Improving online conversion – removing header from registration pages

At the moment I am trying to improve the conversion of my products online registration process.

As discussed in previous posts I have tested playing around with how a customer interacts with the landing page to begin registration.

Now I will focus on the registration pages themselves.

I asked myself are there any distractions for customers throughout the registration process that might cause them to leave the process and hence damage my conversion?

The answer was yes, the header that appears throughout the process is full of links that cause the customer to leave the process.

Should I remove this header?

Are customers leaving the registration process because they are clicking on links in the header?

Or does it provide the customer a sense of familiarity as these pages are consistent with previous pages by showing the header?

The answer is I don’t know and I should test it!

I imagine the answer will be it will increase conversion when I remove the header.

Take for example Amazon, well known for their success when selling products online, there is a header shown prior to entering the checkout.

As soon as the user enters checkout where they will pay for the products selected, the header disappears and the user has no option to leave the purchasing process other than the back button on the browser.

So next steps for me are to test the registration process without a header using A/B testing and see what the results are.

This time I will try out Omniture test and target to compare with google web optimiser and see whether it is worth the cash paying for a professional tool.

Tags: a/b testing, google weboptimizer, increase conversion

A/B Testing the results

Button placement of call to actions on a website is critical and considering the main purpose of my site is to get customers to register for the product I decided to A/B test some different placements to see what impact it would have on conversion.

As discussed in my previous post, my success criteria was a 20% uplift in conversions. Defined as customers who successfully register for the product.

I tested 3 new versions of the product’s website with the call to action buttons placed in different spots.

The winner lifted my conversion to 8.05%, a huge 24% more effective than my original!

The winner:

So why did this version beat the others?

Some theories:

• In this version the call to action buttons are at the top left of the screen. We know from research that customers read from top left.
• The buttons are above the fold regardless of size of device.
• Possibly by the time customers reach this site, they have already made the decision to purchase the product so putting the register button in top position satisfies most use cases.

Version 2 (least successful)

So why no luck with this version? Is it the fact that nothing in the top left of the screen is grabbing the customers attention?

In the next version, the buttons were again at the bottom, however there was no gap between the header and the product information, which suggests that you shouldn’t waste the top left space on your page!

Version 3 (came 2nd)

Notice here the register button is twice the size of the register button in the winning version. Goes to prove that placement of the call to action is more important than size. 

So in summary A/B testing has been a huge success for my product. In the space of 2 weeks I have uplifted conversion by 24%! All using free tools and by simply moving the call to actions.

Next I plan to try some testing on copy and images.

Tags: a/b testing, button size, call to actions, increase conversion

Getting A/B testing up and running

A/B testing is one of the webs greatest tools for product managers, its cheap and easy to implement and lets us try out our ideas to be sure it’s the right move.

Getting A/B testing implemented in an organisation which has a severe case of the bike shed effect was not an easy task so I thought I would share my experience and the lessons learnt along the way.

You can use A/B testing for many purposes, like testing marketing messages, fields you capture in forms, images you use on pages, the list could go on and on. I have chosen to start my A/B testing on the placement of 3 crucial buttons (register, reactivate, manage) on my products homepage for 2 reasons:

1. Getting customers signing up for my product and ultimately the cash is my number one priority. Customers clicking the register button is the start to this.
2. I can bypass the “bike shed” effect, its not like I am changing marketing messages or copy, hence no long tedious unnecessary approval process needed!

I would recommend this to all product managers trying to get A/B testing implemented in your organisation.

Three simple rules for picking what to test first:

1. Pick something people won’t care so much about or want to change
2. Something that impacts as little departments of your organisation as possible
3. Most importantly that will improve one of the products KPIs hence getting you results and a reason to promote the benefits of A/B testing.

I did hear some arguments against A/B testing in my organisation, below is how I responded.

• “It’s too risky, we could lose registrations.” – results can be monitored real time using reporting. If one of the versions being tested is under performing we can see this real time and turn off this version instantly.
• “What if customers come back to the site and notice it is different? We don’t want to confuse customers.” – the tool remembers which version you were served the first time you visit so they will not see the other variations.

Once you have picked what to test you need to pick the tool you are going to use. For me the choice was simple, google web optimiser , its free, it includes reporting, its super easy to set up and give people access to its reports. It is the perfect starting out tool. However of course there are lots of other tools out there like Omniture test and target or optimizely to chose from.

Before you start your test it is super important to define what success is and the sample size you will test on to ensure you get reliable results.

1. What is success? The easiest measure is conversion and for me conversion is the number of people who end up registering for the product. You can of course use other measures such as button clicks, opened emails, etc. My current conversion rate is 6.5%, I wanted to increase this by 20% which meant a 7.8% conversion rate.
2. What is the sample size? Don’t just say I’ll do it for a week and see how it goes, either hit the statistics or use one of the free calculators that use the algorithms for you. I used the one available here and it told me I needed 3488 views of my landing page or 1744 views of each option being tested to be 95% confident of a relative improvement of 20% and that should take me about 4 days. So helpful!

Now you know, what tool you will use, what success means and what the sample size is you are ready to go!

Next post I’ll talk about what I found out…

Tags: a/b testing, google weboptimizer, increase conversion, sample size, thebikeshedeffect

Lets get straight to the point…

Who am I?

Product manager working in IPTV in Australia for a company I’m not going to name (not sure I’m allowed to disclose but I’m sure you could probably figure it out!) Fairly new to product management and super keen to make some useable entertaining products that change the way Australians consume entertainment.

Why write a blog?

I’m learning new things every day, and like many others I am trying to master this Product Management business so I thought why not share my rants and learnings.

Why painting the bike shed?

The single most frustrating thing I have found as a product manager is product development being slowed down by people who want to add their 2 cents every 2 seconds to the simplest of concepts and the bike shed effect proves the crappy effect this has on projects. The bike shed effect says that for every discussion had on a change it is inversely proportional to the complexity of the change. Perfect recent example from my workplace, managers from pretty much every department have argued about the name of one menu item for over 5 months. In this time we could have built the entire product they are still arguing about.

Tags: introduction, Parkinsonslaw, thebikeshedeffect, triviality, whoami