Single-Page Checkout
14 Oct 2009 - 6:36am
5 replies
1485 reads
Paul Bryan
2008
In a meeting today, a client said they wanted a single-page checkout
process. The checkout tunnel is an area that I prefer to have as
vanilla as possible so that customers have no surprises, but I am
open to change.
I've never done any testing on one-page vs. staged checkout, and I
haven't been able to find any data, only opinions. Does anyone have
any non-proprietary data they'd care to share about one-page
checkout vs. multi-stage checkout in terms of abandonment analytics,
A/B or usability testing?
Thanks,
/pb
Paul Bryan
Usography (http://www.usography.com)
Blog: http://www.virtualfloorspace.com
Linked In: http://www.linkedin.com/in/uxexperts
Get the feed
Comments
Single-page checkout generally results in lower conversion:
http://econsultancy.com/blog/3095-gap-com-s-single-checkout-not-a-good-idea-report
http://www.getelastic.com/ecommerce-checkout-report/number-pages-ecommerce-checkout-process/
I think the real question is whether the customer wants visitors to
evaluate the entire form at once or just prevent a page refresh. The
latter issue can be addressed with AJAX and was used by Scott
Thomas' team to build an online voter registration system (extensive
multi-stage funnel).
On the reporting front, I'd look to event tracking. It can be
tedious to implement but the resulting data is more granular than
basic analytics. A/B testing is a fantastic idea but it could also
confuse returning customers the remember their past site experience.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Posted from the new ixda.org
http://www.ixda.org/discuss?post=46724
I have read reports that go in both directions. My personal
experience has favored certain implementations of single page
checkout.
I think the question cannot be simply "single page" vs. "multi
page" checkout. The total experience brings in a lot of variables.
I have implemented single-page checkout which consisted of
essentially an accordion of "steps" to complete within the
checkout. The result was a very responsive "wizard" of sorts,
which only revealed subsequent options as a result of previous
selections, which also is a way of utilizing progressive disclosure
and "wizard"-like behavior.
I'm sure the response to this will vary by audience, among other
things.
If at all possible, I would at least do an a/b split test between a
"multi-page" checkout and your "single-page" checkout.
A step further would be an experiment pitting several different
interaction designs to find the optimal scenario for your audience.
IMO, even a multi-page checkout might benefit from loading subsequent
pages via ajax, to increase responsiveness, and reduce delay, but
there again, a well-planned/executed pause might end up working in
your favor.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Posted from the new ixda.org
http://www.ixda.org/discuss?post=46724
Also, split testing should take into account the potential confusion
for customers, at both the design stage, and on the software layer.
Split Test Accelerator and Google Website Optimizer will load a
cookie that is used for the duration of the experiment, to ensure
returning visitors will see the same version as was seen in their
previous visit to the page.
Of course, by the end of the experiment, you will have chosen a
winning combination which will become the new standard version of the
page, so you will want to plan around releasing the resulting design,
perhaps in the same way you would release a new feature, depending on
how big the change is.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Posted from the new ixda.org
http://www.ixda.org/discuss?post=46724
Read: The number of "steps" perceived doesn't need to translate
into "pages" let alone entire page reloads.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Posted from the new ixda.org
http://www.ixda.org/discuss?post=46724
I feel a BLOG post coming on this one...
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Posted from the new ixda.org
http://www.ixda.org/discuss?post=46724