Curated by: Luigi Canali De Rossi

Monday, February 26, 2007

Blast Your Web Site Conversion Rates To The Max With The Google Website Optimizer

Sponsored Links

Google Website Optimizer is a new, free tool designed to provide online marketing managers and web advertising specialists with all the tools needed to carry out effective conversion testing on their site.


This means that any webmaster or blogger who has a strong interest on how the pages on his site perform in terms of conversion, can now create a number of alternative layouts for any web page and have Google help him test them against real visitors.

Yes, you have read the above correctly. Google itself allows webmasters to show their online audiences multiple versions of the same web page, to measure precisely which solution in terms of ad positioning, copy, color or one among hundreds of other variables makes the most significant difference when it comes to convert web site visitors into customers.

"Let me help you, so that you help my business grow", is the mantra that Google seems to whisper alongside this new launch.

Whether you want them to find their way to your affiliate-marketing online store, the form for subscription to your email newsletter, or simply your RSS feed, here is a service that is going to be of a lot of interest to you, not least because it is totally free.

If you are trying to make a living from your blog or website, optimization of your website content and design are key factors in making sure that your visitors have a better chance of finding the materials that you want them to. When visitors touch down on your 'landing page', they generally make a lot of decisions in the first few seconds as to what they're going to do next, or more importantly where they're going to click. A well optimized site will help these visitors follow some kind of pre-determined path no matter whether this leads them to subscribe to a RSS feed, click through to your online store or make it easy for them to add your new ebook to their shopping basket.

In all these occasions, Google Website Optimizer can make a true, very significant difference.

Here at Master New Media testing and optimizing conversions is a daily work in which two or more people devote daily time. There is no week that goes by without some split testing and all of the good results we have been to achieve in the last year are mostly due to the outstanding job that our conversion optimizer director has done in planning, coordinating and evaluating an ongoing number of tests.

If it had not been for this kind of work, for which we haven't had at our disposal Google Website Optimizer, Robin would be certainly making a lot less than its present $20,000+ month advertising number, which, besides, keeps growing at a steady rate.

So whether you are paying out for advertising to appear on other sites with the goal to receive extra traffic to your site, or a online advertising director wanting to further expand the impact and scope of your conversion testing Google Website Optimizer is going to be the key strategic instrument, to uncover and identify which variables and specific solutions influence the most the conversion rates on your web pages.

Nobody visiting your online store? Use Website Optimizer to find out why. Hundreds of people signing up for your newsletter, but nobody interested in your RSS feed? Again, Website Optimizer might bring you some fresh insights.

In this review, I analyze for you the new Google Website Optimizer key features, and how you can best put them to use in your own online business.

Google Website Optimizer - overview

Google Website Optimizer is currently open to beta-testers who have an existing Adwords account. While Website Optimizer primary target audience is made up of AdWords advertisers looking to increase their conversion rate, the service has plenty of users for ANY web site owner whose goal is one of understanding and optimizing page components to obtain pre-determined visitors actions, like clicking on a certain ad or proceeding to view a specific section of a web page.

In a nutshell Google Website Optimizer facilitates the testing and analysis that alternate content combinations can generate by tracking the user responses to these very changes.

The user of a Google Website Optimizer sets a default page of images, text and headings, and then tries out different variations of this same content - changing the images used or their position, creating variants for headlines, even rewriting whole chunks of text.

According to Google, the key content components that can be manipulated effectively to influence visitors attention are those seen in the first few seconds of landing on any web page:

  • Headlines
  • Prominent images or graphics
  • Simple promotional text
  • Calls to action


It is on the use of these elements that the tests one would run with Google Website Optimizer need to focus. As explained, multiple variants of some of the above key content items are tested in a number of different combinations over the course of a period of time to determine which one has the highest conversion rates.

Google Website Optimizer tracks users re-actions to these variants, and compares them against the default version. After that it reports in detail to you which variants are more successful in driving your users to a URL that you have selected. This could be the URL of your online store, your PayPal payments page, your RSS feed, or wherever it is you are hoping for your visitors to click.

