Every digital marketer stands to gain valuable insights and improve their campaign results by using A/B testing.
It would be easy to spend all the time in the world testing different variations on your content – which font size scores the most conversions with over 50s in Slough?, you might well wonder. Going in-depth with your A/B testing is tempting, but once you’ve investigated the most crucial variables you will likely fall subject to a law of diminishing returns.
To make the very most of your experiments, aim for major, transferable insights from the word go. Here are four shining examples to point you in the right direction:
Faced with poor banner ad performance in an advertising campaign for customisable laptops, a marketer at Sony Europe used A/B testing via Optimizely to identify the a better approach to improve click-through-rate and shopping cart adds.
The campaign’s existing ads intentionally avoided emphasising the customisability of the laptops. Was this the problem?
To find out, Sony created two new variants of the ad. One placed full emphasis on customisability; the other advertised a promotion. All three were pitted against each other in an A/B/C test.
The results were revelatory, with the first new variant – the personalisation-focused ad – achieving the best results by a significant margin. Compared with the control version, it increased click-through by 6%, and shopping cart adds by 21.3%. If you consider the volume of units Sony Europe sells, it’s not hard to make an educated guess at the value of this insight.
Promoting personalisation aspects
Never underestimate the power of personalisation. Whether consciously or not, many consumers regard the brands they buy into as facets of their own identity. If the product or experience can be tweaked according to their personal taste, the effect is amplified.
Is there a customisable side to your product or website? User profiles to create? Custom product choices to select? If so, try setting up an A/B test with a variant that places the emphasis on those personalisation aspects – Sony’s success here suggests it’s worth a shot!
Social proof is one of the factors you’ve got to get absolutely right in order to achieve a good conversion rate – especially if some of your visitors are newcomers to your brand.
There are several social proof mechanisms in use around the web, prominently including client testimonials, star ratings, industry authority badges and client logos. Your job is to work out which can do the most to nurture trust in your site and boost conversions.
Media analytics firm comScore used A/B testing to shed light on this very problem. The company’s homepage was already offering social proof in the form of a testimonial, but the element was not displayed prominently, and the source – Microsoft – was not backed up with a company logo.
comScore tested three new variants of their webpage against the original – each with a tweaked design, one of which featured the client’s logo:
The variant displaying the Microsoft logo emerged as the clear favourite of a 2,500-strong audience sample, with a product page conversion rate 69% higher than the original version could offer. The lesson here is clear: if a company like Microsoft is willing to get behind you, you should make a song and dance about it.
Note how much more convincing a testimonial headed with a logo is than the testimonial alone. There are a number of explanations for this phenomenon:
Establishing social proof
There’s no such thing as a universal recipe for perfect social proof – every business and every industry differs in its demands and opportunities.
Identifying those demands and opportunities is tricky, but by running a few A/B tests you can steer yourself in the right direction. Here are a few examples of the experiments you might run:
From top-earning mobile gaming apps to influential e-learning apps, ever increasing numbers of digital products and services are being offered first as a basic or trial product, free-of-charge. It’s only when the customer is hooked that the question of payment is raised.
Design feedback experts Unveil use exactly this kind of model – they offer their product for free over a trial period, after which the customer may continue using it by making monthly payments. Unveil ran an A/B test to investigate how removing any references to these payments from their homepage and focusing on their free trial offer would affect conversions:
The results of the experiment were conclusive, with the new version yielding 31% more sign-ups than the control.
Should your website mention pricing?
The answer to this question depends entirely upon the type of products or services you’re selling.
If your site is an online shop, you should almost certainly be listing a price next to each item in Shop view and on product pages. That’s simple enough, but what’s less clear is whether listing prices next to featured products on your homepage will help or hinder your conversion rate? A/B testing will give you a good indication of the answers to this and other questions of a similar nature.
For service providers, the issue is more complex. Some prospective customers may be discouraged if they see your prices before they are fully convinced they want to work with you; others may be frustrated if they have to struggle to find information on your rates.
If you are selling a single, digital product or service and offer a free trial period or basic version, Unveil’s successful A/B test would suggest that foregrounding your free offer and banishing mentions of pricing to deeper pages may help to increase engagement.
Which of these sign-up form designs achieved the best conversion rate?
You may be surprised to learn that the victor – by a not-insubstantial 12.6% – was the second version; the version without the TRUSTe security badge.
Earlier we discussed the importance of establishing trust – but once the customer has already decided they trust the brand enough to consider signing up, they probably don’t need to see a security badge when they reach the sign-up form. The badge will simply clutter the design and complicate the user’s thought process – they may even be prompted to question whether or not they really do trust the company.
This seems clear enough with the benefit of hindsight, but at the outset it would be a very difficult issue to spot. Meticulous A/B testing of the key features in your conversion funnel can often unearth surprising results that point to CRO opportunities where you would least expect to find them.
Some extra split testing tips: