What I learned from sleeping with the band


I have been using the Microsoft Band since it was brought to market and I am a big fan of it. Although I actively use all its features, I have learned the most about myself and made the most changes to myself based on the data captured during my sleep.

(be sure to read the part -> I am not a machine at the bottom)

I wasn’t always happy with my sleep and woke up many times not feeling like it was a good one. After I started tracking for some months, as shown in Figure 1, it became obvious that actually, I was sleeping pretty good after all.

227
Figure 1, Microsoft Band sleep monitor graph

I could see from the collected data how long it took me to fall asleep, if the sleep was deep or light, how many times I woke up, my resting heartbeat, duration and many others. Just seeing the graph with the metrics and looking deeper into my own individual patterns helped me feel better about my sleep, even if it was only physiologically, I did start to feel better about my sleep. I must assume that the unknown pattern of my sleep led me to believe I wasn’t sleeping well, when in fact…apparently I was, as you can see in Figure 2.

228
Figure 2, Microsoft Band, Microsoft Health sleep analysis and comparison

I noticed that my sleep was more efficient and more restful than the average person in my demographic group. NOTE: when you sync your data with your phone (the Microsoft Health app supports iOS, Android and Windows Phone) and you have the online Microsoft Health running, you get to see how your metrics compare with others in your demographic. This is where the real learnings can come from. In my case, one which jumped up specifically was the amount of time I was sleeping, as shown in Figure 3.

229
Figure 3, I sleep too much, not always with the band

I noticed that I was sleeping 1.7 hours more than the average person in my demographic. Combining that duration information, Figure 3 with the chart metric, Figure 1, that showed it often took some time to fall asleep (orange color), I decided to make a change. The change was that I wake up about 1 hour earlier each day. The outcome of the behavior change is that I love the additional hour and I go to sleep much faster.

Using the Microsoft Band has added 1 hour per day to my life!

The above story is true and I really did that. The experience though, made me think more about the impact that using my device (Microsoft Band) to capture data had on my life. I thought also about the fact that others are doing the same, and had our data not been placed into a larger pool of big data, analyzed and presented, the changes I implemented likely would have never taken place and I would still be thinking that my sleep was keeping me from performing optimally.

The problem or opportunity is that I had to analyze the results and come up with the change myself. It would however be simple to write a program algorithm to look at the results and make that recommendation. It is just a simple “if time-sleep > 20% average and time-to-sleep > 30 minutes then wake-up-earlier!”

I am not a machine

Take it one step further and assume that I was machine, I know that a machine cannot be tired and they don’t need sleep, but ignore that fact for a moment. Assume that the machine had a device that measured efficiency and restfulness, assume that information is sent to a big data repository and analyzed. Then, an automated real-time analyzer looks for a condition, discovered by running the algorithm I mentioned above and when found, another algorithm automatically makes the machine more productive by using it for 1 more hour per day.

In the case of a machine, you can get more use out of it and likely produce or perform more. From a human perspective, I used one of the additional hours to write this article, I hope you like it, the reaming hours… are mine.

Comments (2)

  1. Alan Stoops says:

    Fascinating Benjamin. I too have had a Band 2 since October and the comparisons against your my peer group are truly interesting - and, as you say, some sort of guidance algorithm would be useful - medical disclaimers aside.

    Some of the terms confuse me though. In my case how can it be that my sleep efficiency is low (62 vs 87), my time asleep is low (4.7 vs 6.1), my restful sleep is low (0.6 vs 1.6) and yet my sleep restoration is better than average at 44 vs 38? I don't know whether to have more coffee or less.

    1. @Alan, I am not able to provide this kind of information as I am just a user of the device like you. I am confident that over time much more information will come from this band and that we will learn more about ourselves.

Skip to main content