Kissmetrics Blog

A blog about analytics, marketing and testing

Built to optimize growth. Track, analyze and engage to get more customers.

A Beginner’s Guide to Google Website Optimizer

If you want to make real progress in increasing the conversion rates on your website, randomly trying different headlines, images, and other elements on your pages and crossing your fingers isn’t going to get you very far. You need something that can give you real data about what’s working and what’s not.

Even better is a free tool that can manage your conversion testing. Leave it to Google to provide just such a tool: Website Optimizer.

What is Google Website Optimizer?

Google Website Optimizer is a free tool for creating experiments for conversion rate testing. You set up experiments for various elements on your pages and then Google randomly serves the different versions of your pages to real visitors who come to your website.

There are huge advantages to this method of testing over other sites that test your pages with people who may or may not fall into your target market. With Google Website Optimizer, you know that the people who are participating in these tests are your target market (or at least your target visitor).

A/B Testing vs. Multivariate Testing

Website Optimizer allows for both A/B testing and multivariate testing. With an A/B test, you test two versions of the same page. This is a great option if you want to test two completely different designs or layouts.

google website optimizer experiment

Multivariate tests can test multiple elements on your pages simultaneously. This is the way to go if you want to test multiple headlines, calls-to-action, images, or other elements (or all of the above). It saves time to take this route if you have a lot of elements to test, rather than doing them one-at-a-time in A/B tests.

What Should My Conversion Rate Be?

So you’ve figured out what kind of test to run, but I’m sure a you’re wondering what your conversion rate should be. Some of you probably have an idea already, as you know you need a certain conversion rate to take the next step in your business. For those who don’t have a clue, though, I offer this simple rule:

Your conversion rate should be double your current conversion rate or 10%, whichever is higher. So if you’re only converting 1% of your current visitors, you should be converting 10%. If you’re already converting 10%, you should be converting 20%.

If you’re already converting a lot higher than 20% of your visitors, you may not be able to double your conversion rate, though it’s still a good goal to have. And if you’re already converting over 50% of your visitors, then what are you doing reading this article?

Don’t forget that KISSmetrics has powerful tools for tracking your sales funnel and figuring out just where visitors are getting hung up.

What Do I Test

Now that you’ve got your conversion rate goal in hand, it’s time to start figuring out what you should test. The answer is both simple and complicated. There are a lot of possible things you can test:

  1. The size and position of your logo
  2. The size, position, and color of your headline
  3. The overall color scheme of your page
  4. Headlines and call to actions
  5. The body copy (including both the copy itself and the size and font used)
  6. The layout
  7. Images used and how many to use (this is especially important for product pages)
  8. The call(s) to action
  9. Related products (how many to show, where to show them, how to show them)
  10. The length and field labels of any forms on the page

I’m sure you could come up with more than that if you look at your pages. It’s going to be different for each site, but deciding what to test is an integral part of running a successful test. If you’re still not sure or want more ideas, check out these two articles: Top 50 Google Website Optimizer Tips on What to Test and 64 Tips for Getting Started with Google Website Optimizer.

Shouldn’t I Just Start Somewhere? Anywhere?

You hear this all the time: just dive in and start testing! It doesn’t matter what you test, just start testing!

Sorry, but that’s wrong.

That’s like getting hungry, walking into a grocery store and grabbing the first thing you see. Sure, it might be something you like, something that will give your body the energy it needs, or it might be those plastic corn cob holder things. And what are you going to do with those in December?

google optimizer results

Think through what you want to test in terms of where you can have an impact. Make sure you’re testing the right pages (most likely your landing page, homepage, or product page, at least initially). Make sure you’re testing the right elements on those pages (your footer likely has little impact on whether someone makes a purchase, so why spend time optimizing that first?). The point here is to increase conversions. To do that you need to test things that actually have an impact on your conversion rate!

Deciding Where to Start

The best way to figure out what to start testing first is to look at your current conversion funnel. You likely have a homepage or landing page, some sort of product page, and a shopping cart page(s). So which page is bleeding visitors the worst? If 70% of people aren’t making it past your homepage, start there. If you’re losing half your visitors at the shopping cart, then test and optimize that first.

It only makes sense to optimize the worst areas first, as that’s where you stand to gain the most.


Once you know what pages to test, and possibly what elements, it’s time to brainstorm the possible variables. Come up with half a dozen alternate headlines or calls-to-action. Then pick the best ones and test those (or test all the ones you came up with). Look for different images to use on the page and add those to the test.

website multivariate testing

Remember, Website Optimizer will handle the serving of all these things, so once your experiment is set up, you don’t need to manage dozens of different page versions. It’s easier to set them all up in one test than it is to run additional tests over the course of weeks or months.

One thing to keep in mind is that very similar test variables tend to provide inconclusive results. You’re better off trying out radically different variables instead of variables where only a word or two is changed.

