26 February 2010
I was recently asked to work on some designs for a registration and checkout process. Now, I’m a responsible and thorough person, so I set out with my copy of LittleSnapper and a handful of URLs to document some best practices. Here’s what I learned.
You’re not so different
My scientific selection criteria for this study started and ended with the requirement that someone, recently-ish, had recommended the site. As a result, I looked at the checkout processes of Amazon, Apple, Dell, Gap, Ocado, and Tesco. Those sites cross into a few different industries, but their processes are essentially the same; if we take out the question of account-creation, a high-level model would look something like this:
And there are a few lessons we can take from that.
Ask for what you need, when you need it
Almost without exception, the sites asked for registration at the point that a user clicked ‘Checkout’; before that point, users are free to browse and shop uninterrupted. The exceptions to the rule are the two supermarkets in the study – Tesco and Ocado – who need to know more about you before they let you shop.
If you have to ask, don’t expect an answer
All of these sites have recognised that people giving you money trumps people giving you demographic information. When they interrupt a user’s shopping process to create an account, they make it as quick and painless as possible, and keep it focussed on shopping; most sites just ask for a name and some contact details and do it all under the heading of billing information.
You’ll also notice that most sites handle the account creation process without mentioning registration; it’s simply a question of being an existing customer or a new customer. There’s a great case study from Jared Spool in Luke Wroblewski‘s Web Form Design about just how painful users find the idea of registration:
Without even knowing what was involved in registration, all the users who clicked on the button did so with a sense of despair. Many vocalized how the retailer only wanted their information to pester them with marketing messages they didn’t want.
Ocado is the only site here to ask for broader marketing information, and it has the good grace not to require a user to answer those questions. The form could use some design attention, but the sentiment is bearable; even better would be to put those questions into the account page and let users answer them later.
None of the sites asked the ridiculous questions that used to be on any self-respecting registration form: are you a farmer, fisher or builder.
Focus on selling
I was surprised to see that most sites used a basket page.
Personally, I hate these; I can’t remember ever finding anything useful in the cross-sell areas and I resent the interruption to my shopping. But they do give the sites a good opportunity to promote other products and, working on the assumption1 that most of the sites in the survey are attracting customers who are primarily buying one thing at a time, there’s no significant downside.
Once a user is in the checkout process, all this cross-selling drops away to put the focus squarely on completing the process. Amazon even kills its own navigation.
What to take away
If you are selling things, you only make money when someone buys them. If you’re going to force a user to create an account before they help you make money, make that process as quick and painless as possible. If, as these sites do, you combine it with your checkout process, the user only has to fill in a couple more fields at a point when they’re already in form-filling mode.
Whatever you do, don’t dial back the clock to 1999 and ask them to select their hobbies.
At this point, I am abandoning any claim to science. There is nothing backing this up beyond personal experience and a guess that this financial crisis has dampened the online-shopping impulse ↩