Increased Profits When Testing Segments
Last year we saw that having our contact number on the site was very helpful for many of our users. We spoke to our lead designer about the data and asked for some suggestions on how to take advantage of the insight.
That was when things got a little crazy. You see, the company up to this point had only used the contact number in 1 of 2 places – the header, and under the H1. So when we took the direction to the designer, he was not sure what else to do with the number. With a little collaboration and some creativity, we were able to find additional placement opportunities, here is what we did:
1. Place the phone number in all the places you can find – especially the ones that look like a bad idea.
2. Remove only the worst design offenders.
3. Present the results.
The designer came back with the results and I was amazed at the ways we could use the number – in ways we had not thought of before. You might be asking “aren’t the numbers in less-than-optimal locations on your page now?”.
A problem with experience
No. You see, when we are working with similar problems every day, we start to see the same answer to those problems. Even thought over time the problems change, if we are not careful we end up thinking the same solutions will solve these new problems.
With about 6 phone number placement options at our disposal on desktop, we set to testing. I am all for cluster variable testing, but in this case 6 seemed a little too much for a single experience. We decided to split the new number locations up into 2 experiences and ran them against the original page.
We found an experience that did not hurt conversion rates noticeably online, but had a slight increase in our phone call conversion rates. Shown here: (Can you find all the places we added the number?)
Translate Learnings
We thought this did so well on desktop, we should try it on mobile.
We set to translating the phone number placement ideas into the mobile website and the results were astounding.
We experienced a 14% drop in our mobile online conversion rates, and we saw a 14% lift in our phone call conversion rates. You might think that is a wash, but you would be assuming conversion count is the same online vs the call center… but it isn’t.
The Learnings Drive the Added Revenue, Not the Ideas
We learned a very important lesson, our mobile users were not just a little different from our desktop users, they were completely different. You see we get a lot of our call center orders from users on our mobile site, and our mobile site orders are very small by comparison. By testing multiple times around what we believed to be different audiences, we were able to find lifts that contributed positively to the company.
This has lead to amazing testing opportunities since, and once again, represents large gains in profit for the company.
Scientific Web Optimization 