Re: [Tails-dev] results for usability tests of first-time us…

Delete this message

Reply to this message
Autore: sajolida
Data:  
To: The Tails public development discussion list
Oggetto: Re: [Tails-dev] results for usability tests of first-time use (#18074)
Zen Fu:
> sajolida <sajolida@???> writes:
>> We hadn't tested this entire flow since 2015 and so many things happened
>> since then.
>
> That sounds to me as a too long interval for such important work, right?
> I wonder if you or someone has an evaluation re. the amount of UX work
> needed vs the amount of resources/focus allocated to it, past and
> future.


It's good that you're making me reflect on this :)

The biggest issue that I discovered during the tests is that people
missed following the rest of the installation instructions after
downloading.

I think that this comes from 2 main changes done since 2015:

- (Mostly), the split of the download (and verification) page as a
separate page from the rest of the installation steps, done in 2017.

I did this split when we migrated the verification extension to
WebExtension. I tracked this to the download of the image itself not
being integrated in the verification process anymore:

https://lists.autistici.org/message/20170609.175300.2fbfc35e.en.html

We thought that having 2 different pages would help people verify
between downloading and moving on with the rest of the installation.

I did some usability tests on paper prototype of the new download
page, but I skipped doing usability tests of the entire process again
on software after the release by lack of budget and time. It was also
the first time that I was testing things on paper prototypes.

Testing only a paper prototype mislead me because they sometimes
influence too much what people pay attention to, compared to being on
an actual web page.

- (Also), the addition of the warnings at the bottom of the download
page, done in 2020.

They made the page harder to navigate by hiding the link to the rest
of the installations steps in the middle of the download page instead
of at the bottom.

We clearly didn't pay enough attention to how this change could impact
the UX of the entire page. The good news is that this regression was
only introduced 6 months ago :)

Regarding what I could have done differently since 2015 and lessons
learned, we should:

- Always test on software on top of testing on paper prototypes.

If we can't test on software right away, we should schedule tests
without waiting for years.

- Consider testing again our critical paths if *anything* changed.

Seemingly unrelated doc updates can break a user journey.

- Work on outsourcing usability tests.

Testing in-house have a lot of value but, in this case, it would have
been better to test our installation instructions again in 2018-2019
than to wait until 2021. In this case, testing the installation
instructions is also probably one of the simplest tests to outsource.

In 2017, we didn't have budget for that but now we could. In December,
I started discussing with external contractors to outsource usability
tests of our donation process. It will good to try.

Regarding the short term plans, we'll solve the biggest issues with
#16808, hopefully in January.

--
sajolida
Tails — https://tails.boum.org/
UX · Fundraising · Technical Writing