Designing a more flexible pledge progression

Could the pledge minimum be tied to the amount of goals you have? I feel like that could be helpful for people that have a lot of smaller Beeminders.

3 Likes

I like how you think! Like it’s not a per-goal minimum but an aggregate minimum on how much you need to have at risk. I think we have to resist things that add any complexity, which might include this (and maybe half of what’s already in the proposal!) but this makes sense at least in theory.

[thinking ensues]

I’m noticing more monkey wrenches with the idea, like what happens as your number of goals changes, how can the pledge floor be made transparent to users, what about loopholes with dummy goals, …

So I’m currently less optimistic about this but, again, love the way you’re thinking!

1 Like

I think users will be happy to get some more granular pledges and more ability to customise! Also, that ability to drop pledges all the way to whatever in one single 7-day drop is a good plan; at the moment we shirk-n-turk that when users ask, because people could just archive their goals and start anew in that time.

I would very much like to see pledge decay happen; I think that does introduce a bit of opportunity to feel you’ve achieved something with the goals. That said, I don’t think we’d need to include pledge decay in order to get this out of the door!

I still dislike the idea of not setting my cap at the outset. Feels that, no matter how we rationalise it, it will sound as if we’re counting on you forgetting to set your pledge cap to get more money, whatever our actual intentions are. I especially don’t think there’s a good way to bring the actual reasoning across to users; the very clear ability to set a pledge cap is part of what reassures people that their money is safe with Beeminder.

I know what the pledge should be for every goal, and I rarely need to change it. Can’t think of a goal where I’ve needed to change it in… at least the past year? That is a benefit partly of a single inflexible schedule: you can learn where on that schedule motivates you in general for goals with varying difficulty. I’d miss that aspect if I had to do more tweaking of my own!

3 Likes

Sincere question: is it really so?

Saying that jumping straight to one’s Motivation Point makes Beeminder not make any money is the same as saying that we’re in fact homo economicus and we’ll never fail once meaningful consequences are on the line

Personally, the concept of a single motivation point isn’t really accurate, and I suspect that this is the case for many other Beeminder users.

For a given goal, that potential sting needed to motivate me is going to change significantly per day, based on what the goal is and how I’m feeling. If I partially built up a buffer yesterday and only need to do half as much work, then that’s less effort for the goal. If I slept well the night before or I have more time left in the day, or I have fewer non-Beem chores, then I’ll need less motivation. And all of those can be inverted to lead to me needing more motivation on a given day. That is to say, there are many factors, and my motivation point is going to vary widely.

Because of this, a steep pledge schedule will lead to the pledge point being much higher than my average motivation point. Consider a goal where my current pledge is $40. That’s generally a pretty motivating point for me, but on a rare few days I’ll derail on it. If I’m using a schedule where my pledge quintuples every time I derail, then after derailing, my pledge will be $200. This will be way too high for me ever, and I’d almost certainly weasel out of it.

Thus, we should want a shallow pledge schedule where I increase from somewhere in the low end of my motivation range to somewhere that’s higher in my motivation range, but still in it, since it will be low enough that I won’t weasel, but high enough that it’s motivating.

This also has implications for short-circuiting. @dreev said above that “you can’t jump straight to your Motivation Point because Beeminder never gets paid that way.” However, if your pledge is set to your average motivation point, or even a bit above that, then you will derail (and thus pay Beeminder) on days when your motivation point is higher than that value, which won’t be uncommon. Thus, short-circuiting has no reason to be a Beemium feature, at least for anyone whose motivation point isn’t the same every day (which I’d expect is literally every user, since we’re all human).

5 Likes

Also, I think that it’s a must-have for pledge caps to be optionally set in goal creation. Users, especially Newbees, want to be able to say from the beginning that they won’t be paying infinite dollars on a goal, and having the pledge cap be set later means that it’s too easy to forget to do it. Plus, explaining to Newbees that it will be set later is going to be confusing

2 Likes

I’m surprised so far by how little love steep pledge schedules are garnering in the poll! Maybe people are thinking of shallow vs steep as mutually exclusive? (User-me voted for both – shallow pledge schedules are good for some goals and steep pledge schedules are good for others.)

That’s certainly one result that could change our mind on this proposal, if everyone just wants more granular pledges.

