A/B testing: Ensuring a long-term relationship between your website and customers

Imagine you could repeat a first date as many times as you like, playing that same night over and over again until you steal your date's heart. From the things you say to the clothes you wear, you could test every possibility and become the perfect partner.

Would your date prefer it when you dress more formally, or is this a casual thing? Will they be more impressed if you order red wine or prosecco? And, once you've discovered the formula for the perfect date, imagine if you could flip a switch and life would move forward, your budding love blossoming into a long-term relationship.

This might not be a reality just yet but the good news is, thanks to A/B testing, you can do all of the above to ensure a loving relationship between your website and your customers.

A/B testing is a simple way to test changes to your website against its current design and determine which ones produce positive results. It is a method to validate that any new design or change to an element on your webpage is improving your conversion rate before you make that change to your site code. As such, there's only one way your conversion rate can go with a little A/B testing: up. If only it was so easy for humans as it is for websites!

Website optimisation and A/B testing can often be seen as analogous to going to the gym – it's something every marketer knows he or she should be doing, but actually getting around to doing it can be a challenge. But new technologies are making the old excuses begin to ring hollow – legacy software, excessive cost and lack of training can no longer be blamed for not fully optimising and testing your company site.

If we compare visitors to your site with a boyfriend or girlfriend, you don't just want a meaningless and fleeting tryst (we hope). Instead, you should be looking for a long-term relationship with visitors to your site. That means not always looking at click-throughs as a metric of success. Instead, conversion into purchase or membership should be the target measurement – a true sign of commitment.

Take the case of the Guardian's dating site, Soulmates. The Guardian uses A/B testing across many of its webpages and properties. Its dating site found ads driving users to their site were performing well in terms of overall click-through, but the majority of these visitors were not converting into members. Using A/B testing, Soulmates decided to test a "people first" strategy on its landing pages, giving users more options and a wider variety of information. The variation landing pages still showed the individual profiles featured in the ad, but also displayed a few other similar profiles of the same age range and gender. The variation landing page, which included search functionality, user testimonials, and related profiles, improved conversions by more than 46 per cent.

Similarly, on the landing page, a site visitor's top matches do not change frequently, so there was the thought the site could look stale and lifeless. Soulmates wondered if changing the default from "Your matches" to "Newly joined members" would garner more engagement and increase click throughs to individual profiles. After running the test for just a few days, Soulmates saw a staggering 137 per cent increase in click-throughs from the "newly joined members" variation.

It's clear that website optimisation and A/B testing can really boost the attractiveness of your website to make it stand out from your competitors, especially if you're predicting spikes in traffic, in areas like online shopping ahead of Valentine's Day. We've put together a few carefully selected tips to help you make sure that this Valentine's Day your website will turn its visitors' heads, permanently.

Start small, go big, go small again

Always start with small iterative changes and concepts for website change. But don't be afraid to "Go Big" with a larger redesign and then "Go Small Again" to iteratively improve upon it.

Don't forget about targeting

Testing your site can involve more than just visual changes – think about displaying targeted content or layouts, for example. Targeting can be based on geo-location, cookies based on the user's account type, pages visited, where the person was referred from, the type of browser used, and much more.

It's only a test

You haven't made any permanent changes yet. It's OK if your test proves less successful than your existing version, as this will help you work out what works and what doesn't. Optimisation is achieved through trial and error, and no matter the outcome of a test, there is always something to learn and take away.

Get in shape

Sometimes with your website, less is more. A lot of successful tests involve removing materials like sidebars or excessive text to slim the site down – removing that additional information often improves user experience through simplicity.

Case study: Nationwide Vehicle Contracts

Nationwide Vehicle Contracts is one of the UK's leading providers of car leasing and van leasing for new cars and vans to businesses and private individuals. It tested its website without side banners and saw an uplift of 39.1% compared to the original, while a version including side banners but without other additional imagery from the original site performed even better - resulting in a 51% conversion rate increase. Interestingly, some of the results saw significant differences across form factors – the simplest designs saw a 62.4% increase when viewed on tablets. That has since led to further tests by NVC that could lead to even further personalisation based on device.

Keep it simple

In marketing circles, one particular bad habit is turning solid leads to smoke by confusing users with their call to action. A key to getting users to convert is reducing the amount of steps and information required, while satisfying the demands of your CRM.

Case study: Iron Mountain

Working with iProspect, Iron Mountain wanted to increase the quality of inbound leads acquired from the lead generation form on its website. Many site visitors were completing the form, but most submissions contained incomplete or irrelevant information. The storage and information management company clarified the form's purpose from "Contact us today" to "Request a quote" and significantly amended and simplified the fields in the form. The variation form resulted in a 140% improvement in lead quality, representing the highest lead quality Iron Mountain's sales team had ever seen - freeing up the sales team to focus on selling to highly qualified prospects, increasing revenue significantly across the board.

Be smarter with your money

Website optimisation and A/B testing shouldn't break the bank. New software such as Optimizely has democratised enterprise grade testing for businesses of any size – and for marketers of any skillset – no coding knowledge required!

So what are you waiting for? Your dream date isn't going to wait around forever, and neither are your customers.

Matt Althauser is European GM of Optimizely, the world's most widely-used do-it-yourself A/B testing tool.