Opinion: Beta is Better
The most effective web stores are works in progress, using real customer behaviour to drive small, iterative improvements on an ongoing basis. This approach not only removes the guesswork, but also the pressure of trying to be perfect first time, says Mark Simpson, Founder and MD of Maxymiser.
A recession is no time to be taking a stab in the dark with promotional activities of any kind, yet, with the exception of some forward-looking companies, most businesses are still taking a very hit-and-miss approach to the design and navigation of their web sites – arguably the most powerful shop window they have. The same is true of the offers that are being displayed.
In a bricks-and-mortar store, the layout of the aisles and displays will have been carefully designed to steer customers to get the most from – and buy the most during – their shopping experience. The high-street retail experience has been researched in great detail, to maximise this, so that, not only does the customer fill up their basket, they will keep coming back for more. This means they need to enjoy the shopping experience, and be able to find what they need easily.
On the web, all of this applies equally. So why is it, then, that web sites are still being designed backwards – without any feedback from customers until it’s too late?
A sub-standard shopping experience
E-commerce has been a booming business for a long time now, yet the majority of online store experiences remain dire, and the very opposite of intuitive – from failing to include search facilities, to wiping clean forms if the customer goes back to a previous page, and numerous shambolic omissions in between. The result is that customers are still being bounced around poorly put-together web sites in such a way that they are left dizzy and frustrated by the time they shut down their PC, with their basket still empty.
Given that many people now shop exclusively online, this lack of customer-centricity and quality control is shocking.
It doesn’t matter how ‘expert’ the web designer. If they are not monitoring customers and tracking their preferences as they move around a web site, they are missing the point.
As noted, bricks-and-mortar stores invest heavily in testing the layouts of their stores, so customers are guided through the premises in a logical and easy way, while being lured towards additional extras along the way.
Test as you go
But one thing that retail store designers can’t do, which web developers can, is test out thousands of slightly different variations in a live environment to see which yields the best results. It is also very easy and affordable to do – and has been shown to represent much greater value for money than search optimisation.
‘Best practice’ is old hat, certainly in web terms. Best practice reverts to the lowest common denominator. It’s generic and makes assumptions – assumptions that have typically been agreed by ‘experts’ in a closed room, based on past experience and the input of a few random focus groups.
Wouldn’t it be a lot more effective to see what happens when real customers are trying to navigate the site in a live environment?
Multivariate testing allows exactly that – enabling designers to be as creative and innovative as they want to be and then, crucially, seeing if it works, by gradually feeding out test pages to designated groups of real customers.
Amazon, the master
Amazon does this all the time, testing every tweak to its page layouts and content on its customers without them being aware of it. Different page variations are piloted with small customer samples and the impact on conversions is monitored, enabling each tweak to be progressed or rejected accordingly. The result is a continually refined web experience for the customer, and maximised sales for Amazon and its partners.
What’s exciting about live, multivariate testing is just how well it works. The smallest and most subtle change can have a surprisingly significant impact on conversions and customer stickiness. One of the most astounding success stories saw a leading news and careers website gain a 1,000% uplift in completed registrations simply by changing the structure and layout of its forms. The content itself remained the same.
Subtle text changes deep in the checkout process for National Express, meanwhile, boosted the flow through that page by 14%.
The reason it’s so important to let customers help refine web pages through live testing is that the factors which make the biggest difference can be small and subtle – something a marketer would almost certainly miss.
Marketers miss the point
No example illustrates this more clearly than that of DIY supplies chain, Wickes. Five different designs of a proposed sign-in page were shown to a sample of around 500 marketers. They were asked which one customers were likely to prefer. When the responses were compared with the results from live customer testing, it was discovered that only 4.6% of the marketing experts had guessed right!
Just flipping the positioning of the company’s log-in and ‘register’ boxes was shown to have a 10% impact on conversion rates.
This is the kind of thing that can’t be predicted. Just as, when Laura Ashley was implementing improvements to its web site, testing showed that its ‘continue shopping’ button was too prominent, potentially deterring some visitors from progressing to complete transactions. The company responded by adjusting the size of the button to achieve optimum results. Mass personalisation
Every customer is different, too. In a bricks-and-mortar retail environment, this necessitates bowing to the majority. In a web environment, these restrictions need not apply. Indeed, personalisation is now huge on the web. Refine the experience in favour of individual customer preferences, and not only is the customer more likely to return, they will buy more – because their time on the web site is now focused and productive.
Amazon, once again, is probably the best example of this, with its personal recommendations, ‘people who bought this also bought this’ suggestions, favourites, wish-lists, and one-click ordering.
Leading-edge web sites recognise this and have honed web personalisation to such a degree that they are able to present a different web experience to different groups of customers, or better still to the individual. Beta is better
Beta is better
This flies in the face of the one-size-fits-all, ‘best guesstimate’ web site, which has been polished and perfected before it is launched to the public.
What a risk these companies are taking, with their flashy glorified marketing brochures! The most effective web sites are works in progress, which use ongoing, iterative changes to continually hone the customer experience.
Being innovative needn’t mean making radical or costly changes. It’s about paying more attention to the detail of how customers navigate and behave on your web site, and then responding with enhancements that can be delivered quickly – without having to wait for the next development cycle.