A smart home automation system that makes your home a cozy place

Istabai develops sensors that extend your home capabilities, managed from the palm of your hand.

The Role

I was part of a shared design resource in a startup incubator where this hardware startup was founded.

I got to research and design several new product ideas as well as refreshes for existing products. To achieve this I needed to switch between several roles.

Product Designer

Early prototypes were a key part of building a shared understanding with the team and showing our thinking to the CEO. They helped to visualise the spec document and resolve almost all disagreements with the direction.

UX Researcher

We couldn't afford proper user testing so I went with Guerrilla User Testing to get some conversation with potential users and test the prototype. A conversation with any user is better than no conversation at all.

The Approach

Discovery

Gathering business objectives and technical requirements to see what features to plan

Pen & Paper sketching to get a feel of the temperature slider feature

Organising discovery workshops to learn about the product space and book development resources.

Research

With Competitive Research, we found that we can stand out with better design.

In Phone interviews, we found that people like to pre-heat their homes while in the car coming back from work.

A mobile-first approach so it also works for the web portal.

I tested the app by interviewing people at work and on the street.

Analysis

In our research, we found:

  • The prototype validated our design decisions as users could navigate round successfully

  • Most participants enjoyed using the product

  • Seeing all your devices upon opening the app, gives the user a sense of calm and feeling in control

Design

The goal of the design was to facilitate calmness in the user and to make them feel in charge.

User tests did not find critical usability issues which allowed a quick design path from prototype to code.

Taking this in mind, the goal for the interface was to give the user as much feeling of calm as possible. They needed to feel that they are in charge. After all, it is their home.

Challenges

Technology limitations

Due to a lack of resources were were limited to web technologies instead of native apps. This limited what features were feasible and since we needed a mobile first approach, we went with a webview container to be embedded in native apps.

We had a bare-bones backend and because of its limitations several value adding features were not supported at that time.

Defining user types

We know that the product is mostly used by:

  • Private individuals in their homes with just a few sensors

  • and businesses with thousands of sensors.

What we did not know was HOW people were using the product. All kinds of edge cases were found but mainly users did things like:

  • Just set it up and forget about it

  • Tinker a lot with the app to adjust heating and presets

  • Companies use sensors as a health check for other hardware.

Design

Everything at a glance

The idea was to create a calming experience with a sense of control.

  • Useful real-time data from sensors is always visible

  • All sensors in your home in one place

  • Soft card design and happy colours to evoke sense of calm

  • Toggling eco mode is easy even in a car

Calibrate your climate

Half-degree adjustments for temperature provided needed precision which corporate clients appreciated.

This was greatly appreciated because most competitors at the time did not offer such a degree of precision.

A Room’s view also includes a Scheduler to set automation in action and Statistics to see how everything is going.

Heat fluctuation report

The set temperature is compared to the actual in set intervals to give a continuous graph of heat fluctuation in the room.

If there are sharp fluctuations in the graph, it might mean a potential issue with the boiler that can be addressed now before causing issues.

Historical data is visible by scrolling left and right.

Automate your heating

This was one of the more challenging features because of the flexibility users had.

Examples of features:

  • Different schedules for each day of the week

  • Different schedules for each room of the house

  • Each schedule can have up to 20 different settings based on the time of the day

  • Each schedule can be overwritten by any other schedule or eco-mode

  • Dealing with schedule conflicts and when to resume the overwritten schedule

We wanted to include the ability to regulate all sensors and overwrite all schedules at once but this was deemed too complex for mobile and was moved to a special place on the web dashboard.

Temperature Presets

Each schedule consists of presets. The idea is that the user sets their default temperatures and builds schedules for different scenarios.

The most common use case was, as shown in this screen, with 4 settings:

  • Comfortable temperature — when waking up so it’s easier to get out of bed and do your morning routine

  • Economical temperature — it lets the house cool down a bit when no one is home and saves on heating costs

  • Comfortable temperature — when people are at home

  • Night temperature — Most often Eco and Comfort were chosen to have a lower temperature for sleeping

Each setting will end when the new setting takes effect.

Adjusting a schedule

To set a mode was moved to the bottom part of the screen for easier single hand use.

Navigate the hours by swiping on them and select which preset to switch to at the time.

It was necessary to account for the time needed for a boiler heat up and cool down. Normally, it is about takes about 30 minutes so this is the increment size we chose. But it can depend on aspects like:

  • Outside temperature

  • Difference between temperature settings

  • Boiler capabilities

  • Heat isolation in the house

Everything affects the energy necessary to heat up or cool down a space.

Device management

At the beginning, users were able to switch between rooms using a dropdown the same way a user can use to switch between houses to give a little more isolation between controls. Given that the most common use case was just 1 or 2 sensors in 2 to 5 rooms, it made sense to leave more granular settings necessary for larger clients to be taken to the web dashboard.

In this view user can see if a sensor is running out of power or has any issues with connectivity to the base station.

When going into a view of the sensor, the user is presented with basic information about the sensor, its status and the ability to disable it.

Final thoughts

The incubator was an amazing experience because of the fast-paced, product focus and startup culture.

Being able to design not only this but other products taught me a lot about the product lifecycle, product market fit and how to treat people in an agency type of environment.

It gave me the freedom to use a User Centric approach but also find and develop other skills like project management keeping projects very interesting.