How intelligent do we want our thermostat to be?

Nantes, 2012 - 2013


Build and test the first minimum viable product version of the Qivivo thermostat to learn what matters to home owners when selecting and using a smart thermostat.

  • Design lead & user researcher

  • Deliverables: Project plan, recruitment, web app wireframes, design specs for devs, participant screener, interview guides, research insights and recommendations.


Feedback from 100 French families used to develop and launch a commercial offer, the Smart Qivivo Thermostat.

Qivivo was acquired by COMAP in 2018.


Create a smart heating system that people can trust

In 2012 I moved to France to live with my current husband. I spoke school-level French. At most French companies, people speak... French. My resume didn't mean much to most people. But it did resonate with an engineer and aspiring entrepreneur from Nantes who had just visited the silicon valley on a cultural exchange trip, learning about IDEO and design thinking.

He told me he wanted to help people heat their homes in a better way with a smart system that reduces energy use and provides more comfort. But he didn’t know how people felt about this, nor how they would like to use an intelligent thermostat. I was looking for an opportunity to improve my French on a daily basis, and was eager to get involved in the design of an IoT device.

I was his first hire as “Director of User Experience”. (Anyone who has worked at a start-up or co-founded one knows that regardless of your title you’re pretty much a Swiss army knife.) My first job in France, my first at a start-up. I came up with the name Qivivo, and we went to IKEA to buy some desks and chairs for our small office in the center of Nantes.


1. Define research objectives

"What do we think we know?" and "What do we want to learn?" The first questions to ask for any user research project. In this case the main questions evolved around adoption of "intelligent" features, and how to explain and market these.

As there were also technical concepts and algorithms to be tested, we decided to run a long study during the winter month and equip research participants with a working prototype including the hardware "box" with presence sensor, and a web dashboard to control and follow the temperature.

2. Design prototype

I started my exploration by sketching out some key features. Then I created wireframes for the MVP and contracted a visual UI designer to bring these to life.

Rough sketches - exploring key features and screens

MVP Interface - section showing the heating program

3. Recruit research participants

After having created a participant screener, we recruited participants online via a web form.

Research participant survey - Web form created to select participants

4. Visit research participants at their homes

Interview activity - create your ideal heating system dashboard

Research observation - the "smart home" of the future is also just a home...

5. Synthesise and present findings

Research synthesis - collaborative reach download session

Research findings presentation - screenshots of some slides

Mobile UI - design based on recommendations


Running a longitudinal research study with a working prototype is a fantastic way to learn more about real usage, not hypothetical use. However, the prototype needs to work. Well. The two times in my career that I've conducted this type of research, in practice there were a lot of hick-ups caused by bugs and non functioning features. As a researcher you become a type of help desk employee and trouble shooter, and it requires a lot of time to engage participants during a period of multiple months. I'd recommend starting companies to at least have a dedicated resource for ongoing communication with participants.