Your website isn’t set in stone. Savvy online marketers are always experiementing with their website’s appearance. What was generating sales last week, may not be generating sales today. Paula touched on the importance web design has on your visitors in her awesome post about bounce rates. In this post, I’m going to talk about how to experiment with your design to see what works best.
Google has implemented Experiments into Analytics. Experiments let you show different visitors different versions of the same web page. For instance, you may show Visitor A your current “Buy Now!” page, and Visitor B a new design that you’ve created. Analytics will randomly alternate between the two pages and record which one performs better in terms of conversions.
Sounds interesting! How do I do it?
Log into Google Analytics, then click through to Content -> Experiments to set up your experiment.
Here’s a step by step guide on how to do it. I’m going to experiment on E-Web Marketing’s own contact us page for this example.
Here at Choose experiment pages, I enter http://www.ewebmarketing.com.au/contact-us/ as the original page I want to experiment on. I also input http://www.ewebmarketing.com.au/free-proposal/ as the variation to be tested against the original. What I want to see is how our free proposal form measures up against our contact us form.
On the next page, I Set experiment options by giving my experiment a name. I also specify what percentage of visitors to test at 50%, meaning that 50% of visitors will be in the experiment. (This does not mean it will show 50% of visitors the free proposal page).
Now that the experiment has been set up, the only step left is to add the special experiment code the page. If you’re comfortable with editing HTML yourself this is extremely easy to do. Take the code that’s given to you by Analytics and copy and paste it into the ORIGINAL page, before the line that reads </head>. In this example, the original page is http://www.ewebmarketing.com.au/contact-us/.
Alternately you can have Google email your web developer to have them implement it.
Once you’ve done this, the final step is to let Analytics verify the code is correct and start the experiment!
Press Run Experiment now, and you’re done!
You can have more than 1 experiment running at 1 time, and you can test between 3 or 4 different pages in the same experiment. it’s a good idea to start slowly though so you don’t get overwhelmed with seemingly contradictory data.
After a few days/weeks/months, you’ll be able to help improve that extremely important conversion rate and start turning more of your visitors into customers.
Let me know how you go with your experiments and if you need any help interpreting the data.