Through experimenting with key content combinations you can see over the course of time just which "setup" is going to work best to achieve the goal that you have set for a particular page of your website. Whether this is to sell the item featured on the page, as advertised elsewhere, or to ensure yourself future advertising revenue through user subscriptions, it remains up to you.

This is not a tool for everyone. Google Website Optimizer is squarely aimed at those looking to seriously optimize their web site content and needing to improve the ratio of visitors that buy, read or click selected content.

While the process is relatively straightforward, it did require until today very extensive amounts of time to design, plan and coordinate the execution of such optimization and conversion tests.

The process involves also copying and pasting HTML code that the Google Website Optimizer generates into the HTML of your web-pages so that "variants" can be played in an automated and systematic mode using Google own engine and infrastructure.

Certainly some less technically adept website owners may prefer to ship this task out to their webmasters, or online advertising specialists. In all cases the Google Website Optimizer is best implemented in conjunction with somebody having some knowledge about page design and layout, usability, accessibility and SEO-SEM principles.

Creating your conversion testing experiment

Google Website Optimizer is now integrated into your existing AdWords dashboard, and it is from here that experiments are created and added to a test list. This list can be filtered by those experiments being edited, currently in progress, or completed, so should you have a number of different experiments on the boil, locating them become a very simple prospect.


When you choose to create a new conversion experiment, a wizard guides you through the process step-by-step. This process involves:

  • Identifying the pages to be tested
  • Adding tags - inserting the appropriate tags into the HTML of your pages, so that Google Website Optimizer knows which parts of your page are being tested with multiple alternatives, and need to be tracked with precision
  • Creating variants - adding and specifying the different combinations of content you would like to use in your experiment
  • Reviewing settings - a last chance to review your experiment settings before launching the experiment on your website

Let's take a look at this conversion testing process in a bit more detail.

First of all, you are going to identify two key URLs - the URL of the page that you want to test, and the URL of the destination you want to send your visitors to.

This second URL is of key importance in tracking your conversion rate, or monitoring how successful your layout is in helping readers achieve your pre-determined goal.

Next up comes the technical part, the adding of tags to your experiment pages. Tags are added first to both the test page and destination page, and then added to each individual layout element that will be varied in the experiment. It is here that some users may wish to delegate the code copying and pasting to their webmaster if they are not so proficient in wrangling with HTML themselves.


One, at first non evident limitation, is that having to add code to the destination page ultimately precludes Google Website Optimizer from being able to help you to track your outgoing advertising or affiliate links, unless you have the cooperation of your site sponsors. As this is not what the tool was designed for, it is difficult to fault Google for this limitation - Google Website Optimizer has been set up to assist you in tracking site visitor actions from one part of your site to another, and this is indeed what it does well.

Once your tags have been validated from the Website Optimizer control panel, you can now go about creating your key content variations.

Using your default setting as a basis, you can now set about creating as many variations as you see fit, bearing in mind that the more variations you create, the longer your experiment is likely to take.


Variations are created directly in the HTML code itself, so that you can dip in and change text content, size and colour, image source and any other variables you might come up with. These previews can then be previewed from the control panel.

Finally, when your variations are ready you can quickly check that your URLs are correct, decide what percentage of your site visitors will be exposed to your experiment, check which variations are going to take place, and then activate the whole works and put it out live on your site.

After that, it is really just a waiting game, and the time your experiment takes will depend on your site traffic and the amount of variations you have created.

The total of your individual variations will be rolled together into different combinations, so that, for instance, if you had three headline variations and two image variations, you will have a total of six combinations that will each be tested on your users.

As you can imagine, then, adding too many variants can quickly snowball the amount of combinations that will need to be tested, thus further protracting the time necessary to complete the experiment.

Analyzing your results

When your test is completed you are presented with a lot of useful report data to work with.

Reports come in two varieties:

  1. Page section reports - these reports give you exact details of the individual items you have been playing with in your tests, so that you can see, for example, how well one image performed versus the alternatives you had selected.
  2. Combinations report - While the page section report will help you to see which individual items work better than your default set up, the combinations report is potentially more conclusive, given that it can show you exactly which combination of page section elements proved most effective in terms of conversions.