I’m thinking hard about how to square the circle of shepherding newbees to their motivation points without being too prescriptive or annoying to veterans. I’m certain the best of both worlds exists! It might take some experimentation to find it. For example, how annoying would it be to veterans if there were magic links in legit check emails that said “click this to make your pledge stop increasing or click this to drop it back to the previous level”? Or maybe there’s some other way to give you full and immediate control over your pledge in the 24 hours following a derailment.

Yeah, great point, I don’t want to lose that! I think this proposal can avoid losing it in that you could just never tamper with the default of pledge doubling.

Ah, yeah, this is a big philosophical question! But empirically and anecdotally it feels like people putting crazy high pledges on things from the get-go doesn’t work well, neither for users nor Beeminder. There are exceptions! Allowing that on the Beemium plan is the compromise and we feel pretty good about that.

PS: @gbear605, good point about how variable Motivation Points are! You’re using that to argue for shallower pledge schedules and more granularity, and that makes sense, but I think there’s also an argument that the wide variance of Motivation Points across one’s goals means that a steeper schedule can be valuable. That’s how you can quickly and with less cost, reach high pledges on the subset of goals where you need high pledges.

I guess you’re focused on how much one’s Motivation Point changes even for a single goal. That feels less true for me, as a user, but I’m trying to put myself in your shoes and think through the implications if I’m the atypical one on this (as I often am!).

PPS:

I think there may be a way to do it nonconfusingly. A couple ideas I’ve had: (1) Get a global pledge cap as part of account creation when it’s more clear that we’re asking for more of a theoretical upper bound on how much to ever have at risk, where it won’t be confused with “how much do you want to risk right now on this goal?”. (2) The magic link idea above. (3) In all our marketing and onboarding copy, really double down on the exponentially increasing pledges aspect, and kind of make it intentionally scary sounding. Maybe this one doesn’t accomplish much other than helping with expectations management. I agree that we need to make sure no one ends up with a pledge that makes them want to weasel out or flatten their commitment because they forgot to cap their pledge.

Now I’m thinking it might just be all about the UI. Maybe something fancier than just “pick initial pledge” and “pick pledge cap” that forces the user to understand, without reading webcopy, that those are different things and the latter is best left frighteningly high, or on the edge of frightening. It’s a subtle thing! You’re asking the user to imagine a counterfactual world. “You want to have $5 or at most like $30 at risk here, but if HYPOTHETICALLY those weren’t motivating enough after a while (and they might not be as you come to appreciate the awesome power of Beeminder and find that paying $30+ occasionally is no big deal for how much you accomplish from beeminding) then what amount, in that state of the world that you can’t currently fathom, would be the most it would make sense to risk before you’d rather throw in the towel?”

Would having to draw the progression on a graph induce newbees to take that kind of mindset? It seems really hard compared to just saying “we’ll adjust your pledges to maximize your motivation for you – just say when they should stop increasing”.

Further idea: We make it part of the normal derailment protocol: “You just derailed and are paying $30! You have 24 hours to pick your next pledge – anything from $10 to $90. (If you don’t pick something then the default amount will be $60.)” And we make that as in-your-face as possible to minimize forgetting.

2 Likes

I wouldn’t worry much about annoying veterans with too much information and options when they’re creating a goal. Afertall, no one creates a new goal every day and it won’t be significantly annoying to be extra descriptive for newbees.

2 Likes

I’m with @gbear605 here; my motivation points varies massively from day to day. If it weren’t so, then no one would ever derail on a goal that they’ve been succeeding at for a long time! But every now and again, I derail on my $90 goal, because some days I’m busy/exhausted/stressed/have enough going on that $90 seems like the lesser of two evils; that is, it’s not sufficiently motivating on that day, even though it is sufficiently motivating most days.

Maybe it makes more sense to think of it probabilistically. Every pledge amount (for a given goal) results in a certain probability of me being motivated to do that goal. For some goals, $0 results in functionally 100% probability of me doing that goal (I’m entirely intrinsically motivated, and beeminder is just reminding me – thanks beeminder!). For most of my goals, $5 results in about 97% probability of me doing that goal (I derail about once a month). And for some goals, I don’t really get good odds of doing the goal regularly until I get up to $30 or $90, because they’re inherently hard (like doing 5 hours of paid work a day). But for any given pledge, it’s still a probability that it will motivate me – most of the time $10 motivates me just fine for my “5 hours worked” goal, because I have other extrinsic and intrinsic motivations for that, but I’ve found that on days when that other motivation wanes, $10 is not enough for 5 hours of work, which is why I set it to $30.

