Better Website Testing: Five Steps To Knowing What To Test

Website testing can help you figure out what’s actually converting, and what’s not, on your landing pages.

You spend all that time and effort optimizing and managing your PPC AdWords campaigns to drive targeted traffic to your site and landing pages, so take advantage and optimize for the conversion, too.

By testing out all the various parts of your website and landing pages you can end up with a well-designed and intentional converting machine, which can help you increase your business ROI.

We’ve partnered up with VWO to bring you these five website testing ideas so you can you start to optimize your landing pages today.

Set Your Optimization Goals

You’ve got a long list of things you want to test out and need to start attacking the items on the list, in order to optimize your landing pages.

Wondering what to test first?

Set your optimization goals and figure out which order to test your various components. There’s a handful of prioritization frameworks out there to choose from. Here are four of them:

1. PIE – Potential, Importance, Ease

This one’s pretty popular and measures: how much improvement can be made (potential), how valuable traffic is (importance), and how complex the testing will be (ease).

2. ICE – Impact, Confidence, Ease

There’s two ICE frameworks. This one helps you prioritize what to test first by: measuring the impact of the test (impact), confidence level in the test working (confidence), and how easy the test is to implement (ease).

3. ICE – Impact, Cost, Effort

The second ICE framework helps you prioritize by: measuring growth and company benefits (impact), determining the cost of implementing (cost), and understanding the amount of resources required to test (effort).

4. PXL – ConversionXL’s framework

Our friends at ConversionXL came up with their own prioritization framework, where they make the potential, impact and ease of implementation categories more objective.

By prioritizing your optimization tests, you can have a better idea of which optimization goals you want to test for and reach first.

Remember to match up your optimization goals with your overall business and revenue goals.

Improved click through rates and conversion rates can be beneficial but only if they contribute to improved revenue streams.

If the optimization process isn’t going to make you more money in the end, then  you may have missed the mark.

Identify What to Test

Drumming up a list of which items to test can be a tricky part of the process. This takes place even before prioritizing your test list.

VWO recommends testing out almost anything on your website that affects visitor behavior.

Here’s their recommended list of 12 things to test out:

1. Headlines – Is your headline explicit and clear?
2. Sub headers – Does your sub header further support your unique value proposition?
3. Body copy – How’s the paragraph text readability?
4. Testimonials – Are your testimonials accurate and realistic?
5. Call-to-action (CTA) copy – Does the CTA threat level match your audience intent?
6. Links – Are you shooting for a 1:1 attention ratio?
7. Images – Does your hero shot portray your offer benefit being used?
8. Content near the fold – Which is the best landing page length for your offer?
9. Social proof – Are your social proof numbers high enough to publish?
10. Media mentions – Does inclusion of media mentions build or hurt your credibility?
11. Awards and badges – By including awards and badges are you building trust?

There are numerous other things to test for to improve your conversion rates. For more ideas on usability testing, check out 27 usability testing tips to help you win more conversions.

By combing through all the various parts of your landing pages and determining which pieces of the landing page anatomy to test, you can uncover some items with pretty significant improvement potential.

A/B Test Your Website

With A/B split testing you can isolate one factor at a time to see which of two variants accumulates more conversions.

Here’s an example of what an A/B split test can look like, where you have two variants testing for the same thing. In this case, Novica is testing for which email capture format is optimal:

By conducting the A/B split test, Novica was able to find out the email capture in Version A had an overall 67% lift in email submissions over Version B.

Definitely worth the testing effort…

VWO suggests following these six steps when A/B testing:

1. Study your website data – By analyzing metrics within your Google Analytics reports, you can uncover insights that can point you in the right decision-making direction.

For example, an attribution report like this can debunk your judgments about visitors coming in from your PPC display ads:

Although the display ads didn’t bring in direct conversions, they definitely played a hefty role in assisting with conversions. Now you can dig deeper into your reports and find out more specific trends.

Use these stats to figure out which parts of your site are high-performing and which aren’t converting as well. Then use those issue areas as testing topics to hypothesize areas of needed improvement.

2. Observe user behavior – Usability testing can help you find out how real people are actually interacting with your websites. By moderating in-person or recording a remote session, you can observe your participants’ behavior and look out for any major bottlenecks that might stand out.

You can observe user behavior using three usability testing methods: moderated in-person, moderated remote, and unmoderated remote. Here’s what a moderated in-person session looks like:

You can observe and record the session to analyze later – image source

3. Construct a hypothesis – After checking out your Google Analytics stats and observing your users, you can come up with more specific ideas to test. Maybe there are certain hypotheses that come up just by analyzing the data you’ve gathered thus far. Hypothesize on any glaring stats first.

4. Test your hypothesis – This is the stage where you implement your A/B testing and create variants for your subjects to test out. Find out if your hypotheses are true or not.

To calculate the appropriate test duration for your monthly visitors, current conversion rate, and expected change in conversion rate, you can use VWO’s Bayesian calculator.

5. Analyze test data and draw conclusions – Collect the results from your A/B testing and decipher what all that gathered data means. You should have a much clearer picture of which variant will optimize your landing pages and website.

We have a winner – image source

6. Report results to all concerned – Sharing is caring, especially when it comes to internal squad members so share your reports and findings with those that need to be in the loop.

When team members from your marketing, IT, UX, UI are all clued in, you’re likely to all be on the same page and ready to move forward as a collective group.

Multivariate Testing for Combination of Variations

Let’s say you have numerous items and combinations of each that you want to test, but it just won’t work with A/B split testing.

No worries, multivariate testing is the way to go. Here’s how VWO draws it out for us:

Testing for headline and image combos – image source

You can test out combinations of different items that you want to test, and you can test it in one session.

This can help you save time by knocking out various test items in one fell swoop. Just be sure your website testing process is calculated, intentional and fits into your optimization goal.

Split URL Testing for Heavier Variations

To test out more broad variations of your website you can use split URL testing.

With a split URL campaign you can test multiple versions of your website hosted on various URLs.

With VWO’s split URL testing feature, you can have two landing pages for your website hosted on two different URLs and split the traffic between the two to find out which one converts better.

Here’s an illustration of what that could look like:

Testing two landing pages on two different URLs – image source

You can take it a step further and set conversion goals for things that you want to track on your page.

Here’s how you set it up in VWO’s dashboard – GIF source

Tracking things like page visits, engagements on a page, form submissions, click on links or elements, revenue generation and even custom conversions, can help you optimize your specific URLs even more.

Closing Thoughts

Now that you’re up to speed with optimizing goals and knowing what to test, you can try out A/B testing, multivariate and split URL testing for yourself.

By sticking to your goal plan and conducting your testing process diligently from start to finish, you should have the highest converting version of your website and landing pages.

Now get optimizing so you can bring in that ROI…

Cynthia Meyer

Content Marketing Manager