Confused much? Here’s what you need to know about optimisation and A/B testing
Us digital types live and breathe marketing technology to the extent that we sometimes fail to realise a lot of the lingo we speak is complete gobbledegook to other business owners.
We are so used to waxing lyrical about algorithms, click-throughs, PPCs and lead funnels that we forget these are terms which the rest of the world isn’t aware of.
First up – sorry if I’ve ever confused you with lingo and please pull me up if I ever go too deep into a subject without giving a proper explanation!
Second, let’s break down a term you’re bound to come across if you take the brave step of moving your business into the digital realm as part of a plan to cut costs and improve profits.
A/B testing is also known as split testing.
What is A/B testing?
A/B testing is also known as split testing.
According to a website called Optimizely, “A/B testing (also known as split testing or bucket testing) is a method of comparing two versions of a webpage or app against each other to determine which one performs better. A/B testing is essentially an experiment where two or more variants of a page are shown to users at random, and statistical analysis is used to determine which variation performs better for a given conversion goal.”
Get it? Perhaps not.
Let’s make it clearer.
You have two buckets of apples, one full of Pink Ladys and the other Granny Smiths. A lineup of hungry school children awaits (this is your target audience). You have enough apples for every kid to be happy, no matter what their choice
Line the kids up and get their pick of apples. Do they pick red or green? The apples which are selected the most are the ‘winners’ in these circumstances.
Given the choice of Pink / A or Green / B, eighty per cent of the kids grab a Pink apple, so it is a clear winner.
This is great. Good for you, Pink.
But now let’s try again. Scrap the Granny Smiths apples (who likes them anyway?) and offer a Red Delicious in its place. Now you have another A/B test but it is based on the assumption that children prefer sweeter, rose-hued apples.
You will probably find the percentage gap narrows. Perhaps the Pink Ladys are still popular but it turns out that given a choice of red vs red, 60 per cent of kids will go with a Red Delicious.
This is a very basic explanation of A/B testing but should help you get the drift. A test like this would show a school which apples they should order for the canteen so they have less waste.
The difference with A/B testing in the digital realm is you create two options and serve each one to your target audience at random.
For example, 50 per cent of your digital audience see a blue BUY NOW button, and 50 per cent see a red BUY NOW button. Track which button gets the most clicks over a two week period to see which colour is more attractive. From there, you could experiment by changing the shade or changing the placement of the button on the screen. Perhaps a bright red BUY NOW button which sits at the top left of the page is the ideal positioning.
Why do A/B testing?
A/B testing gives you actual data to work from. While design experts (known as UX / user experience – another term to get used to) have an idea of best practice, they aren’t relying on an exact science.
What’s more, you can gather data from user surveys but quite often a survey responder will give you an answer they think you want to hear, or one which will work in their favour (did you ever see an employee survey where respondents were happy with their current salary?).
As a result of your A/B testing, your website can be optimised for the best results from your customers based on their actual behaviour. This means you know from physical evidence what actions they take on your website and what they respond best to.
For most businesses, it is ideal to engage in regular A/B testing. Audience preferences and requirements evolve over time so what they want now may not be their ideal in two years time.
[Small changes can make big differences Photo by freestocks.org on Unsplash]
This method of online testing has the huge benefit of improving your conversions, your engagement and the amount of money a customer spends in a single transaction. Some companies find one simple change can deliver huge profit increases. Others compound a number of small changes for improved results.
How to do A/B testing
- The question
To get started with A/B testing, think of some of the big questions and areas of mystery for your business.
For example, Why are so many customers filling a cart but abandoning it? Or How can I get more opens on my newsletter? You may want to know the category of products which, when highlighted, are mostly likely to get a customer to make an additional purchase.
Tie your questions to key business outcomes, e.g. You want more sales or you want more enquiries.
2. The predicted answer
You also have to have a hypothesis. Something like: “If we change sales newsletter subjects to say “LIMITED ITEMS LEFT”, more people will open them.” Or “If I add a pop-up offering a ten per cent discount which springs up when a buyer tries to leave my site, they will stay and make a purchase.”
Your prediction could even be “If I change the colour of the ‘Contact Us’ button, more people will go ahead and do it.”
3. The test
Now you have to test your theory. There are several tools which will help you implement A/B testing and to track user behaviour and therefore the results.
You will have to work closely with your web developer, marketing team and designers to create the different options for people to be presented with.
Remember, your website users aren’t making a conscious choice based on A or B. They’re being given one option and taking action based on what they see.
Keep your tests simple so there aren’t too many elements and behaviours you are trying to measure.
[A subtle difference can have a large impact Photo by Jeroen den Otter on Unsplash]
Once your ‘experiment’ is live, give it a few weeks so you can get some consistent and ample data to work with.
4. The real answer
And finally, it is time to analyse your results. You may find that there’s a clear ‘winner’, or you may not notice a measurable difference between options. For example, some companies experiment with sending newsletters at different times of the day or different days of the week and fail to witness an impact on open rates.
If you do get a clear indication towards A or B, you can scrap the less effective option. However, be aware of potential flow-on effects. For example, that popup may improve sales results but you could then notice a drop-off in repeat visitors to the site.
Once you have completed your initial A/B test, you can make further tweaks, similar to our red/green, red/red apple example. There are always potential ways to improve your results and improve usability and customer satisfaction.
[Compare like with like to see what your customers prefer Photo by Benjamin Wong on Unsplash ]
Different areas to apply A/B testing to your website and marketing
Some basic A/B tests to apply to your website include the font, the text size and the spacing, plus the headlines and CTAs (call to action) that you use. There are some tools which also allow you to add C and D to your A/B test.
Colours are another simple way thing to test which could have surprising results and make a difference to customer actions.
The location of ‘buy now’, ‘call us’ and ‘make a booking’ is another element of your website you can optimise through A/B testing for excellent results. Move your button up, down, left and right, and play around with colour and size.
You can test the wording, images and calls to action on Facebook and Google Ads. Perhaps a more motivating “Click to buy” is less effective than “Find out more”. A picture of a frowning face may beat a smiling one, and a frowning person holding an empty wallet may beat a frowning face on its own.
Similarly, the subject line you use in an email can make a world of difference. Does adding the client’s first name pique their interest? Perhaps you’ll notice better open rates if you ask a question in the preview.
You can use A/B testing to see if including videos make a difference to conversions from your landing page or if testimonials make people more inclined to buy.
Special offers are another item you can A/B test. Give half your customers the excitement of ten per cent off and the rest the option for a bonus gift with purchase.
Even the overall design of your homepage could be split tested. Do people prefer a lot of information or does a single image, opening line and call to action spur them to keep clicking?
So many questions! But remember to keep your business objectives in mind. You don’t have to engage in A/B testing just for the sake of it.
I hope this has clarified the purpose and benefits of A/B testing for you. While you may have a good understanding of what it is now, the fact is you’ll still need a hand to get it up and running! Working with a digital expert will help you conduct efficient A/B testing and they can then show you how to optimise your website and online communications based on the results.