I’m pretty in love with this way of characterizing the value of Beeminder! (“Internalizing the externalities of your impetuous in-the-moment self…” Could be a new tagline if we can pith it up a bit.)
Also incredibly grateful for this whole thread, and especially to @nichlemn for starting it. Keep it coming!
Catching up a bit:
Great point that you can talk about, say, banking your results and avoid a lot of nerdery about cumulative graphs! I think the term “safety buffer” is reasonably non-nerd-friendly.
Agreed about the mixed metaphors – bee stings and Wizard of Oz. About the only place they fit together is the field-of-poppies image when our server goes down. So that’s not great! (If you don’t know what I’m talking about, that’s pretty great though!)
Thinking hard about replacing “derailment” with “sting”! The verb is arguably more awkward – “to derail” vs “to get stung”. I’m thinking we really need an active verb there. Like something you may choose to do (derail) rather than something that might happen to you (get stung). Could “take a sting” fix it?
Graph redesign: road.beeminder.com is the future there, but it’s still all too confusing. Working on this! (Super cute/clever ideas for adding pictures and stuff to the graph. Thank you! I’m not sure about actually doing that but I love the brainstorming!)
Beeminder as RPG: I hope when I tell you about Habitica you won’t run into its arms and never come back… Fortunately there’s a great Beeminder/Habitica integration for having the best of both worlds!
OK, you knew about Habitica already And you’re totally right that “Habitica exists” would be a terrible reason not to add RPG elements. Quite the opposite.
More community: yes please! One idea we’ve been sitting on: Group Goals
Brilliant points by @narthur, in particular about taking inspiration from YNAB. And I agree that there is a ton of low-hanging fruit / best-of-both-worlds kinds of things to make Beeminder more newbee-friendly. I think it will be a while before we have to make any painful tradeoffs between what’s best for newbees and what’s best for power users.
Integeriness: Yeah, so confusing to newbees! (And such a non-issue if you’re used to the ceiling() function that we persist in failing to prioritize it! -ish-face)
User-defined session sizes: thinking about this and a possible generalization of the integeriness setting…
@fatdentist97: And, yeah the “why would you be ok giving some company your money” objection is so frustrating and yet so common to hear that sentiment that it means there’s something wrong with our messaging! (Ah, yes, @narthur has made that point beautifully.)
Eek, I knew we were neglecting all the requests to let people beemind in other currencies but didn’t realize there was ambiguity about it all being in USD! I’m loathe to clutter the webcopy with “USD” everywhere an amount of money is mentioned. So I’m not sure right now how best to address this!
@oulfis: Oy, yes, the “I would lose so much money” response. shakes fist I’m often tempted to write those people off as just not Type Bee but you make a great point about more inspiration to draw from YNAB here in addressing people’s emotional/psychological blocks.
Dang, @narthur, your 6 stages of getting on board with Beeminder. So good. I kind of want to describe this with religious language but don’t want anyone to think I’m mocking that (which, to be clear, I wouldn’t be!).
As an economics major, I love it. As the OP of this thread, it seems like it’d make thing worse
(Though the term “internality” already exists in economics, for exactly this).
“Take a sting” sounds awkward, like you’re participating in an Amazonian stinging ritual. I think “You’ve been stung” is fine. If you’re worried about the “choice” element, it just needs to be clear that you were stung because you e.g. remained stationary while a swarm of angry bees were heading towards you,
Haha, the new Beeminder pitch: imagine your goal as a yellow brick road of daily choices… surrounded by swarms of angry bees in every direction. Walk the straight-and-narrow to stay on track, or fall victim to THE BEES!
Another idea: default starter goals. Based on the idea of the paradox of choice, it’s possible that the effectively limitless number of potential Beeminder goals is a real barrier-to-entry, as it’s something you might have to think hard about for a while (a particular obstacle for highly akratic individuals!). Some people may instead have a vague first-order desire to be “happier” or “more productive”, and would appreciate a few well-selected goals chosen for them.
