Web Hosting

The Benefits of Split Testing Your Landing Pages

One of the requests that I often see on the Warrior Forum is for members to review a re-designed landing page and say whether or not it’s better than the old one.

Clearly it’s good to get other peoples’ views on how effective a new landing page is but, at the end of the day, there’s only one opinion that matters:

The opinion of your visitors.

A landing page has a specific job. It either has to get opt-ins or it has to get clicks on affiliate or buy links.

So the only real test of how good a landing page is comes from how successful it is at getting your visitors to do what you want them to.

I usually offer some thoughts, but I always encourage the questioner to test the pages against each other. Because that’s the only way you’ll really find out which is the most effective.

I know, from my own experience, that testing your pages can create a personal challenge, because you have to accept that you might not like the result.

When I first started putting up landing pages I’d pour all my creativity into them. They became a personal labour of love. Testing them raised the possibility that others might not like my labour of love – rejection, horror!

However, if you’re in this to make money, the only opinions that matter are the opinions of your visitors.

I learned to get over myself on this issue – and, if this sounds familiar, you need to do the same.

So, how to test your landing pages?

If you already have a Google Adwords account then you already have an extremely powerful tool at your disposal – free of charge. The Google Website Optimiser.

Website Optimiser allows you to do A/B split testing, where you’re comparing the results of two different pages, or multivariate testing, where you’re comparing different combinations of elements, such as headlines and images.

When I started using Website Optimiser I kept it short and simple. I did A/B split testing of 2 different pages, which were identical except for 1 element.

So how complicated is it to set up an experiment? Actually, it’s very easy and straightforward! Here’s how:

  1. You need an ‘original’ page. This is your existing landing page, which you will use as the control page
  2. You need a ‘variation’ page. This is your new or re-designed page.
  3. You need a confirmation page. This is the page that your visitors go to after they’ve taken whatever action you want them to take. It could be your download page, your payment page or your thank you page.

Both your original page and your variation page must direct your visitors to the same confirmation page.

Once you’ve identified these pages you will need to upload them to your site and then give Website Optimiser the URLs. Website Optimiser will then check the URLs to make sure it can access them.

When it’s confirmed it can access all 3 pages it will give you some code that you need to copy into the HTML source of each page. The instructions that it provides for you to do this are exceptionally clear and easy to follow. So, although you should ideally be comfortable with copying and pasting HTML code, you really only need the most basic understanding to do this successfully.

After copying the HTML code into your pages you will need to upload all 3 of them again, and tell Website Optimiser when you’ve done so.

Website Optimiser will then check that the pages can still be accessed and that the code has been correctly installed. Assuming everything is OK, you’re all set to go.

You continue to promote and drive traffic to your control (original) page and Website Optimiser will split it between your control page and your variation page. It does this for all traffic hitting your control page, whether or not it comes from Adwords.

After a while you’ll begin to see which page is generating the most conversions. Clearly the bigger the sample you can measure the more accurate the results will be. Website Optimiser will tell you once you have enough data to see a valid result, but you can continue the experiment as long as you like in order to validate the initial feedback.

The beauty of doing split testing in this way is that you can continue to improve the effectiveness of your landing pages over time.

After your initial experiment you can convert the ‘winning’ page to the control page, create another variation page and start a new experiment. And you can repeat this process over and over again.

In keeping with the Keep It Short and Simple principle, here’s how I do my testing:

I take my initial landing page and make a copy of it. I then change 1 element and set up the experiment. So what I’m testing are two identical pages except for the 1 element I’ve changed.

I don’t change more than 1 element per experiment. If I changed several elements I wouldn’t know which one made the difference.

So what have I found so far..?

Well, here’s something that’s been repeated many times: small changes can make a big difference. It’s always been something that I’ve grappled unsuccessfully with mentally. If it’s a small change how can it be significant enough to make a large difference?

Well – I don’t know the answer to the ‘how’ question. But I’ve just discovered that it does.

I’m currently running an experiment on which I made one very small change: I removed the value of one of my opt-in bonuses from my variation page.

Although the experiment is only two weeks old there is already a huge difference. The page without the value of the bonus is getting more than twice as many opt-ins as the page with the value of the bonus. Now that’s a result I wouldn’t have expected in a million years!

Here are some other things that I have planned for future experiments: background colour, headline, opt-in box colour, font.

By taking the winning page each time and converting it to the control page for the next experiment you’re going to be able to greatly increase your conversion rate over time. And that equals a continuously shrinking cost of acquisition.

So – set up your split testing, have some fun and increase your opt-in rate. Just change small things, and only one thing at a time.

And be prepared for some surprises.

How do you test your different landing pages and what results has testing them brought you? Leave a comment and let us know!

Web Hosting