Frictionless Tracking with Beeminder Autodata (QS 2015)

Saturday, January 23, 2016
By dreeves

Daniel Reeves at the QS15 podium

This isn’t exactly new news but a few months ago we posted the video of Beeminder CTO Bethany Soule’s talk at the 2015 Quantified Self conference and we finally got the video for my talk at QS2015 as well so we’re posting that too!

I’m including the whole transcript below after the embedded video but the real heart of it is captured in a few slides which I’ll reproduce here in text for better future searchability. First is a list of forms of friction (thanks to Aaron Parecki for first defining this notion of frictionlessness for self-tracking):

  • Entering data
  • Charging a device
  • Syncing
  • Taking devices on and off (e.g., if not waterproof)
  • General usability

Mostly I’m focused on the first item in that list, entering data. (The other items are why I like, for example, my Misfit Shine for step counting.) Which is to say automatically tracking data. Here’s a list of metrics that can be automatically tracked and sent to Beeminder along with the gadgets or services you need to do so:

  • Steps, sleep (Fitbit, Misfit, Garmin)
  • Weight (Withings, Fitbit Aria)
  • Time spent (RescueTime)
  • Todos done (Habitica, Trello, Todoist et al via IFTTT)
  • Number of workouts (IFTTT location)
  • Tweets
  • Emails archived (GmailZero)
  • Points on Duolingo
  • Blog posts posted (RSS via IFTTT)
  • Phone calls placed (IFTTT or Tasker)
  • Files in a Dropbox folder (via IFTTT)
  • Pocket articles read/queued (IFTTT)
  • Words written/edited (Draft integration)
  • Time spent learning (Skritter)
  • Points earned on Code School
  • GitHub commits or issues closed
  • Instagram photos posted (via IFTTT)
  • Reddit posts (IFTTT again)
  • Toggl entries (Zapier)
  • Bitbucket commits (Zapier)

And for good measure, here are some semi-automatable metrics:

  • Pushups (via touching nose to screen)
  • Time via Timer in Beeminder Android app
  • Steps/sleep via gadgets that require frequent charging
  • IFTTT’s “Do Button” (Trigger is pushing a button on your phone’s homescreen)
  • TagTime (also by Team Beeminder)
  • Tracking by Reply to Beeminder Bot w/ data

Clearly a lot of options for frictionless tracking (and beeminding!) and the possibilities are exploding. This is a big part of the Internet Of Things. It’s also about to get cheap and easy to have sensors on everything. One I’m particularly excited about on the horizon is the Droplet physical button.

Ok, here’s my now slightly outdated talk!

Hi, I’m Daniel Reeves, cofounder of Beeminder, here to talk about Frictionless Tracking with Beeminder Autodata. Beeminder is Quantified Self plus commitment contracts, or data-oriented behavior change. So we’re all about leveraging your QS data to impose discipline on yourself. But mustering the discipline to enter data can be a catch 22. Which is why we love ways to automatically collect data about oneself, and Beeminder has ways to help with that.

I know some of you have no interest in the commitment device aspect of Beeminder. It’s not for everyone. But this is of more general interest, at least for a Quantified Self audience, whether or not you’re into artificially imposed discipline.

Here’s the collection of Quantified Self apps and services and gadgets that Beeminder currently integrates with, the most recent being Skritter, for learning Chinese and Japanese.

Autodata collage

We’re proud to say this list has tripled since the previous QS conference. We’ve been pushing hard to make Beeminder integrate with as much of the Quantified Self world as possible.

In fact, we made a quantum leap in that regard when we recently added Zapier and IFTTT Channels. Now those platforms are doing much of the work for us. Every time another QS tool gets on the IFTTT or Zapier trains, that’s one more automatic source of data that can be fed to Beeminder.

For anyone who doesn’t know about IFTTT and Zapier, IFTTT stands for “If This Then That”. It’s a service that lets you connect two different web services or apps together using Recipes of the form “if this happens” then “do that”. You pick a Trigger — new instragram photo, rain forecast, you name it — and IFTTT will do some specified Action whenever the triggering event happens.

Zapier is the same idea but is more nerd focused, giving you more control over the recipes you set up and is a little less friendly for non-technical users. It also tends to focus on startups as its target demographic, meaning it has lots of channels for productivity tools like to-do lists and time trackers. It’s freemium and if you use it very heavily you’ll probably want to pony up for a premium plan. IFTTT, on the other hand, is totally free, and really impressive in its simplicity and slickness. Both are powerful tools for automating data collection about yourself.

Back to Beeminder, the goal is to eliminate the step of actually entering data. Going to a website to enter data is a pain. We have smartphone apps, which may help. In some cases replying to our email bot with data can remove most of the friction. But truly frictionless self-quantifying means you just do the thing you want to measure and never give a single thought to the actual measuring.

Maybe you wear a Fitbit or a Misfit or a Garmin Vivofit or a Jawbone UP. If you connect any of those to Beeminder, letting Beeminder read the data that those gadgets are collecting, then Beeminder will automatically make a graph for you showing your cumulative number of steps or hours of sleep.

