Hard-Committing To Do Something “Soon”

Thursday, October 22, 2015
By bsoule

A student sprinting to turn in a paper right before midnight

Deadlines are really important. Without them things don’t ever get shipped. But they’re also — if you can hard-commit to them, which you have to or else they’re pointless — kind of awful and arbitrary and stressful.

Epiphany: Beeminder gives a way to get the key advantage of a deadline without the stress and arbitrariness. What you do is, instead of “I hard-commit to finishing X by date D” you hard-commit to spend half your time on X until it’s deployed. It’s like committing to finish something “soon”, but in a way that’s actually meaningful. You’ll be diligently working on it until it’s done.

I’ve done this twice so far. The first was for committing to finish several new autodata integrations and, recently, for revamping Beeminder’s reminders. Let me tell you about that last one, as promised at the end of that blog post.

A word from Bee on the $1215 she spent revamping reminders

Did you know you can win money from the Beeminder founders’ pockets? Because I paid out $1215 total over the 175 hours and 3 months I spent revamping reminders, including my highest ever Beeminder derailment at $810 [1]. Let me assure you, lest you suffer sticker shock on my behalf, that it was well worth the price. To be sure, I was literally very unlucky with the final $810 derailment — there was a more than 5 hour TagTime ping gap, which is a rare event, happening less than 1% of the time. [2] Even so, compare it to the price I’d pay to contract out the work to someone else, and it sounds like a steal! And taking a slightly broader view, there are many of our meta goals (my meta goal, Danny’s, the UVI goal, Infras, my previous autodata goal, oh, and the goal for this blog) that are literally vital to Beeminder’s forward progress, and which would not have happened without beeminding them. Total pledges paid to users on all of those meta goals over multiple years is still just a few thousand dollars! A pretty good price for the existence of Beeminder. And all of that doesn’t even account for the good will it generates to pay our pledges out to you users.

Up next: how we wish we had thought of this a year ago when we hard-committed (for another $810) to finish several big new features for our 4th anniversary post.



[1] Congrats to Geoff Hubbard for winning it!

[2] This sentence might not make any sense. If you’re curious what the heck I’m talking about, check out Danny’s other blog. If you don’t feel like reading all that, just focus on the phrase “very rare event” and let the rest slide.


Illustration by Kelly Savage originally for a Messy Matters article about setting deadlines for students.

Tags: , , , , ,

  • olimay

    I’ve come to the same conclusion: for complex projects (or a project that’s sitting in a complex mix of priorities) you’ve gotta Beemind the inputs before the outputs. Putting the stakes up earlier in the causal chain makes for better maneuverability. Navigating complex systems means it’s better frame your journey in small steps. Adriano of Gingko app wrote about dealing with complexity/non-linearity in goal setting several years ago — http://blog.gingkoapp.com/essays/focusing-process-alternatives-goal-setting-resolutions — but he framed it as *alternatives* to goals, and I didn’t know how to reconcile that with still wanting to have goals! But the key concept of smaller steps, sooner accountability, and faster experimentation transfer readily. (We find them in lots of popular methodologies, notably Agile.) That’s the power of setting and meeting a daily Beeminder quota–it ensures you’re at least exerting control over the parts of the process that *are* linear–the time you put in.

  • http://mbork.pl/ mbork

    Wow, Zeno strikes again!

    This is in fact quite clever. I envision one problem with this approach, though: you have to define clearly what “half of your time” means. I’m sure you did, Bethany, but not everyone subscribes to the Quantified Self lunacy. (I do, btw, not 100% yet, but more and more.)