My Role: UX Researcher and Product Designer

Project Duration: Two Months (2019)

Client: [Confidential]

Methods: Competitive Analysis, Conceptual Model, Contextual Inquiry, User Interviews, Mental Model Diagrams, Experience Maps

Output: Findings and Recommendations Report, Thermostat Design

Synopsis: A client approached Balance with an interest in creating the easiest possible installation experience for a smart thermostat. We took a human-centered approach to understand what issues people experienced when installing a smart thermostat, and how added, subtracted, or modified components might help solve for those issues. Methods included 4 contextual inquiries with homeowners with high/low DIY skills. The result was a smart thermostat designed to make installation easy.


Installing a Thermostat is Intimidating

Our client had been in the IoT space for a while and had found installation to be a major barrier to entry. Paying for an electrician is expensive and there are technical and educational barriers that could be overcome through better technology. Inexperienced homeowners are intimidated by playing with their home's wiring, with fear of damaging their home or themselves causing them to 'satisfice' with subpar HVAC experiences.


Competitive Analysis

My colleague installed seven competitive thermostats which we analyzed. We used this to inform a competitive matrix. We took notes on every stage of the installation process, included tooling, and instructional materials. This helped us get a baseline understanding of the requirements to install a smart thermostat, and the nuances between the different models.


Conceptual Model 

Armed with the understanding from the competitive analysis, I designed a conceptual model of the installation experience using Omni-Graffle. This model would be used to help understand the installation process at a high-level, and assisted us in coming up with questions that would need to be answered through our user research. It was a useful tool to return back to throughout this entire project.


Contextual Inquiry with User Interviews

Myself and two colleagues ran four contextual inquiries with four homeowners. Due to budget constraints, we selected four friends and family members who were willing to participate. I started by writing a strategy document which included an outline of goals, objectives, and research questions. This document informed the facilitation guide. Each contextual inquiry lasted approximately an hour and a half, and included all stages from unboxing the system to pairing the system with it's mobile application. 

Data from each interview was combed and coded in Google sheets to help inform the project's artifacts.


Experience Maps

Experience maps were made to outline each interviewee's installation. The maps included the stages of the installation process along with discreet "thinks", "feels", and "dos" based upon the data gathered for each participant. This helped to empathize with each participant, while providing insight to high's and low's experienced throughout the installation. 


Mental Model Diagram

Using the data from the user research, I created a mental model. This honestly was not very useful because the experience map simplified a lot of the represented data.


A final report was created to outline process and recommendations for the final product. Results of the final product are currently confidential and will be shown after the product is formally released.



Doing and mental model diagram and experience maps is redundant. Where I originally thought creating both of these artifacts would help conceptualize the installation experience and help develop opportunities, I found that there was too much overlap between the artifacts. It would have been more worthwhile to do an experience map and formative usability testing. Time-boxing the installation experience might have helped to establish a baseline for task completion times, which would help create a benchmark for summative evaluation in the future.

That being said, the mental model diagram helped to identify dozens of opportunities which made it into the final report. It's just that the overlap between the experience maps and the mental model diagrams made things a little redundant.