Split Testing Overview

Split Testing Overview

Do you run an online store or service, and have you been actively split testing?

Have you ever noticed when you buy on a site like Amazon that sometimes you see an offer presented one way and then a few weeks later the same offer on a completely different page layout? They’re testing which style gets more sales across all their users. It requires some special software set up in order to do it smoothly but those extra sales can be critical. If your advertising is returning more sales, now you can afford to buy more advertising.

It’s been accepted practice that even in the days of magazine adverts you would work on at least 2-3 versions of the same ad and the headline of the ad will have had multiple versions to consider. With a magazine though you have to eventually agree to go to print with a final copy. Split testing turbo charges that process and allows you to actually test those copies you would have thrown out. What if one of the ones you would have thrown out actually made more sales?

In order to run a successful A/B split test you do need some special software installed, and the type of software is going to depend on whether you’re split testing your main page or an internal product page. You don’t want a customer to see one page, come back 10 minutes and see a different page. It’s both confusing and it ruins the test because they’ve now seen both versions so their response may be influenced by either.

As with everything else, and core to our company’s strategy, it must all link back to the analytics tracking. We don’t just want to know that one headline performs better, we also want to know that it performs even better if the customer typed into google "buy product" instead of "reviews of product".

That extra data could influence later decisions in the marketing process. Perhaps we’ve found ourselves ranking well for reviews but not for keywords relating to "buy". Well maybe our goal of making a sale is a little lofty for that keyword and we should scale back and just try and get them to register with their email & phone. Yes it’s a longer game and requires a bit more thought but we can build trust with the customer and make sales in a few weeks instead of immediately.

Of course split testing would also extend to us trying to get their email and phone. We could try a few different forms. Different colors, different positions. Maybe people are uncomfortable typing their phone number in right now, so we’ll try the form without it. Perhaps your sales team find they can close customers with phone numbers at a higher rate. That would reach the limits of what would be easily tracked with split testing and require a judgement call as tracking sales from offline avenues can be complicated to manage.

Contact us to find out more about ecommerce optimization strategies and Facebook marketing.

I need more:



Services We Provide

We cater to the problems and issues of small to mid sized businesses.

Marketing and Advertising

Marketing & Advertising

We use a statistics first approach to identify where you are losing sales. Statistics help us reliably get inside the mind of your clients.

View Details
Ecommerce Optimization

Ecommerce Optimization

Are people abandoning their carts before paying? Ask us to help you find out why.

View Details
Site Optimization

Site Optimization

Every single second longer a page takes to load loses increasingly more customers. Your plugin heavy site might look great but is it costing you? Find out how we fix it.

View Details