Making the move from bricks to clicks in the retail space can be tough for traditional retailers as they try to apply offline knowledge to online offerings. It's no use trying to replicate the shop window experience or create a website and hope that it will run itself.
To make sure a site works as hard as possible, retailers and brands need to understand their customers and how they react online. This requires the creation of a solid testing programme that allows online retailers to gain actionable insights into their customers' behaviour and capitalise on this. Having devised a number of testing programmes for clients such as Myla
, Matthew Williamson
, here are a few dos and don'ts for retailers looking to firm up their online offerings.
Don't spend a fortune on test programs
Installing a test program used to be an expensive undertaking but today most analytics software offers this either as an add-on or additional part of the package. Google's Website Optimizer
is free and easy to use. It works brilliantly with or without Google Analytics installed and it is easy to set up. If you know what you want to test and have all assets ready, you could be up and running within an hour. Every tool you use for testing will automatically calculate the number of visitors per variation that converted, display the correct variations and so on. All you need to do is provide assets and design follow-up tests.
Do set realistic goals
If your goal is to test how many clicks it takes to achieve a conversion, make sure that you don't set your targets too far away from each other. If you are multivariant testing your homepage then the goal page shouldn't be much further away than the product detail page. The further away the tested page is from the conversion page, the more likely it is that you get skewed results as there are too many drop off points in between.
Don't forget to test your information architecture and usability
Your site's information architecture is critical as it dictates the way the site elements are structured and the way you want to move users though the purchasing process. Users will not follow the planned route and will always behave erratically on your site. You need to test that your IA caters for this behaviour and offer the right information at the right time.
You also need to test site usability. An example would be the "delivery information" section on a product detail page. Good information architecture would suggest a link to the details and good usability would offer this information in the form of an overlay rather than a link that leads out of the cycle.
Also remember to test your blanket email marketing, the order confirmation email, customer support emails and so on.
Do make sure that your copy is working for you
Even without investing in high-end technology you can do simple things to improve your online proposition. You should ensure that simple buying psychology is applied to all copy on your site and that you know what grammatical structures to employ — and avoid.
Also, place copy effectively and apply the appropriate copy style. When copywriting stick to the core principles: never use a passive voice; start your copy with a question that the customer has to answer with a 'yes'.
As a simple rule of thumb, when writing a product description make sure the first paragraph appeals to the emotional, the second focuses on product features and the third points out the benefits.
Don't stop testing
Tweaking your website after an initial test may lead to a spike in customer engagement — but don't stop testing at this point. It is important to come up with a test plan and to evolve the tests you conduct. So start with variations and then test the winning one against further variations. Matthew Williamson, for example, started to test showing price versus no price on emails. We are further testing different copy and different subject lines to find the ideal combination for Matthew Williamson. Once we feel comfortable with the emails we will continue to test landing pages against variations of these emails.
Do make it someone's responsibility within your organisation to be in charge of testing
Someone within your organisation has to own this space. Who should it be? Marketing, IT or the intern? There is no right answer to this, but it is advisable to use someone who encompasses both marketing and IT skills and has the necessary level of authority to make changes happen quickly.
Do be prepared to embrace change
Be aware that a test may lead to adjustments to your site which some people in your business won't be happy with. While a swanky Flash intro may please your brand manager, your test results may show that your money would be better spent improving the information architecture and usability.
Don't think that this advice doesn't apply to you
No matter what the size of your business, you always need to be active in your approach to new technology and constantly learning. Matchesfashion.com tested 'home link' against 'no home link' in the main navigation. 'No home link' proved to increase their product detail page click throughs by about 20%, and this is the page where you really start to making money.
Retailers and brands need to be courageous about unifying their entire business proposition both off and online. If a customer has a bad experience surfing your website, it is likely to affect how they perceive your company in its entirety. So test, test and test again!• David Hefendehl is the ecommerce manager at independent creative digital agency Pod1, a 50 strong collective of creative, strategic and technical specialists passionate about delivering cutting edge creativity. The firm's clients include Kurt Geiger , Uniqlo, Burton, Jigsaw, Kenwood, Net a Porter and Anya Hindmarch.