17/03/2015

By Nick Keating, VP EMEA, Maxymiser


From retailers to gaming companies, travel firms to banks, online failure is now headline news. Whatever the cause, from the repeated outages at RBS that left customers high and dry to the 8.1% drop in sales attributed to the new Marks & Spencer website, a bad online experience affects not just individual customers but brand reputation and shareholder value.

So why are companies taking such huge risks with current ecommerce re-platforming activities? When companies have spent years honing the existing site, improving content and journeys to optimise every aspect of the customer experience, it makes no sense to risk turning on a new site without testing any aspect of the user experience at all.

Whether the revamp is reskinning the old site, changing the back-end ecommerce engine or swapping in a new content management system, the risks associated with delivering a poor user experience cannot be ignored. As Nick Keating, VP EMEA, Maxymiser explains, when it comes to re-platforming, risk mitigation is key – and that means measuring, testing and optimising as much as possible before, not just after, the event.

Online Disasters

When Marks & Spencer (M&S) announced that the 8.1% drop in sales and resulting share price dip was due to the launch of its new website, Ecommerce Directors and UX teams across the world took a collective deep breath. But M&S is far from alone. The catalogue of disasters – or near disasters – that litters recent website development history makes sorry reading.

Consider the broadcaster that undertook a six month investment with a third party agency to create a compelling rebranding of its back catalogue, only to find the new service garnered less response than the previous one. Or the retailer that spent months creating an entirely new checkout process, only to discover customer conversion dropped and the old process had to be switched back on.

The broadcaster was, to an extent, lucky. By testing the new back catalogue branding before full roll out the company was able to pull back from launching it and avoid both bad publicity and negative customer feedback. Similarly for the retailer, early testing of the new experience – and a willingness to accept that the new process wasn’t working – minimised the overall impact on both customers and revenue.

Flawed Thinking

But how did this situation even arise? A website increasingly underpins every aspect of the business – from sales to share price and brand value - and companies in every market are spending millions on capturing customer data in order to create that perfect, personal and relevant customer experience. It makes no sense to launch a new experience without understanding how it will impact the end user; or invest heavily in a new online platform and fail to include the learnings from the last platform.

For any retailer it takes around six to nine months to launch a new site, creating a lead time of a year if you take into account the Christmas lock down. No business can afford to spend this long on a development that is not going to be hugely successful. When the new platform is launched it is essential to have confidence in every aspect of the user experience and journey.

Optimisation has become a strategic component of incremental website development activity over the last few years, with companies continually using A/B and multivariate testing to check the response to new content, calls to action, even key business strategies. So why wait until a new website development project has been completed to ascertain the customer response to the new experience? Waiting until an entirely new checkout process has been designed leaves a company with few options if the worst happens and customers give a wholehearted thumbs down. It is so much more effective to apply the principles of website optimisation that have become standard in recent years and test each aspect of that process incrementally.

For example, break down the checkout process into different aspects and identify those that are different to the current process. Is it a journey change; a longer or shorter process? Is there more information; or is the entire process being streamlined and de-cluttered? There are so many aspects that can be individually identified and tested to validate the process. Are the stages as good as or better than the old? Is the branding better? New content copy more successful? The new security sign-in; or payment process? Unless every aspect of this process is being tested and optimised using A/B and multivariate testing, a company is simply delivering this vital new business channel based on a set of assumptions.

By breaking the entire checkout process down into a series of tests a company has far more control over the entire process and, critically, is not going to face the shock of an investment not working as planned when it is unveiled to the public.

Changing Attitude

Of course, many developers argue that it is impossible to test a new site when there is no traffic. True. But, as outlined above, it is more than possible to test aspects of the new content, processes and journeys on the existing site. Yet right now, when a new site is under development far too many companies actually stop optimising the existing site, arguing that it would be a wasted investment on a site reaching end of life. Many add that staff are required to focus all their attention on, firstly, developing the new site and secondly, firefighting once that site is live due to the level of problems expected.

This makes no business sense. How can a company decide to switch on a newly developed website – whether that is a new look and feel, responsive design or a upgraded ecommerce platform – with the expectation that it will perform badly and that there will be problems? Companies are admitting, in effect, that a business channel that is critical to revenue and brand value, is going to be served up to the customer base underdone.

Even if there are some aspects of the new design that cannot be tested on the old site, organisations need to be ready to test the new experience as soon as possible. This will identify problems fast and enable proactive remedial action, rather than ill-targeted firefighting. As the experiences of the retailer and broadcaster highlight above, a new experience can be bad enough to make it far better to revert to the original site. Either permanently or temporarily. Just consider the dire impact on the business if on top of failing to test the new platform, the original site has also been effectively ignored for the past six months. So much for the much vaunted customer experience.

Optimisation is not just about looking for a sales uplift – it is about mitigating risk and avoiding disaster. Building in testing and optimisation as part of the process radically reduces the risk associated with major website change and redevelopment. When a company will not so much as change a call to action button without testing it first, how on earth can any Ecommerce Director sign off an entire untested website? Fingers crossed anyone?