For example, you might have a “Wellness Starter Package” that consists of three goals: “10 minutes meditation per day”, “Write a gratitude diary entry every day”, and “30 minutes of exercise a day”. The user is free to customise those goals to better suit their needs, but the key thing is that you’ve got a) something that should be fairly helpful to most people and b) something that you can get started on very quickly, without a lot of thought.
–
More ambitiously, I’ve for some time been intrigued by the potential for entirely algorithmic goal recommendations. It seems like there are a number of evidence-based interventions for improving well-being that people just aren’t familiar with. Or people may have a specific goal in mind, but they’re unsure of the optimal way to attack it, or how ambitious a target to set. It seems to me that a combination of evidence from the behaviour change literature and data from actual users could be utilised for this purpose.
I do wonder, though, if algorithmic goal recommendation may be infeasible, either because it naively assumes that success can follow a “recipe” when this is not the case, or that it may be theoretically be possible but would be in practice require extremely advanced AI beyond any current capabilities.
I feel like the opposite is true - people know exactly what their beeminder goal should be. It’s “work on that project I’ve been avoiding for 15 minutes a day.” Working on it might be sitting down and thinking about how to go forward. The obstacle is getting over the avoidance and trusting yourself enough to take the risk and put money on the line by making the goal.
There’s definitely a barrier to entry to making goals, and I really wish the process were easier. I like the idea of having pre-made goals. Instead of a package, how about a menu of 10 sample goals that you can choose from?
I also think the interface and lack of ability to do things from the apps is a problem as far as appealing to a wider audience. People want apps that are easy to use. If I were the Queen Bee I would deprecate the web interface and move towards being mobile-only. It’s ridiculous that you need to open a web browser and go through as many steps as you do to make a new goal. I should be able to open the app, click “new goal,” be presented with a menu, and make a goal with a single click.
The problem, though, is that people really have to want to do their goal. If you put in the Wellness Package (and sorry, this is me being me, but I’m simultaneously gagging really hard at the gratitude journal and groaning at powerful self-destructive spiritual practices with intense side effects being treated like eating more veggies) it’s something someone else is telling you that you should do, and we all know how well that goes over. I think for beeminder to work, you really have to want to do your goal.
And yeah, I’m skeptical about evidence-based interventions for improving well-being or algorithmic goal recommendation.
I think having a set of goals that work together for a variety of common things newbees want to do is an amazing idea. Beeminder does now capture intent when newbees sign up, so it’s even feasible to figure out what the most common goal families are.
I guess I’m a little skeptical about doing this algorithmically. I think a curated list of common goals + what they’re intended to achieve, how you can modify them for your own needs, etc, might be more useful to new users.
Also, re: getting rid of the web interface, I guess I have a different perspective on that. There’s currently nothing in the web interface that I can think of that couldn’t be made completely responsive (aside from the new visual road editor, which would be pretty stinking hard to make responsive).
Once the desktop interface is completely responsive, you could package those web views into the apps, ensuring the apps have all the functionality of the website without having to duplicate the features in multiple places. And that doesn’t prevent you from adding more mobile features to the apps along side the web views.
Also, having a completely-responsive website means you could start working towards implementing Progressive Web App features, meaning, as browsers and mobile OS’s get better and better at supporting PWA features, the website could potentially replace the apps altogether.
People know what they want to accomplish but I’d dispute that they know the best way to implement the goal.
There’s a lot of forum conversations and blog posts about “Should I beemind the number of times I do X?” and we usually say “No, you should beemind the amount of time you spend on X.” and then “Oh so I should beemind the time I spend reading?” and we usually say “You got us! No, you should beemind the number of pages you read because odometer goals are great”. And on and on.
The value add of default goals or goal templates would be to help new users fall into the pit of success of defining their goals in the way that makes the system function the best.
