Reading the room before the redesign

The brief said the interface felt dated. That is almost never the real problem, and it is never a problem you can solve from a screenshot. So before opening a design file, we asked to stand behind the counter on two busy evenings.

Three things no screenshot shows#

1. 01
   Glare

   The counter faces a glass front. For roughly forty minutes each evening the screen is washed out, and the staff had learned to work from memory and muscle position. Our elegant low-contrast secondary text was, for that window, simply not there.

2. 02
   One hand

   The other hand is holding a card terminal, a printed ticket, or a phone with a booking reference on it. Every two-handed interaction in the interface was being performed as a sequence of one-handed ones, badly.

3. 03
   Interruption

   The average transaction was interrupted 1.4 times — a question from the queue, a colleague, a phone. The interface had no concept of a task you come back to, so staff had built one out of sticky notes.

What changed as a result#

• Contrast floor raised across the board. No secondary text below 4.5:1, even where the brand would have preferred it. The glare window decided that argument.
• Everything critical reachable in the lower-right third. Where a thumb is when the other hand is full.
• A resume state. Any transaction can be parked and picked back up by name, which replaced the sticky notes — and turned out to be the single most-used thing we shipped.
• Bigger targets on the two destructive actions, further apart, because both were being hit during interruptions.

The workshop we did not run#

We were offered a stakeholder workshop instead. It would have produced a prioritised list of features, all of them real, none of them the three things above — because the people in that room are not the people holding a card terminal at 19:40 on a Friday.

Research does not have to be expensive to be decisive. Two evenings, a notebook, no recording equipment, and permission to be in the way. The redesign that came out of it was smaller than the one in the brief and it tested better on every measure we had.