Machine Learning (and Big Data) is Playing a Huge Role at Nest

It may only be for sale in the United States and Canada, but Tony Fadell reported earlier today at GigaOm’s Roadmap Conference that Nest thermostats have been activated in 65 different countries around the world.  With an IP-connected product, Nest immediately knows as soon as a new thermostat comes online by looking at IP gateways for every device.

Here are a few highlights from the interview - so far, Nest has gone through three major software upgrades (and eight minor updates), each of which has been pushed out to customers for free.  Major upgrades have included Airwave, which runs air conditioning more efficiently, and upgraded AutoAway algorithms (based on field usage data) that far more accurately predict when people have left for work and automatically saves energy.  Big data played a huge role in building and tuning these algorithms - learning that during the workweek, AutoAway could jump in after just 30 minutes rather than waiting an hour or two.

Second, Nest has been actively reaching out to users in countries outside North America - to ask them about where they would have wanted to buy their thermostats (among other questions), and asking about how they installed them, thereby fine-tuning Nest’s future international go-to-market strategies.  Still acting like a scrappy 20-person startup and staying close to their users.