Whenever you’re in danger of going off track, meaning your average number of steps (or whatever you’re measuring) has fallen below the goal you set for yourself, Beeminder will warn you. If you don’t heed the warning, it charges your credit card (though we’re very friendly about it and you can easily cry foul, unless you weaselproof yourself, but that’s another story).

We recommend automatability as the top criterion for what metric to beemind in the first place. Like maybe you have a goal of getting in better shape. Body weight isn’t the greatest metric for that, for a few reasons. But it is very easily automatable, thanks to Withings and Fitbit’s Aria scale, both of which we integrate with. You literally just stand on the scale and your graph is updated. And, especially if you’re well above your ideal weight, it’s a fine proxy for fitness.

Or take gym workouts. Thanks to the iOS and Android location Channels on IFTTT, it’s dirt simple to set up a recipe to automatically send a +1 to Beeminder whenever you come within some radius of your gym. A personal favorite of mine is built in to the Beeminder Android app: I do pushups by putting my phone on the floor and touching my nose to it with every repetition. Beeminder counts the pushups that way and magically updates my graph.

For productivity, we highly recommend RescueTime, which beautifully automates time tracking and integrates with Beeminder. We have a vaguely similar tool called TagTime that’s much less friendly and tries to hit the sweet spot between fully automatic tracking and manual time tracking. But since this is all about frictionless tracking, RescueTime is where it’s at.

One more option for Android users is our Tasker integration. Tasker is like IF This Then That for your Android phone. You can set up Tasker tasks that trigger based on any conceivable event that your phone is aware of — the screen coming on, any app opening or closing, location, you name it — and cause data to be automatically sent to Beeminder, either directly, for counting events, or by starting and stopping a timer, though RescueTime is generally a simpler way to measure time spent.

Finally I’ll quickly mention that Beeminder has an API so the sky is really the limit for automating feeding your Beeminder graphs. We also intend to keep adding more official integrations. So makers of other QS tools, please talk to us! As our current integration partners will attest, it’s a pretty nice feature to offer your users who may want to use your tool more than they do, if only they could construct a small kick in the pants to stay on the wagon.

Tags: , , , , ,

  • John H. Long

    Here is one I came up with recently. I am running behind on processing paperwork. I have a wemo insight and a shredder. I get a Beeminder point for every minute of active shredding I do. Since I scan and shred, there is always enough shredding if I’m processing my paperwork regularly. I can set an aggressive goal for now, then back off after I get through the backlog.

  • Grayson Morris

    Coming in to comment on the wake of the new Beeminder integration features recently built into the Complice daily planning system. These made using Beeminder MUCH easier for me, because:

    1. Beeminder autotracking only let me use one automated input source, without even the option to enter data manually from time to time. So if I wanted to make progress on, e.g., my guitar-playing goal, I *had* to open the Rock Prodigy app I’d coupled my Beeminder goal to and keep it live. This led to some stupid sessions where I was practicing something else that was an equally valid step toward my guitar goal (but whose time I couldn’t log in Beeminder) while I had the Rock Prodigy app open and would occasionally move the mouse to keep it live and tracking to Beeminder. Kind of the opposite of frictionless.

    2. So that left manual input, which also felt Hard and Overkilly, since I was already using Complice to track daily intentions like “practice guitar”. Cross it off there, then go enter the time spent into Beeminder…frictiony. Very frictiony.

    3. Complice had simple Beeminder integration, but it only let me connect a given Complice goal to a given Beeminder goal. But the two don’t usually align; my Complice goals are generally bigger-picture things like “be badass at 80” while my Beeminder goals track much narrower metrics. For example, I have at least four Beeminder goals related to my “be badass at 80” Complice goal, which track steps, pushups, and other fitness and health metrics.

    4. Now Complice (as of a week or so ago) has added an “arbitrary task” feature that lets me link a single Complice goal to any number of Beeminder goals, using text matching and other features to distinguish them. (I know this all sounds like Klingon if you don’t use Complice; in that case, just read this as “AWESOME.”) So, for example, if my “be badass at 80” Complice goal is number 6, I can tell Complice that any task for goal #6 that contains the string “pushup” should send the number in that task to my Beeminder pushups goal. Now, if I add a task to my daily Complice intentions that reads “6) do 4 pushups”, Complice will send the number “4” to my Beeminder pushups goal when I mark the task complete. Lovely, perfect, frictionless automated data.

    (It’s even cooler than that, but I’ll spare you the details — *snort* like I haven’t been too detailed already.)

    (Also, don’t laugh at my paltry 4 pushups. They’re MAN pushups, I’ll have you know, and I’m just starting. And I’m 5’3″ and 115 pounds and a grrl. Actually, if I were a grrl I could probably eat those pushups alive. So make that girl. Woman, really, seeing as I’m almost fifty. But I digress.)

    Anyway, the point is that this new Complice integration has removed a ton of friction for me, and I’m happily back in the saddle again in both apps.