Air purifier automation feature   

Dyson - 2017

Dyson was rebooting their mobile app (Link) which monitors and controls all their connected appliances. It was to be written from scratch and branded with a new design system another team had been working on. Part of this rework was to redesign the scheduling functionality which enabled users to automate the operation of all their connected Dyson air purifiers.

My role

I worked with an industrial designer and a UI designer, a few software engineers as well as key stakeholders. My responsibility was to work with the industrial designer who conducted data research, prototype ideas, test and implement with a UI designer for the new branding. My contributions to the team and product involved

Storyboarding use case scenarios
Mapping personas
Wireframing user-flows
Brainstorming logic models with engineers 
Prototyping low fidelity interfaces

Presenting to stakeholders
Conducting guerilla testing
Planning and observing usability trials
Directing a UI designer

Design objectives

Understanding user behavioural patterns

First, we needed to understand what users were doing from data reports in order to understand usage patterns 

Air purifier average weekly schedule start time

Proportions of typical air purifier operation times

Listening to use case scenarios from users

Competitor research

I looked at what other companies interface looked like in order to what common user patterns were out there.

Defining new system logic

In order to create a new scheduler feature that works across all machine types, I came up with a new model that used events as a parent with the selected days as children. That would be easy to edit and satisfy all tasks users wanted to complete.

Benchmark I created that used all edge cases captured to design and test against

The new logic system I worked on with the technical product owner

Prototyping and testing user tasks & goals

I built dozens of prototypes, experimenting with various schedule entry methods, editing some of which that aren't included here for NDA purposes. Below are a few of the tasks users needed to complete. We tested these prototypes in a lab to validate:

Viewing events
Adding a new event
Editing an event
Deactivating an event

Deactivating the whole event
Deactivating the whole schedule
Indicate overlaps

These are two of many prototypes showing a set of presets a user can select to speed setup. It also shows two different entry methods. The user can then set up the time and machine preferences.

Removing a day from an event

Deleting an event

Disabling an a whole event

Disabling the whole schedule

Scrubbing the week schedule

Iterated, stripped back and shipped in time

After some time restraints and engineering resources, we had to cut the weekly view in order to make the build and ship it. After receiving the usability test results, I iterated and simplified the design. I worked with a visual designer to apply the newly built design language to my wireframes.

Selecting a preset

Adding an event

Viewing the schedule

Editing the event

Lessons learnt

A few things could have gone better. I could have spent less time researching and hypothesising and more time working with the engineers. Unfortunately, the company has not been structured this way creating a lot of waste work and major assumptions. In the end, we had achieved major goals that we set out, namely,

The outcome