Smart Buildings · IoT
Commercial Smart Thermostat
At a glance: Joined late as solo designer with one week for the home screen and eight weeks for the full thermostat. Adapted to late-breaking hardware limits in real time and shipped a working prototype that leadership praised.
Small commercial building owners and their employees were stuck with outdated thermostats: functional but clunky, with complicated configuration and dated interfaces.
The device and its original third-party interface already existed. Our job was to rebuild the interface and functionality around our own product needs, designing within the realities of their platform.
Timeline was the biggest constraint: I joined late with one week to deliver home-screen designs, then eight weeks total to design the full product for a working prototype at a leadership conference.
To deliver and test the home-screen designs
1 wk
Home-screen concepts explored, then narrowed through testing
8
From kickoff to a working prototype at the sales conference
2 mo
Home-screen concepts, in a week
I aligned with the PM on the two user types (owners managing settings, employees making quick adjustments) and their core tasks: checking temp, adjusting setpoint, changing modes, viewing schedule status. Then I studied competitor thermostats and patterns on Mobbin and Dribbble to map the opportunities.
- Designed eight home-screen concepts, pulling components, colors and fonts from existing products so it felt like family.
- Matched the zone sensor's interaction patterns for cross-device consistency while keeping the thermostat's richer visual language.
- Teammates rated all eight on clarity, intuitiveness and high-tech feel, narrowing to four.
- Ran a dScout feedback session with eight business owners rating the four on the same criteria.
What we learned: users had clear preferences, but the highest-scoring concept for usability wasn't the highest-scoring for modern, high-tech feel.
What I changed: the PM and I chose the concept that scored highest on modern/high-tech, even though it wasn't the top usability pick. That aligned with the product's positioning and its differentiation in the market.
Designing the full product and adapting to hardware
After handing off the home screen, I moved to the rest of the thermostat. I analyzed the requirements, clicked through the existing product to understand its functionality, built a sitemap and reviewed it with the PM before designing the remaining screens and flows. In my first meeting with the third-party engineers, I asked whether there were design constraints to keep in mind. They said no.
What we learned: As I handed off designs week by week, constraints surfaced: the hardware couldn't support scrolling, gradients, transparency or more than six font styles. Engineering said they just didn't know what their platform couldn't do until they actually ran into the issues.
No scrolling → re-architected the IA
The biggest impact. Screens designed as single scrollable views were split into multiple discrete screens with navigation between them. I cut lower-priority content, moved secondary actions into submenus and added hierarchy to the sitemap. Settings and scheduling became nested flows with category screens acting as menus.
No gradients or transparency → a flatter visual style
I leaned on solid color blocks and borders instead of subtle depth cues, giving the interface a more segmented, intentional look.
Six font styles max → hierarchy through layout
With limited type variety, I created hierarchy with spacing and layout instead. A tight component system let me absorb each change quickly, adapting in real time without restarting from scratch.
A focused, hardware-ready system
The final design is a clean, segmented interface tuned to the hardware's limits: discrete screens with logical navigation, a flat visual language of solid blocks and borders, and hierarchy carried by spacing and layout. Owners get full control of schedules and settings; employees get fast, obvious access to temperature and modes. It mirrors the zone sensor so the two products feel like one family, while keeping the thermostat's richer, high-tech look.
Outcome & reflection
I delivered the complete thermostat design on time despite the aggressive timeline and late-breaking constraints. The engineering team built a working prototype on the new hardware, integrating every screen and flow, and brought it to the leadership conference just two months later.
Leadership praised the prototype. The modern, high-tech aesthetic landed as intended, differentiating it from outdated competitors while staying intuitive and easy to use. Reworking the IA around the no-scrolling constraint actually left the structure cleaner and more focused: every screen had a clear purpose, and the component library I built meant future iterations could move fast. Adapting to constraints in real time, without slipping the timeline, also built real trust with stakeholders.
Into the field
The thermostat is heading into field trials, where it will run on real hardware in real buildings.
I'll support the rollout by conducting usability testing and collecting metrics from live usage, then partnering with product and engineering to turn those findings into prioritized improvements, so the next iteration is shaped by how people actually use the device, not assumptions about how they might.
What I took away
- Design systems save you under pressure. A tight component library let me absorb changes without rebuilding each time.
- Consistency doesn't mean identical. Matching interaction patterns while optimizing visuals per device created one cohesive family.
- Stakeholder priorities aren't always user priorities. Choosing modern/high-tech over the top usability score taught me when to advocate and when to align.
- Shipping is a feature. Getting hardware in front of leadership on an impossible timeline built trust and momentum.