Over in the other blog post, we discussed the idea of introducing telemetry to measure a feature’s success and inform how to move forward with the feature or solution. Ideally, we would know early on in the development process, what telemetry we’d need, to make further decisions, and ensure health and usefulness of the feature.
We had designed a thermostat earlier already, with a crawl-walk-run approach. So let’s try to look at the thermostat that we want to deliver for the crawl state, and plan which insights we want to get, in order to plan the walk and run states after, shall we?
The thermostat in “crawl” as a minimum viable product that delivered:
- Very basic UI with a 3-state dial: “Off”, “Cool”, “Heat”.
- Pictograms for each state that explains each function
- The selected state runs until it’s shut down by the user – there’s no timer.
- LED indicator that shows operational state
- Selected mode (cool/heat) starts for “slightly cold” or “slightly warm” and progressively gets colder/warmer over time, until it reaches the maximum cool or warm temperature and continues in that state.
So – what do we need to look for, to figure out how to progress?
- Measure the time how long users stay in a given state – to see how “well” they find their feel-good temperature – per room and across all rooms.
- Measure how often the thermostat is turned on and off and on and off again – to figure out how well the “progressively colder/warmer” air flow is working out, and whether there needs to be a change prioritized – per room and across all rooms.
- Measure the room temperature over time, to correlate the “start/stop” actions with room temperature that users were looking for.
- Correlate the temperature development over time also during night, to understand for how many rooms there will have to be a “over the night, keep the temperature” function
- Track what the temperature “extremes” are that guests use for a 95th percentile or 90th percentile.
- Measure how long rooms in “extreme” temperatures take to come back into “feel good” temperature
- Track when the guest shuts off
- Be able to correlate thermostat data across the whole hotel or all set of rooms, to figure out if there’s patterns in different locations (i.e. patterns the south-facing sunny side generates – as opposed to in-between buildings sides)
- Since for “Walk”, we want to add a timer function that allows “+30” and “-30” minutes to schedule start and stop better, let’s find out if 30 minutes is the right size for adjustment. Or is it 20 minutes? 10 minute slides?
There may be a few things more after some more brainstorming and chatting to your Engineering team.
You may also want to ask guests for their opinion – either on a provided feedback form, that asks about their general experience in the hotel upon checkout:
- How easy was the thermostat to use – on a scale from 1 to 5.
- How easy was it to find the “feel good” temperature – in a scale from 1 to 5.
- How fast did the temperate adjust? – in a scale from “too slowly” to “too fast”.
FRICKELSOFT 