Database Reference
In-Depth Information
Co-Discovery (Two-User) Testing
Usability testing with two users at once is not, strictly speaking, a paper prototyping technique, but most
of the tests I've conducted have been co-discovery because it seems to work quite naturally. In a co-
discovery session, the two users work on the tasks together, discussing them with each other (and the
facilitator) as they go.
Benefits
There are several reasons why co-discovery can be useful:
More comfortable for the users.
A solo user who gets stuck might blame himself or herself, but a
pair of users who are stuck realize, "Okay, obviously this isn't just me." Especially when you have
in-room observers, having users work in pairs can help them relax.
Easier for the facilitator.
Contrary to what you might think, facilitating a test with two users is
often easier than working with one. In co-discovery, users talk to each other as they work through
the tasks, and the facilitator may end up saying relatively little. This is fine. With only one user, the
facilitator usually needs to take a more active role and be more conscious of providing positive
feedback.
More data.
With two users, you're getting two people's thoughts and reactions at the same time.
Note that you're not necessarily getting two
independent
viewpoints—User B may quickly agree
with what User A said, making it hard to determine what User B really thinks. But if both users
provide a plausible rationale for why they agree, that can give you more confidence about the need
to change something. And when users disagree, then you usually do have two valid data points.
Although this isn't very scientific, I tend to split the difference and view co-discovery as providing
about 1.5 times the data of a single-user test.
Scheduling.
In several hundred usability tests, I estimate that I've experienced an average no-
show rate of about 1 in 10 users. If you've scheduled two users and only one shows up, you can
still hold the test. That's useful if you're testing on a tight schedule or have observers traveling to
attend the sessions. Double no-shows are very rare—perhaps 1 in 100 tests—so with co-discovery
there is little chance of having to cancel a test.
Drawbacks of Co-Discovery
Although it has its advantages, there are also some drawbacks to co-discovery:
Discrepancy in experience or confidence.
If one user knows substantially more than the other
(or thinks so anyway), it can be awkward for the other person. Ideally, you'd avoid this during the
recruitment process by not pairing people with a mismatch in experience, but this makes
recruitment more complicated and it's hard to do perfectly.
Peer pressure.
Some participants may be more concerned about looking foolish in front of
someone else in their profession than they are in a roomful of strangers that they'll never see
again.
Slower pace.
Sometimes two users will take longer to complete tasks because there's more
discussion. This drawback balances the "more data points" benefit—it's a trade-off.
Dominant personality.
Most people are pretty good about sharing and will work out for
themselves who does the clicking and typing. Some pairs will automatically switch roles after a task
or two. But a few people don't get good marks in "plays well with others" and tend to dominate the
Search WWH ::
Custom Search