How to Set Up Your Tests

As already mentioned, Website Optimizer has two different testing methods: multivariate testing and A/B testing. They have technical guides for setting up each kind of test, and either you’ll need some basic knowledge of HTML (and access to your site’s code) or you’ll have to get help from someone who does. You can find the A/B testing technical details here, and the multivariate ones here.

ab testing experiment

But setting aside the technical details, how do you run a test with Website Optimizer?

The first step is to decide what you’re going to test and the alternatives you’re going to use. You should have already completed these steps based on the information above. Select your test page and your conversion page (what the final destination should be, such as a receipt page or a confirmation page).

Next you’ll want to set up the content to test. Remember that trying radically different variables is likely going to give you more conclusive results than slight variations. So your original call-to-action might be something like, “How can we improve?” and your new one might be “Tell us your ideas”. Effectively, you’re saying the same thing, but in two completely different ways.

Website Optimizer gives you the option to limit the test group to a certain percentage of your website’s visitors. This isn’t recommended, though, as it not only takes longer for the test to run and can give you less meaningful results. This is especially true for pages with lower traffic numbers.

When your test is set up in Website Optimizer, you’ll have to insert the code it gives you into your test pages before the tests can run. After that, all you have to do is wait for the results!

10 Tips for Making the Most of Website Optimizer

  1. Don’t just focus on completed sales or sign-ups. Focus on getting visitors from one page of your conversion funnel to the next. Making small gains at each step along the way can result in big gains in the end.
  2. Don’t overlook site wide tests, too. Try testing things like your header, navigation, or search feature to see if they make noticeable gains to your conversion rates. Improving overall usability almost always has a positive effect on visitor satisfaction.
  3. Consider segmenting your visitors to serve different tests for different kinds of users. You can do this based on things like the traffic source, the country or location the user is coming from, and more.
  4. If in doubt, test. Can’t decide between two different headline options? Is your company split over which image to use? End arguments with coworkers and others by testing out different ideas in the real world.
  5. Try what your competitors are doing. If your competition is trying out a new website design or sales style, why not try something similar and see if it actually works better?
  6. Run simultaneous tests. It saves time, which in turn means you can get your optimized page up and running for all your visitors much faster.
  7. Test big changes. We’ve already talked about testing radically different versions of elements on your page. But why not test radically different page all together? You might be pleasantly surprised with radically improved conversion rates!
  8. Don’t run your test for too long. Website Optimizer will let you run tests for as long as you want. But there’s really no point in running a test for longer than a couple of weeks, at most. All you’re doing with longer tests is opening up your results to normal fluctuations, or influences based on what your competitors are doing. (Obviously if your biggest competitor runs a huge sale, you may see a drop in your sales, but you don’t want to make changes based on those kinds of influences.)
  9. Make sure you have your goals clarified before you start. This goes beyond just your conversion rate. What are your company’s long-term marketing and positioning goals? Make sure your tests are reinforcing those things, and not deviating from them.
  10. Test regularly. Tastes change. The economy improves or declines. New products emerge. Make it a point to test new ideas and improvements on a regular basis. And especially test out new ideas if you see a sudden drop (or even a steady decline) in your conversion rates. Better to test some alternatives when your sales are only off by 1% rather than waiting until they’re off by 10%.

Be sure to check out our previous article, 35 Resources for Getting the Most out of Google Website Optimizer. Have other tips for using Google Website Optimizer? Share them in the comments!

About the Author: Cameron Chapman is a freelance designer, blogger, and the author of Internet Famous: A Practical Guide to Becoming an Online Celebrity.

  1. Kenneth Kowal Nov 10, 2010 at 10:37 am

    Thanks for the post. Getting people to your site is only half the battle. Getting them to take action is the hard part.

  2. I totally disagree with you on

    “Your conversion rate should be double your current conversion rate or 10%, whichever is higher. So if you’re only converting 1% of your current visitors, you should be converting 10%. If you’re already converting 10%, you should be converting 20%”

    Its not practical to get aggregate 10% rate in ecommerce.

    Lead signups can get 20%. but getting 10% aggregate conversion rate for e-commerce is not possible

  3. 10% is rare, true, but it really depends on your industry and your conversion goals. Non money related goals can have huge conversion rates. I’ve seen free newsletter signups well over 50%.

    Even e-commerce, in today’s well optimized sites with high quality traffic flow have some sky high rates. The top 10 e-commerce sites in 2010 (mid year):

    Schwan’s – 41.7%
    ProFlowers – 26.5% – 24%
    Woman Within – 22.4%
    Blair – 20.5%
    Lands’ End – 19.5%
    Doctors Foster and Smith – 18.6%
    Office Depot – 18.4%
    Roaman’s – 18.4%
    QVC – 18.3%

  4. Continuous runs race between “mechanics” of the company that enhance Google search algorithm, and consultants to optimize the links to adjust the new algorithm.


Please use your real name and a corresponding social media profile when commenting. Otherwise, your comment may be deleted.

← Previous ArticleNext Article →