So to get into the nitty-gritty, the $30 to $90 range is right around my motivation point for Hard Things, but that’s a HUGE range! I basically have to pick between “the price of a takeout meal” and “NEARLY ONE HUNDRED DOLLARS”. Having more levels between those two in particular would really help me balance between “scary enough to motivate me” and “not so scary I don’t stretch”. There are definitely days when $30 isn’t motivating enough, but $90 would be too scary just across the board (I’m considering my 5 hours worked goal specifically). Being able to pick $50 would be great.

1 Like

:eyes: hides own graphs guiltily

(It’s not quite every day, but it’s whenever I start reading a new book! Which is almost every day, sometimes, or sometimes multiple times a day – I read four on Saturday. That said, extra stuff for newbies wouldn’t bother me as long as the actual process is smooth, and I’m just popping up to be an obnoxious reminder that There’s Always One. :joy: Right now I long for a lot more during goal creation – like adding the end total, much as @dreev would say that’s something few people need during goal setup because most goals continue infinitely, and setting my autoratchets – so I couldn’t really complain if we gave it some more love.)

1 Like

Here’s my illustrated argument for lower pledge caps being better. I hope it doesn’t out me as being a terrible Beeminder customer. beechart

The red line shows the max derailment fee I’ve paid in each year, the blue line is the total I’ve paid Beeminder each year (including the green line which is my subscription fee). The bars show the number of times I derailed each year (right axis) and mostly for my own amusement I’ve split the derailments into which rough category of life the goals were helping with. The most I’ve been stung for is $90, about a year after I started using Beeminder, and all that seems to have done is put me off. I derailed entirely on that particular goal, and though I would have said I got a lot of value out of Beeminder over the next few years it’s clear that I derailed less and paid up less. I mostly made everything safer (though with a non-negligible “I got better at doing things” in there too).

During 2020 I decided to push myself more, skate closer to the edge, risk derailment a lot more, but reduce the sting of each derailment. I created a lot more new goals to try ideas out as well. I still have pledges capped at $30 by default but mostly reduce them down and try to reevaluate whether my targets are reasonable regularly (I should probably beemind that!). I’d definitely say that Beeminder has more value for me this way. And it looks as if I have more value to Beeminder this way as well. I’m not just thinking of the slight rise in fees I paid Beeminder in 2020 compared to 2019 but also that the long tail of the graph is probably sustainable income for Beeminder for years to come rather than trailing away to nothing. I’m a happy repeat customer and hopefully not one who takes up e.g. lots of support resources.

Whilst I understand how the pledge-tripling schedule should motivate me in practice I find the gaps between the amounts too big to climb above the first couple of steps of the ladder. It’s pretty much how one part of my brain understands I’m safe looking through the glass at the top of a tall building, but another part is just “aargh, let’s pretend I’m not here” (which in Beeminder terms is something like “let’s pretend and enter fake data” which then ruins everything).

7 Likes

Am I using it wrong? I personally have been much less motivated by “do it or pay up” and much more by the positive feedback loops articulating what I want (in the next week), making it happen, seeing that it happened (the graph), knowing that I did it, and building up the confidence that I can do these things reliably, at least using this tool.

If I recall correctly, at some point the idea was floated around (and if not, now it is) of having multiple pledge schedules available and users can pick either per user or per goal which schedule to use.

current one:
5 → 10 → 30 → 90 → 270 → 810 → 2430

or one that allows repeating at a level (perhaps with a time component - derail again within xx days?)
5 → 10 → 10 → 10 → 30 → … and so on

or one with a decay:
5 → 10 → … 5 → 10 → 30

or one I would like to see based on the Fibonacci sequence, start it at five if you want:
5 → 8 → 13 → 21 → 34 → 55 → 89 → 144 → 233 … and so on

I can image a concern being that charging less than the user was willing to pay is money left on the table and thus it makes sense (or seems to anyway) to bump the pledge quickly so as to collect the 5+10+30=45 through three derailments instead of merely 5+8+13=26.
This of course assumes some limited amount of pledge capping and some particular rate of derailing.

