Multivariate testing diagram
There is a lot more to running a website than getting the code right. To make sure your website is getting the views it deserves, you need to understand who exactly you need to reach and how to target them. This can be achieved by using several tools.
To be frank, this is entirely new territory for me. Indeed, I have heard of SEO, or Search Engine Optimization, but I was not aware of the many tools that fall under this domain. The two in particular that I looked at this week are split tests and multivariate tests.
You might be wondering what split testing and multivariate testing are. I certainly did. Thankfully, the required class reading filled me in on these very interesting mechanisms that optimize websites.
First, split testing allows different visitors of a webpage to see different versions of one of the individual webpages. Through this technology, a designer can test out changes to their webpage before rolling it out full-scale. For instance, half of visitors will see an existing headline for a homepage, while the other half will see the new headline for a webpage. Now you are probably wondering how anything is actually “tested”. This part, I cannot explain in detail but what I do know is that software keeps track of which version of a page generates the most positive responses or conversions over a period of time. A conversion is an industry term that can mean anything that the website is aiming to do—have customers place an order, click somewhere, register, or sign up, etc. So now you are probably wondering how this is helpful. Ultimately, you can test two different versions of a webpage and find out which one is more successful, thereby allowing your website to better reach its objectives.
Next, is multivariate testing, which is basically the same as split testing but taken a step further. Split testing can take time to be completed depending on how often a website receives conversions. This is where multivariate testing comes into play. With this technology, you can conduct several tests on different pages versions at once. So if you need to test two different buttons along with different conversions, this is possible.
Example of Google Website Optimizer data
Thankfully, for novice web designers or for people who are on a shoestring budget, testing can be free through Google Analytics, which can perform split testing or Google Website Optimizer, which can conduct split testing and multivariate testing. However, people with larger budgets can pay for more advanced multivariate testing services. These can range from hundreds of thousands of dollars to a few hundred dollars to set-up testing and maintain it.
After delving into this topic I have a few lingering questions. What percentage of websites use split testing or multivariate testing? How accurate is it? When could you justify paying hundreds of thousands of dollars on these testing mechanisms? Would these tests be worthwhile for a portfolio website?
Since I will be taking a second stab at designing and coding my online portfolio per course requirements, I will definitely consider these testing mechanisms. I look forward to learning more and implementing these optimization tools in future designs.