|
While search engine optimization is absolutely essential in today's online marketplace, and can be - by far - the most effective mechanism for attracting highly qualified, targeted traffic to your website, the next largest question always remains: What do you do with your traffic once it gets on your website?
Every website has goals, and in some way or another an eCommerce website will have goals that can be translated into revenues and profits. For the sake of this article, our goal is the online sale of a birthday cake. In this example, our online bakery and fulfillment center provides cakes, desserts, and cheesecakes to its online customers, and relies on being found on search engines organically as well as through pay per click campaigns. So what does the traffic find once they get on our example website?
Well, an online bakery would most likely have the ability to browse and purchase a number of different dessert products, recipes, and accessories. For example, the site may have its best profit margins on the sale of chocolate cheesecake. If this is the case, a reasonable suggestion would be to sell more chocolate cheesecakes!
So how do we do that?
The answer is testing. A/B split testing, where we compare one variable against another, or the more powerful multi-variate testing that allows us to compare relationships between a number of variables, allow us to see what presentation of a website gets you the most conversions on your goals.
Let's take an example. If we wanted to sell chocolate cheesecakes on our website, what do we think would happen if we chose the featured product on our homepage to be diabetic desserts? We might expect that we'd sell more of our diabetic desserts and less of our chocolate cheesecakes. A simple test would be to change the featured product on your homepage. The first step is to obtain statistics over a statistically significant sample size and then show half of these visitors a chocolate cheesecake featured product and half of the visitors a diabetic desserts featured product. The next step is to record conversions (raw and as a percentage of traffic - easily achieved with today's analytics packages) and see what happens. A few calculations will quickly tell you if your featured product does indeed have a significant effect over how many people purchase your chocolate cheesecakes.
To illustrate the idea of multi-variate testing, let's add another variable into the mix. Perhaps the background color of the site itself. It is certainly a feasible hypothesis that having a chocolate brown background to the site would entice more people towards chocolate cheesecakes, and a simple test would show the results. We can run a subset of the various combinations of these variables to see how they interact with each other - put simply, what is the ideal combination of background color and featured product. The major advantage of multi-variate testing is the ability to test the relationship between variables. Say you found that individually, the diabetic desserts make you more conversions and more money, and that the chocolate background was better than the white background. BUT, when put together, the chocolate background makes people shy away from diabetic desserts and the combination of these variables reduces the overall conversion rates of the whole site! Relationships can be key in this type of testing.
Remember, it's almost always easier to convert the traffic you already have on the goals of your website than it is to get new traffic. So next time you think about adding birthday cake delivery to your list of offered products, make sure to do a little testing to find out if that's truly the way to go!
|
|
|