The main source of interest in "page sections reports" is what Google calls a 'Relevance rating', a score from one to five that rates a particular piece of content in terms of its overall impact on your combinations. A low rating suggests that the given item has little or no impact on the success of a particular layout combination, while a high rating indicates the important role a given content item plays in the combinations it features within.


If you have one image with a rating of five, and another with a rating of one, you can be certain that those combinations featuring the five-star image will be creating much better conversion rates than those that feature the one-star image.

Colour coding helps you to quickly analyze your data, so that well performing elements will show a majority green bar in the 'Estimated conversion rate range', while poor elements will show a majority red bar. Likewise, the percentile figure of the observed improvement column is highlighted green for improvements on your default layout, and red for those elements that fared worse than the original. This means that even in scanning your results, you can pick out key performers and weed out the deadwood at-a-glance.

Other key data available includes the exact conversions and page impressions ratio, which will give you a clear idea of how well a given design element fares against its competitors.

The combinations report though is really where you find "it's at" in terms of seeing which combination works best overall. While you would think that the most successful combinations would be those comprised of the most successful individual elements, this isn't always the case, so it is worth waiting on this more conclusive experimental data for the final word on what really works.


If there is clear winner in terms of conversions, it will become instantly evident in this report, and once again, all your variations tests are compared with the original layout by means of color coded green or red bars. The greater the green bar, the greater its advantage over the default setup, the greater the red bar, the poorer its performance in relation to the default original.

Added to this visual data are probability percentiles gauging the given combinations chance to beat the original, its chance to beat all other combinations, the percentage of observed improvement over the original, and finally the conversions and impressions ratio.

Using this combination of data, it becomes much easier, less time consuming and more precise the precise identification of the highest performing page layout combinations, titles, headlines or selection of images that will guarantee the highest conversion rates possible.


For those serious about increasing their ad conversions, or finding the most effective key content combinations for their websites, Google Website Optimizer could prove an invaluable tool.

For sites that like the mini-network Robin Good has created, this is a true godsend, as there is nothing more valuable to a commercial web site to know how to continuously improve and refine the way its content pages are composed and setup.

While Google Website Optimizer will in no way do the design work for you, it will provide a straightforward way of testing out content and page layout variations and the effect that they will have on your visitors' actions.

As this is not a tool for the layman, not so much in terms of technical difficulties it presents, but rather more because the sciences it integrates range from statistical analysis to SEO-SEM, usability and accessibility principles, it is highly recommended that it is placed in the hands of individuals having some background and understanding of these areas to be used to its highest potential.

Also, keep in mind that the process of setting up experiments involves a fair amount of adding code to the HTML of your web-pages, and while this can all be copied and pasted into the appropriate places (with the assistance of Google excellent tutorial presentation) some will feel more comfortable assigning the task to a knowledgeable webmaster.

Nonetheless this may be superficially considered a new enemy by web page designers of all sorts, Google Website Optimizer does take the guesswork out of the design process and puts it to the service of achieving specific, quantifiable results. Whether in helping your visitors to decide to make an impulse purchase, or making it easy for them to subscribe to your site feed, this is valuable information indeed.

Google Website Optimizer is a breakthrough analysis and optimization tool for the serious online commercial publisher wanting to seriously increase web-based conversion rates.

There is nothing I know of that does anything close to this, and positively none that does it at the same price (zero).

If you run a blog, news or commercial web site that earns money via advertisings, sponsorships or through the sale of particular products and services, this is a tool you can't do without to find out what and where needs to be changed to increase your revenues by utilizing a highly scientific, calculated approach.

Google Website Optimizer is in Beta testing at the time of writing, and is currently accepting sign-ups from people like you.

If you are serious about making money online, and you have the resources to use it, you MUST go give it a try.

Additional resources

If you would like to learn more about Google Website Optimizer, you might want to check out the following links:

Originally written by Michael Pick for Master New Media as:
Optimize Your Key Website Content To Maximize Your Profits: Google Website Optimizer Shows You How

Readers' Comments    
blog comments powered by Disqus
posted by Michael Pick on Monday, February 26 2007, updated on Tuesday, May 5 2015

Search this site for more with 








    Curated by

    New media explorer
    Communication designer


    POP Newsletter

    Robin Good's Newsletter for Professional Online Publishers  



    Real Time Web Analytics