ROI - Page 3
April 15, 2002
The Return on Investment (ROI) for usability testing is very
high. The Usability Company quotes an average figure of 800% over
a twelve month period, meaning that companies can expect an
increase in revenue equal to eight times the cost of the usability
testing.
Clearly, if a site has no decent revenue streams to begin with,
usability testing isn't going to make the situation much better,
but financial service and ecommerce sites usually find that
usability testing is a very profitable exercise.
A practical example was Smile-on, the UK's leading online provider
for dental practices. The company noted some customers had
difficulty using its online E-store. They commissioned a single
usability evaluation with eight users, concentrating on tasks
associated with purchasing through the E-store, starting with
registration and working all the way through search and the
shopping basket system to completion of order. The evaluation
identified 46 issues and made 60 recommendations.
Smile-on subsequently relaunched itself and its site. Monthly
orders improved by over 100%. The number of new clients increased
by over 240%. Some of this improvement was down to factors other
than improved usability, but the vastly improved conversion rate
showed how well the usability study had worked. The return on
investment (within one year) for the usability project was
estimated at better than 10:1.
Sites outside the finance and pure e-commerce sectors have also
benefited - mtv.co.uk gained a 30% increase in core traffic after
usability testing.
New Sites - Greenfields
The more switched on companies in the world are beginning to
recognise that usability must be incorporated in any new site
right from the beginning, and this includes usability testing of
prototype sites. The quality of a Web site begins with the process
of how it is originally created, and for top sites usability
testing is an essential part of that process.
Getting usability incorporated into a site at the development
stage is estimated to be even more cost-effective than adding it
retrospectively to existing sites. Apart from increasing take-up
of the launched site, it is claimed to reduce overall development
costs by cutting down on redevelopment and costly corrections.
For new sites, there's likely to be more flexibility in the
testing process. For example, with existing live sites it's usual
to stick with the same tasks and not change the site during the
test period. With a Greenfield, the developer may change the site
between tests, and the tasks might be adapted too, real time, to
get the best value from testing. In a real example, a task that
took 30 minutes at the beginning of the day was reduced to just
ten minutes by close of play as the site evolved.
Other kinds of user testing for prototype sites include card
sorting, which is used to identify useful section headings for a
new site. If testers don't see the headings as intuitive, they
may make their own suggestions for better section names. This
appears to work well whether the intended users are highly
educated or not.
Another variation for greenfield site testing is user group
testing, with one person navigating and the screen projected on
to a wall for other users to comment on. This is useful for early
prototype sites, or when a client has a selection of alternative
designs to try out.
Other Usability Testing
A further type of testing, mainly for live sites, is competition
benchmarking - again as heuristic evaluation or alternatively as
user testing. This involves testing a site against its main
competitors to find their relative strengths and weaknesses.
The Usability Company also carries out usability testing for
interactive TV and wireless applications. It may appear that
the company has generously given away all its commercial secrets
in this article, but of course that isn't the case. While it has
allowed us far greater insight into the usability testing process
than any Web publication has been allowed before, there are two
crucial aspects of testing that can't be copied with ease from
this short article. One is the formation of user tasks, and the
other is correctly interpreting user performance on these tasks.
A little mystery remains.
Additional Resources:
International Usability Engineering
Why You Need to Test your Site with Real Users
Why You Only Need to Test Your Site With 5 Users
WDVL design section, including usability tutorials
The Tasks
Usability Testing in Practice
|
