I’ve long been a fan of tools like Optimost, Verster, and Offermatica which allow marketers to do multivariable testing to find the best possible converting landing pages for their ad campaigns. Yet at the same time, I wondered in the back of my head how long it’d take for Google or Yahoo to integrate a tool that does it for people instead of having to use a third-party service.
Google has now launched Google Website Optimizer, a tool that’s part of Adwords that allows websites to test the various pieces of their landing pages to find the landing page that converts the best. When your winning page version is determined, you can then shut off the experiment until you’d like to test new combinations in the future.
How does it work?
You break your landing page up into sections using tags/code provided by Website Optimizer, each section is considered a variable. For example, the headline on the page is a section. You then provide multiple versions of that headline for Google to test. You can do the same thing with buttons, images, text, page location, etc. Website Optimizer then creates multiple combinations of your landing page and serves them to users and tracks how many conversions each version of your landing page create.
You then view data to see which combinations of your landing page perform the best, and how much improvement they generate over your original landing page. You can also see the data for each section to see how each individual section affects the conversion rate. When the winning combination is found, you select it to run all the time and let the conversions roll in.
What does this mean?
This looks to be a very powerful tool, and it has some ramifications if it’s successful. First, if advertisers manage to use Website Optimizer to improve their conversion rate, it will mean that they can potentially up their ad spend with Google Adwords. This is really why Google has built the tool.
The nice thing about this is that it can also trickle down to provide more revenue to Adsense publishers. If advertisers are spending more money, that money can trickle down to the publisher.
There are challenges
Multivariable testing isn’t easy. One of the nice things about the companies I mentioned above that have been offering services like this for years is that they are experts and will help advertisers set things up correctly. When creating a test, you need to go through the work of creating different headlines, text, images, and buttons to be tested. The more you create, it exponentially grows into a higher number of combinations needed to test. This means you need more traffic (which costs money) in order to get enough data to determine which combination was the best. And it’s no guarantee that many of the combinations won’t perform worse than your original landing page, meaning you could be spending money and getting worse results than before.
Basically, it’s quite easy to screw these tests up, it’ll be interesting to see how Google handles this since they generally don’t provide much personal support.