I strongly agree with this. (For end total in particular, that has a pre-req of having “open-ended” be the default. As soon as we have that, then we can have the best of all worlds and ask for the end total (if any) as part of goal creation.)

Holy cow that is a gorgeous graph. :heart_eyes: :chart_with_upwards_trend:

(And, ha, you’re an amazing Beeminder user on every dimension!)


Thanks for all the feedback and conversations here! It’s gradually crystalizing for me, even if I don’t know the answer yet. Part of it is reconciling these two tricky constraints:

  1. Don’t let anyone’s pledge go beyond where they need it capped (very bad for both user awesomeness and revenue!)
  2. Don’t ask newbees for a pledge cap in a way that they’re liable to confuse with the pledge itself, or ask for it before they can appreciate that pledge caps should be higher than one would think when first starting out.

Or, more pithily:

  1. Don’t let anyone’s pledge get too high
  2. Don’t let anyone’s pledge max out too low

Both, I contend, are very bad for the user (though the feedback I’m hearing here is that #2 isn’t as much of a concern). #1 means making pledge caps easy to set and hard to forget to set. #2 means not asking a newbee for a pledge cap when they don’t know how to think about that yet.

(Ironically, failing at #2 feels more money-grubbing to me, like tricking a newbee into capping their pledge below their Motivation Point so they just keep derailing and paying instead of staying on track on their goal! :grin: I’m not sure how serious I am here but the real point is that we want to succeed at both these things and get everyone to the right pledge!)

It has a catch-22 feel to it but we’ll find the answer! Here’s some copy brainstorming:

A powerful feature of Beeminder is that in the process of beeminding you figure out the optimal amount of money to have at risk for maximal motivation. This amount is sometimes much lower and sometimes much higher than you might think at the beginning! Every goal starts out with a nominal amount at risk and increases each time you derail. We recommend leaving that open-ended (you can always cap it or drop it back down if you find it’s gotten too high) but if you know ahead of time where you’d like to cap it, just click the dollar amount on the goal page as soon as the goal’s created.

Guess I missed this in the Beemail, so I’ll reply here.

Pledge amounts don’t matter much to my effort not to derail. $0 (I have Beemium so I can do that) is a bit different from >$0, $10 is barely different from $5, $30 is a bit different from $10, and I never use a pledge cap over $30, because I’m pretty sure that would not increase my motivation to meet the goal at all, but might make me pay so much money that I just quit using Beeminder. I don’t think I care massively about the multiplier either, although I guess for the pledge levels I ever actually use, doubling would be more consistent than tripling.

Being able to set a pledge cap on initial goal creation isn’t necessary, but having to wait until you derail is user-hostile. Please at least exempt Beemium so I don’t have to do that. Maybe have an account-level flag on whether someone has ever derailed and had a pledge escalate (and preferably a non-$0 pledge so they’re pretty sure to actually notice) so you can ensure people understand the difference between pledge and pledge cap, without artificially restricting them?

I think of the $5 pledge floor as accounting for Beeminder’s overhead of being willing to respond to emails about the goal, credit card processing fees, etc. If not for that, $5 would be way too high; $1 is probably a very common correct amount to motivate the user’s desired level of effort on a particular goal. But Beeminder folks gotta eat, so I don’t mind having a floor (though of course I also bypass it with my Beemium all the time anyway).

I actually like the “user-hostile” stepping down one level at a time. I step down all my goals all the time. Basically if I can get one to $0, that’s a freebie; doing it this way lets me set more ambitious goals with occasional acceptable failures. The stepping down one level at a time makes it much harder to keep all my goals at $0 all the time. However, if you added automatic pledge decaying, I’d probably use that to accomplish the same thing and stop stepping down my goals manually at all, so it wouldn’t be a big deal for me. You’re also right that it’s illogical; you could always just archive the goal and start a new one with the lowest pledge you’re allowed to use, instead, so all it’s doing is making something you could accomplish anyway more difficult. I just used it as such, since intentionally making certain things (not sticking to your goals) more difficult is basically the point of Beeminder in the first place :slight_smile:

2 Likes

I agree with this. Pledge caps should be set right away. Users need the comfort of knowing that they will not have to pay more than X no matter what, and just leaving it hanging is anxiety-producing…

Me too. Though I think after a derailment, or two, is a reasonable place to say “hey, let’s talk about pledge caps now you have more idea of how Beeminder works and what it’s value to you is”. There doesn’t have to be a single correct point in the interface to work out what it should be.

