Is there a formula for helping you decide what parts of your website you should test? This is an essential question when it comes to website optimization. However, an even more relevant question is this one: How can you actually determine what you should not test?


1: Check For Evidence

If you have no evidence there’s a conflict or issue of some sort, then you shouldn’t test anything. There’s a reason such ideas are known as theories. A theory is just something you may come up with, sprinkle some logic on it that works for you and there you go. However, this is more of a brain fart, honestly speaking. Therefore, you first of all need to educate, then inform and finally come up with data for your ideas, because if you don’t, then they cannot be called a theory. They’re just things you think are cool to think about.

One of the main pros of starting to look for proof is the fact that you can remove a theory. It could be that the evidence you find suggests there’s no problem to solve in the first place.

Just check out this example of a theory for an electronic commerce website’s product page: If the Add to Cart button is placed at the footer of the page, there’s a good chance more users will add one or more products to their cart. Initially, this sounds, right, but if someone scrolls down the webpage, this theory will fail.

Just check the heat map or attention info generated by software programs like CrazyEgg. They’re going to tell you how far those who visit your website scroll on the product page. If they don’t scroll that much, then this hypothesis may come in handy at a later date.

When trying to think about what you need to test, every theory you come up with should be rated from one star to five stars for the amount of evidence you have for it.

A theory that’s rated one star has no evidence, so it’s useless. A rating of 5 stars does show that you have a lot of evidence you’re dealing with an issue that your theory can fix.

2: Measure Traffic

Optimizing parts of the website that you shouldn’t optimize is something you need to avoid. Your list of theories needs to contain ideas for page specific and site-wide improvements. For instance, if the structure of the website’s navigation is altered, then this can be considered a site-wide alteration. On the other hand, you can call a page-specific change one where trust symbols are added to the checkout page. If you’d need to rate your traffic’s value from one to five stars, how would you rate these 2 scenarios? In fact, there’s a good chance they’d be rated five stars.

Site wide changes usually have a full impact on every visitor, so it should be rated five stars. But when you change a webpage, it’s something that only twenty percent of visitors see, so you should rate it one star.

In general, only a minute number of visitors will end up checking out. The reason you should rate them 5 stars is because even if their number is small, they do make up for that in opportunity. People who buy something from your website and then check out have proven a major intention to buy and therefore, are very important for your business.

There are of course, pages that won’t get too much attention, such as the FAQ and About Us pages, so in this case, they may be rated two or three stars.

Prioritize theories that have a direct influence on the most interesting or the majority of visitors.

 3: Is Testing That Difficult?

The level of effort for each of your theories needs to be well understood. For instance, adding live chat to a website is a lot more complex than changing the font of an offer.

By using the one to five scale once more, a modification of the copy can be rated one or two stars, but implementing live chat involves paying a vendor, considering integration and hiring new employees to chat with your visitors. This is rated 5 stars.

Never favor easy tests because it’s easy to do so. Likewise, it’s also recommended that you’re patient with your fives and leave them for when the time is right.

 4: What Has Experience Taught You?

Eventually, you’ll need to estimate the impact your theory is going to have. This is something that’s based on what you have learned from the tests you previously conducted. It’s based on the research you have done and on your experience as an internet marketing team.

What about a one to five star rating again? If a theory is rated one star, what you’re saying is that this is just an idea. But if it has a significant impact, it’s going to take you by surprise.

If your theory is rated five stars, then this means you believe the change is going to have a major impact on your website’s visitors, but also the website itself. As a result, you’re expecting a great win.

Intuition though, usually leads us astray and therefore, don’t be surprised when you’re going to find yourself rating theories a lot higher on the impact scale because you’d like to try them and not due to your experience. It can also be that you favor one, since the idea sounds interesting.

However, in the scientific environment we’re creating, these types of sentiments simply do not belong. It’s true, though, that experienced businessmen have a great intuition that cannot be ignored.

This is only 1 of the 4 factors we consider, with the rest being the level of effort, traffic value and proof. A theory can be tipped into the top ten by a great impact score, yet this is only possible if it managed to score positively in the rest of the categories.

Once a theory has been disproven or proven, there’s no more room for intuition. When you have specific data in front of you, you’re always going to prefer it rather than intuition. On the other hand, when thinking about what to test, it may be a good idea to add in some intuition.