A/B Testing the Long Page Version and the Short Page Version

This article comes from a webinar by the Convert Academy,  that saw Carlos Del Rio the formal Director of Conversion, Analysis and Data Strategy at Unbounce, host a presentation on how you can create awesome landing pages and the various factors you need to consider.

Del Rio has also authored a book titled ‘User Driven Change. Give Them What They Want’ that talks about picking web traffic data and what visitors try to accomplish on your site. He used an example of the Unbounce landing page to get his point across on this one, and the various tests they had to run to achieve high conversion. You can view the entire webinar here.

Case in Point

The first step that was used when running the tests for the Unbounce landing page was creating a big page with everything. Basically, a page with lots of content. The page contained:

• Text
• Images
• Videos
• Headlines
• Thumbnails
• Maps
• Links
• Icons

Next, they stripped the page of the all the things that were going to cut in on time, which wasn’t going to happen on schedule, but was needed nevertheless, and they developed a shorter version of the page. That’s how they managed to create the control for their paper book and each has a specific hypothesis about what might make the page perform better.

They started at a conversion rate of 10% as the base, and when they did get around to testing, they discovered all their hypotheses were negative. It’s not something to get overly excited about, but the truth of the matter is that…

[Tweet “in A/B testing, even a negative result can teach you something: you know what exactly needs to be changed, or done differently”]

Figuring they each had to take a different approach, they rebuilt the page making some headline changes, content changes and some link variation which bore very good results. The trick was to make the page shorter and getting the right amount of questions for the particular need they hoped to achieve.

But what exactly is the right amount of questions?

Do you know who your website visitors are? How much do you know about them, where they come from and what they do once they are on your site? How often do you walk through the process as a consumer? Do you enjoy the process? Do you have information and tools available on your landing pages which help the customer make their decision to convert? – Selligent

Simply put, it’s the amount of content that is required to communicate value. Take this an example; If you are taking someone through something that is just as simple and highly actionable such as downloading a whitepaper, then you don’t need to furnish them with lots of information – you just stick to the highlights and how they are going to get it after taking the action they need to take. This is especially good when you want people totally unfamiliar with your brands to make a purchase with you.

Both Ends of the Spectrum

When it comes to your landing page, testing the very short version and very long version of it is probably a good idea. You may not be using the long version but it’s advisable to test it as well since then you’ll have lots of content. Having at least one long version of anything makes it very quick to do the tests. Don’t overlook doing a headline test too, since this could yield a high converting landing page within minutes, as the guys at Unbounce realized with their tests.

Another thing that may not be leading to highly converting landing pages is failure to focus on what the visitor is trying to accomplish. Have it touch on something that is of benefit to them – not just focusing on what the product is doing (features). Most people don’t attach too much relevance to the landing page, or rather don’t really have a good definition of what a landing page is. But one of the positives is that they do know the conversion rate, at least. If you want to find out more about this, or if you want to check out the full webinar, click here.