3 Likes

Proportional pledge
There is one thing that nobody mentioned so far. I’d love to see a way to be charged on what is done.
Let’s take the very common goal of doing daily pushups, let’s say 100 pushups/day. Should you be charged the same amount whether you did none of them or 95 of them ? In both cases you fail your goal and should be charged.

Another example can be for losing weight, maybe you just can’t make your goal, so you might as well make the most of it and just eat whatever you want, after all punishment will be the same, and it may even make the goal easier to eat more as the value will reset to a higher level. This goes against the goal itself, but if the punishment was proportional to the transgression then you would fail your goal but still stay on your diet and exercise regimen to pay less of a penalty.

For me having this combination of higher pledge but proportionality would be the most effective motivator. It would motivate me to push myself a bit more by setting more ambitious goals, and also by trying to at least do part of the goal even if I know I am going to fail.

Pledge multiplier
Short of that, I’d also like to have the option not to have addition of a specific value instead of a multiplication. I mean $X => $X x 2 => $X x 3 => $X x 4 => $X x 5 instead of $X => $X x 2 => $X x 4 => $X x 8 => $X x 16. Having the option to set the multiplier would already be a progress though, I find the current one a bit to steep for me.

Pledge decay
all for it. Good behavior should be rewarded.
We could even imagine a system where you would gain “good behavior vouchers”. Those would be one-time vouchers that get applied to a goal when you fail it to reduce the price you have to pay. For example, for every goal that you didn’t fail in the last 30days, you get a X% discount voucher, you have a limit of Y amount of them that can be applied at once on a goal. So let’s say X=5%, and Y=4, it would mean that provided you have enough voucher you could get a 20% discount on your pledge as a reward for making your life difficult with many goals to keep track of.

Current setup
The way I currently manage my pledges is by considering the respite time and the cap together. For me the frequency at which I can be charged tend to be more effective than the charge itself, just because it stays “top of mind” and it also shames a but more fore repeatedly failing, it’s not a one time thing anymore but a trend. So being charged 9x10$ is more effective than being charged 1x90$. So I reserve higher pledges for goals that are very long term, and I put lower pledges for smaller more frequent goals.
If I put higher pledges for things that I may fail often, I will definitely weasel my way out of it.

2 Likes

Interesting! Proportional pledges kind of violates the bright line principle. It’s not totally crazy or anything but my gut feeling is that we shouldn’t do this. I think we’d keep it separate from this pledge schedule redesign in any case.

True, and good point, though it’s just one day so maybe that’s fine? Also I’m not sure proportional pledges are well-defined in the case of weight loss. There’s no bound to how much you could exceed your hard cap by, so how would Beeminder compute the fraction of your pledge to charge you? I suppose we could just pick something, like exceeding it by your daily rate or your specified max fluctuation is a 100% derailment and anything less is pro-rated. (But again, we’ll probably set this aside from the current pledge schedule redesign. I love how you’re thinking, btw. It’s possible this idea will end up growing on me!)

Proportional pledges might already be too much complexity and I’m pretty sure that this would be! (Still love how you think though!) Actually I may even have a philosophical objection to this one. It fuzzifies the consequences of derailing and I claim that that always backfires. See blog.beeminder.com/catchup and Beeminder’s No Free Lunch Theorem. That might apply to proportional pledges too. I’m still thinking about that; but I guess we shouldn’t get too side-tracked by these more academic questions, fun as they are!

1 Like

I’d like to put in a request to allow small pledges. They’re perfect for things like this, where you want to be charged on what is done. That way, you could set up, say, 10 $1 push-up goals, one for each 10 push ups you do. @xeltros

I am not as concerned with the minutia of how the pledge progression works - my personal view is that people should just be able to freely choose their pledge amount each time, rather than be limited to discrete choices or have to go through a progression.

My main concern is that $1 pledges should be available, at least for Beemium. For large projects where you have many goals, it would be nice to be able to have them all at $1.

1 Like