What if Data Scientists would run the Smart Grid?

No, I am not talking about any data scientists, but the ones that work at Booking.com. And I don’t mean that this army of data scientists should run the peak-shaving algorithms of the smart grid. Although that would be an interesting experiment… When I talk about the Smart Grid I actually want to talk about smart meters. And I would like to focus on an important goal of smart meters namely to stimulate energy efficient behavior. The running part would roughly translate to playing with data.

Are we clear? Smart meters + energy efficient behavior + playing with data? We’re good? We’re good!

How will smart meters contribute to more energy efficient behavior?

Letting end consumers know how much energy they use should nudge them to rationally decide to use less electricity. For this to happen smart meters should be able to show the information on energy use to end consumers. As my colleague Farid Soebrati pointed out this came with a bit of controversy recently. Should smart meters come with connected displays or is it more logical to show the energy related information on a screen the end consumer already owns? I am guessing it might just be the end for in-home displays for smart meters.

More fundamentally, coming from the field of persuasive technology myself and still keeping up with the latest developments, I was intrigued by a paper by my former colleague Sander Hermsen who evaluated the available research on feedback technology (PDF) and it’s effectiveness. In-home displays for smart meters are considered feedback technology BTW. Conclusion of Sander’s research: it’s not proven that this kind of technology changes the habits of people. Not yet at least.

When I started my PhD on this topic the first realisation was that behavior change is very hard. Much harder than the popular press sometimes makes it out to be. But there is hope. Data-driven companies like Booking.com are trying hard to crunch a lot of data to make behavior change happen.

How will data contribute to more energy efficient consumer behavior?

With over 1 million hotel bookings per day and tons more interactions every minute, Booking.com can be compared (with some imagination, come on people!) with a smartgrid of a small country or a large city and millions of interactions with smart devices. Picture that every hotel booking could have been a person turning the thermostat a bit lower. If only their data scientists were running the smart grid…

  • Booking.com’s KPI: more hotel bookings
  • Smart meter/grid’s KPI: more energy efficient behavior

Booking has a very clear strategy on how to change behavior. Having run probably more than 100.000 experiments the last decade (last time I was there the experiment velocity was over 1000 experiments a day, so this estimate is definitely on the low side) they appreciate data and continuously question opinions and assumptions.

Do you like what you have read so far? Don’t miss the quarterly update of what I am busy with.

Take a look at a talk by former colleague Stuart Frisby, head of Design at Booking.com and you will understand how important it is to be humble in your quest to change behavior.

Be very humble indeed and run a lot of experiments. But what experiments should we try first?

What experiments are needed to get energy efficiency right?

As Stuart points out trying to make people do something depends on the context. And your context is not my context. You have to test your own hypothesis. Booking a hotel is not the same as turning down your thermostat. Really it is not.

And I just want to mention; I don’t know either what is going to work. That’s why it’s good that we have multiple screens and touch points we can test. BJ Fogg and Dean Eckles have already shown in 2007, with their seminal book Mobile Persuasion, that smart devices have a huge potential for behavior change.

Good! We have the technological platform to test what works, so we can start. But where did we have to start again? Let’s follow the experts on experimentation; Booking.com has tried almost every insight from behavioral science on their website. And they found that Professor Robert Cialdini’s persuasion principles are a very reliable source to test behavior change in the wild, the same principles that made Opower famous in the energy industry. (surprisingly the cognitive biases made famous by Nobel prize winner Daniel Kahneman in Thinking, Fast and Slow were less effective)

I would say that’s a great start indeed. And now, it’s up to the utilities to test their own hypotheses, making a serious effort in stimulating energy efficiency and at least have a chance of turning around climate change.

Do you like what you have read? Get a quarterly update of what I am busy with.

BIG P.S. over the years I developed my thinking, and no one has influenced me more than my mentor at Booking.com and good friend Mats Einarsen. Thank you Mats for being a rolemodel to me and for so many in the discipline of business experiments.

P.S. 2 Tom Raftery pointed out that making automatic decisioning the default would probably have the most effect on energy efficiency. And I completely agree, it does take away quite a bit of the shine of this blog though, and that’s why I’ve put it in yet another P.S.