Working as a conversion optimization officer for a few years now I want to share the process I use to increase the conversion of websites. This method has proven successful for me on all various websites I have worked with.
Previously I wrote more in general about this process, this is a more detailed how to guide. You will find the previous article here: My A/B testing guide.
1. Start by looking in your web analytics tool.
a. Pull out your top 20 most trafficked pages.
b. Look at those pages on traffic volume, conversion rate and bounce rate.
2. Now apply the segmentation of traffic source – direct, seo, ppc, social media, other websites.
a. How much traffic do you get from each traffic source?
b. What is the conversion for each traffic source?
c. What is the bounce rate for each traffic source?
d. Notice any difference. How could you explain the difference?
3. Pick the page you think has the most potential for improvement.
a. It can be either low conversion or…
b. …high bounce rate
4. Apply any search engine or ppc keywords.
a. Keywords tell you a lot about behavior i.e. if people enter the site with the keyword “buy a bike” that says something about users intention. Compare it to the keyword “best bike” which is not at all as strong user behavior as the first one.
b. How similar are the keywords? Can you group them?
5. Look at your web analytics data for next page flow and the click density. Basically you are looking for user behavior. Anything that can explain why people act as they do.
a. Where are people going after this page?
b. What elements are the most clicked on?
6. Look at your chosen page and try to determine which of these areas you want to improve by testing.
a. Headings – are they crisp and attractive? Do they make people want to take action? (Read my post on the top 3 most converting headings)
b. The offer – how attractive is your offer? What are the benefits for the user?
c. How clear is the layout of the page? Do people know what you want them to do next or are they left with many different options? Do you need to rearrange the page elements and remove some to make the action more clear?
d. How clear are your elements? For example is your buy button dull and hidden?
7. Cooperate with your designer to create a wireframe and a new design to test.
8. Setup the AB test.
9. Measure the effects
a. Where do people click on the new design.
b. How big conversion increase did you have? Are your numbers significant? You want statistical significance to be able to draw correct conclusions.
10. Apply the learning and move on to the next page or redo this page.
Notice you can also apply this on a section level. I have done it many times by looking at our review pages and try to improve the whole section. I do this mostly when the traffic is not big enough to reach statistical significance within two weeks.