The next step in your discovery process is determining what you should be testing. Don’t just jump into testing immediately because you know what might work. 

Step 2: List Hypothesis

You’ve just spent a decent amount of time gathering your data, so use it wisely. Engaging a conversion rate optimisation consultant (or specialist) may help ensure you are on the right track. 

What do you want to test?

Go back through all of your data and look for the common trends in dropouts for your website or the pages that people don’t spend a lot of time on. 

Look at the customer satisfaction surveys and note down any trends with your free text fields. Do people have a common issue with your payment page? Perhaps you need to test ways to make this seem more secure?

Make sure you have a key section in mind that you want to test to make sure you know what your end goal is. Conversion rate optimisation consultants typically use their data analysis processes to identify what is worth testing and what is likely to make a difference.

Who do you want to test?

The next stage is to consider who you need to test to get some accurate and usable results. 

You need to understand your demographic, which we now know that you can pull from Google Analytics. Once you’ve got this you also need to understand why they come to your website. What is it that they’re looking for when they search for your product? And what is their mindset when they click on your site? If you specialise in selling craft items, do you get an uplift in people on your website at Christmas, for example? Their goal might be to buy Christmas presents. If you’re a florist, does your demographic differ by gender or do you get an increase in traffic around Valentine’s day?

The trick is to understand the customer and what they need out of a purchase. You must use websites yourself in the same way, so consider yourself a customer on competitor’s websites and understand what you need from them too, to gain a different perspective. 

You’re never going to be able to segment everyone perfectly because your data isn’t that detailed and the conversion rate optimisation process doesn’t allow for you to check into every single potential customer, but you’ll still be able to get a good idea of the mindset of the wider part of your customer base. 

If you’re able to segment the different types of people and the different reasons that they access your website, you should be able to create test pages to send them to. 

Creating a test page is just generating a copy of your original page with your main URL but changing certain aspects of that page to test the differences in traffic and conversion. You’ll even be able to direct certain segments to your test page and others to your standard pages so you can compare the difference. Optimizely offers a great feature for this and will provide a conversion rate optimisation consultant to help you along the way. 

Which pages do you want to test? 

Obviously, you’ll need to decide which pages you’re testing with. If you’re conducting A/B testing by sending some customers to your original website and others to amended pages as discussed earlier, you’ll need to make sure that the pages have the same or similar information and products available on them. 

It’s no good sending some of your traffic to a page listing a product for $100 and the other half of your traffic to a page with something worth $5. Your customer base for those products is likely to be very different, meaning that you won’t get accurate readings. 

You need to know your exact conversion rate optimisation hypothesis before you make the decision of which pages to test. Use your data from the previous phase to determine the demographic of your traffic, the reasons for their visit and the reasons that they leave the site without making a purchase. Identify the problem pages with the highest amount of traffic first, as testing these will have the biggest positive effect. 

Conclusion

If you’re managing this on your own, there is a lot to consider. It can be beneficial to reach out to a conversion rate optimisation consultant. They can either verify your approach or conduct the analysis for you, helping to speed up the process and ensure everything is done correctly. The key takeaway is that every test run needs a transparent and clear conversion rate optimisation hypothesis.