of the traffic that PPC and Search has driven to your site.

There have always been plenty of pieces of software out there to take care of this for you, but none that I am aware of that fit in nicely with Google Adwords and that are free! Welcome Google Website Optimizer!

So lets get started. I am running a PPC campaign on www.commercetuned.co.uk, so I will use this as my test subject.

The first thing that I need to do is decide which landing page I am going to use for the test. I have been directing traffic from the PPC campaign to an article on social bookmaking that I wrote, so I will use that page. The aim of the page is to educate users and to induce them to request more information about this topic. So how do I work out how many parts of this page to change? The Optimizer tool tells me that I can test up to 1000 variations in one go, but the more variations I have the longer it will take to get meaningful results. In fact the length of an experiment is determined by 5 key factors:

1). Number of combinations
2). Conversation rate
3). Website traffic
4). Estimated conversion rate lift
5). Percentage of participation
These points should all be fairly self explanatory - i.e. the higher the number of combinations the longer the test duration; the lower the conversion rate the longer the duration etc.

I am now going away to work on the two versions of my page, so will let you know how I get on!

Diana Gray - Search engine consultant, Commercetuned.