I feel like this example goes against your point. Beeminding is so personal. I used to beemind pages, but I switched to time, and don’t think I’ll ever be going back, because it works so much better for me, given I read in several mediums. And I’m pretty sure I’m not the only one.
That’s why I think “Here are some useful goal ideas that many people like and how you might set them up” would be much better than “Push this button and we’ll plop in a goal for you to do X.”
That’s exactly what I don’t want! I want the apps to replace the website.
But what do you mean by “completely responsive” and what is PWA?
I think you’re missing my main point, which is that no matter how awesome the website gets, you shouldn’t have to go to the trouble of going to the web browser when there’s an app for it.
I can’t really explain very well, but the psychology is totally different for an app vs a website. Using an app feels like it’s mine, it’s right there at my reach, under my control, convenient, accessible, seamless. Whereas having to use a website is just a big ugh. It feels like an additional hassle, a hurdle, a pain, like it’s not easily accessible, a trivial inconvenience, like it’s not under my control.
And granted beeminder’s website is comically bad, but the principle I’m talking about is true even for the best website ever.
Well, my vision for algorithmic recommendations is that (given a sufficiently large amount of data) you could provide useful tips. Maybe it turns out that “time read” goals derail 11% less than “pages read” goals, so “time read” gets recommended for new reading goals. Or maybe there are some demographics characteristics which predict “pages read” actually being more successful, so people in those demographics can get “pages read” recommended to them. Or maybe “frequency of derailment” is not the only relevant information we’re interested in, but there’s some other (perhaps hitherto unmeasured) distinguishing trait(s), like maybe self-reported subjective well-being.
Alas, it seems the main obstacle to this is insufficient data. If there were hundreds of millions of Beeminder users, you could get some pretty amazing data-driven insights. The social scientist in me wants Beeminder (and/or other life-tracking tools) to succeed on a large scale for this reason. For instance, the field of positive psychology interests me a lot. My earlier reference to a gratitude journal was because this seems to be one of the more promising positive psychology practices. Unfortunately, most social science research relies on relatively small sample sizes. It’d be great to see what works with sample sizes in the thousands or millions.
A key skill is emotional reframing. First, you have to emotionally accept that the derail is final, let’s deal with it. Then you have to reframe it emotionally. Failure hurts. But it also is a learning opportunity. You now know how much you can do, and you can prune your priorities. You have now gained self-efficacy, the ultimate drug.
(Or, you have found out never really wanted the goal that much. You are now free for better goals.)
To make Beeminder more appealing, you would have to show people this skill and hand-hold their way through the baby steps.
Hmm, I wonder if you could create a flow that coaches the user through this process the first few times they derail…
I can’t find a picture of it right now, but on Betterment, when you view your performance, if the current value of your investments is lower than the amount of money you’ve invested, it shows a message explaining that short-term losses are a normal part of long-term investing and aren’t something to panic about.
It takes skill to set good goals and make Beeminder useful. A predefined package is a great opportunity to de-skill or teach that.
In more detail, some common errors:
distractingly many goals. (I can only expend mental effort toward three things.)
measuring something other than the bottleneck, (e.g. people who have trouble getting out of bed might want to beemind sleep hygiene instead of getting out of bed.)
failure to tune the goals to their needs (e.g. derailing instead of knowing how demanding a goal should feel)
metrics that are not actionable (e.g. weight rather than intake of sugar and refined grains)
bee-goals that do not improve their end-goals in an evidence-based way (e.g. eating less calories if they actually want more flirts, e.g. flashcards beyond the first 1000 to 2000 words if they actually want to use a foreign language)
over-reliance on willpower (e.g. directly stated goals instead of using B. J. Fogg’s 3tinyhabits and beeminding systems).
perverse incentives (e.g. browers extensions to track unproductive time, instead of todos done, which leads to time-feeling-like-work instead of results)
And you’re totally right that “Habitica exists” would be a terrible reason not to add RPG elements. Quite the opposite.
-ish-face)
