So what is it, exactly, you're testing for? How many subjects should you recruit for testing, and what kind of platform will you be testing them on? Answer these questions correctly, and you can save yourself a lot of work down the road.
Now that we've established
it's time to get down to
what you'll be testing.
Your test plan will depend
on what you're developing.
You might be working with a
wireframe for a new website.
Or maybe you're developing an app for iOS.
Or maybe you're just creating a few
additional pages for an existing website.
Next, you're going to want to choose
the hardware platform for testing.
Is it going to be an
iPhone or a Chromebook,
a desktop with a 27-inch monitor
or an iPad mini? For websites,
we're going to want to test and use
a variety of platforms and browsers.
For apps, there should be fewer
variations to test. Obviously,
you want to concentrate on testing
devices that are most popular with your
users. Next, you want to select
your test subjects. Ideally,
you're looking for people who are both
likely to use your app or website and are
incredibly articulate. In the real world,
you might just get one or the other.
Sometimes the only criteria you need for
selecting your testers is demographics.
Say age, gender, income.
But you can always get closer to your
target audience by asking them some
screening questions. For example,
do you do most of the grocery
shopping in your household?
Or how often do you go out to the
movies? If your team is using personas,
try to select users that match as
closely as possible to those personas.
Once you have a good pool to select from,
you want to decide how
many testers to recruit.
Studies have shown that five participants
are usually enough to uncover 85% of
the usability problems on a website. So
if you recruit more than five testers,
you're likely facing
diminishing returns. But again,
it depends on what you're testing.
If you're testing for something as
open-ended as user satisfaction,
a large test group could be useful
in drawing out a range of responses.
Before running a full round of tests,
it's never a bad idea to do a "dry run".
Using a single tester running one subject
through your test will help you to
spot any errors in your
plan. So, in summary,
decide what type of product you're
testing. Choose the appropriate platform,
then recruit a pool of testers.
Then decide on the appropriate number
of testers to draw from that pool.
This number will be determined largely
by the type of test you're conducting.
And, finally, test it with a dry run.
If you take these steps in order,
you could save yourself and your team
a lot of unnecessary work. Next up,
building a plan.