After reaching the goal the mechanism has to behave somehow - the existing rules can’t just keep the same. So an “extra” rule has to be in place either way.
2 Appreciations
It is a rough animation. As I said it would have to take into account how the joining of each patron is a discrete step, but remember my example is only showing very few patrons. At the end of the day it does not matter that much. Pause the video and move the time slider to the center and you’ll see that 50% is 50%, at 20% it is 20%, … . I actually just cloned a single shape and used it as patrons and goals on both sides.
2 Appreciations
Oh oh! I get it now. The point of the difference in the animation is that each patron can only push the goal ahead the same amount regardless of their pledge!
So, the biggest pledge in the crowd-goal version actually would just be the same width as the others if this were less-rough, right? It’s not supposed to show different widths, that’s an artifact of quick rough version, right?
2 Appreciations
it does not matter in what way difference in size is shown really. could be height, or width or both. I just went with the exact same shape scaled in both dimensions. It is all about proportions, and those remain the same.
1 Appreciation
“sufficient simplicity” = people easily understand and relate to it, so complexity of the mechanism does not act as a barrier to potentially interested people becoming patrons.
“sufficient flexibility” = people can participate in a way that feels meaningful and attractive to them, which will differ greatly between different people in relation to different projects.
Those properties matter because the success of crowdmatching hinges crucially on it being attractive enough to very large numbers of people who currently don’t donate at all that actual large crowds of patrons form. If either insufficient simplicity or insufficient flexibility prevent it from attracting large numbers of patrons, it will not succeed.
What makes me skeptical about crowd size being part of the math is:
- It can only be simple if everyone donates the same. As soon as you try to add options to donate more, it becomes complex. In the particular way of doing this you’ve proposed, actually only the minimum donation is matched by others. Let’s say the minimum is $1 but that feels way too low to me and I want to donate $10. Then what the rest of the crowd donates is only affected by my first $1 since it’s the same whether I donate $1 or $10. But what I donate is affected by the size of the crowd, so if the crowd is 50% toward the crowd size goal, I’ll donate $5, etc. That is complex to understand.
- Even that flexibility, at a significant cost of complexity, doesn’t really add the flexibility I think we need, since having only the first $1 of my participation matched by the crowd hugely reduces the motivational power of crowdmatching. I don’t think it’s genuine flexibility. It’s really just a slightly different framing of our current one-donation-level-fits-all model with an overlaid option to donate more and match the crowd size, but without being matched in return. $9 of my $10 pledge is not being matched, so how different is this scenario to just asking me to donate a monthly amount without crowdmatching at all?
- A crowd size goal seems arbitrary and doesn’t feel to me like it gives me much reason to participate vs. not having a goal at all. I think if Snowdrift decides to stick with matching people rather than willingness to donate, then I’m not sure that adding a goal helps much, and might actually hurt by adding complexity and vagueness, i.e. no clear answer to the question “why should I get excited about helping to achieve that particular crowd size?”
4 Appreciations
First of all THANK YOU for perfectly addressing my questions. I really do appreciate that.
Concerning flexibility and simplicity we actually seem to be on the exact same page, I value those for the same reasons you do. I second your dedication to underline the easy understanding is key to participation is something we absolutely need for our success.
-
About complexity.
In a recent discussion with David I think finally understood your point here. If I’m not mistaken the problem is best described when we compare the event of a single patron joining – and then comparing the outcome in both cases. Yes, there are differences, and they may have an impact on an attempt to understand matching. But that’s only problematic when there is a preset framing that search for one special variation of matching behavior. The actual matching that does happen and that can be found is actually easy to understand and has the same overall identical impact. I think, if we can assume that the framing fits the actual matching from the get go – there is no problem at all. -
About flexibility
I think you are misunderstanding the crowd-goal variation of the proposal. Matching does happen and to the same extent - just in a different way. Nothing keeps a patron from applying the exact same degree of flexibility in matching. If any portion of any pledge would get less matched there would have to be a difference in the final donation. Again, with a skewed framing it might come along wrong – but that’s what we have to take care of: make sure that the understanding of the mechanism fits how the mechanism really works. -
About the crowd size goal
This is a problem in and of itself that I think we are having problems to solve in both cases. Even if a hard number expressed in currency is something we tend to understand better than a certain number of people it still begs the question: “How arbitrary is it?” The way the mechanism currently is envisioned creates an incentive to set a realistic goal, not a desired goal. And it is subject to be changed all the time – unless there actually is some specific amount of money that embodies a more or less permanent desired end goal. Which I doubt is ever the case. No project can make more or less use with $3989, $4000 or $4011 – especially when the said mark only represents a wish, not the an actual amount of money.
I think relevant discussion to that topic is expected to happen here: Goal adjustment – terminology – framing
1 Appreciation
On that point, your $10 is still there as matching incentive to others. One new patron choosing to pledge gets more dollars out of you in matching than they would if you only pledged $1. So, it’s not the same as giving without crowdmatching.
2 Appreciations
Besides the value of a goal, even though arbitrary, it has one core benefit: that’s the point at which the budget limit kicks into a different behavior. So, the main point is that it’s the budget limit. It’s a statement of “I’m willing to crowdmatch up to a crowd of this size”.
So, yes, it’s like our current model + option of different size pledges + better per-project budget cap approach.
1 Appreciation
Yes, but it only incents them to pledge $1. (And it feels to me like this discussion itself underline the complexity here.)
3 Appreciations
So, yes, it’s like our current model + option of different size pledges + better per-project budget cap approach.
I’m not sure it’s necessarily a better budget cap approach overall. It is good that it avoids having to auto-drop patrons to avoid exceeding their budget cap, but it does that at the cost of imposing the same crowd size limit on everyone. In the current model, people with higher budget limits can continue matching as the crowd size grows beyond what works for some other patrons.
2 Appreciations
If matching is based on [current crowd size n / crowd size goal G], then when I join as a patron the impact on what the existing crowd donates is the same regardless of whether I pledge $1 or $100. In both cases I increase current crowd size by 1, so if previously they were each donating n/G of their pledge, they are now donating n+1/G of their pledge. So only my first dollar is matched by others (assuming $1 is the minimum pledge), since how much they donate does not increase as I increase my pledge above $1.
If matching is based on [crowd total $ pledged / $ goal], then when I join as a patron the impact on what the existing crowd donates is proportional to what I pledge. If I pledge $100, the increase in what the current crowd donates is 100x what it would have been if I’d pledged $1 because I’ve increased [crowd total $ pledged] by 100x as much as if I’d pledged $1.
4 Appreciations
Estimating what things will cost and budgeting for them is a universal practice that all successful organizations have to do. Of course there’s always some uncertainty involved, but the degree of uncertainty in what could be achieved is not even close to the same in these two scenarios:
-
“If we had $X of reliable funding per month we estimate that we could do Y.”
-
“If we had N people each donating somewhere between $1 and $[upper limit if any] we estimate that we could do Y.”
In both cases the degree of uncertainty about what can be achieved with a given amount of funding is the same, but in #2 you’re multiplying the normal uncertainty of budgeting by a lot of uncertainty about how much funding there will actually be.
3 Appreciations
- Actually this is not entirely true. By joining you become part of the crowd and pay a bigger share of your own pledge compared to joining the crowd at a lower level.
- The effect you notice is correct, but two pronged: at the VERY INSTANT you join, a bigger pledge will have a bigger effect, yes. But every smaller pledge will have a smaller effect, evening it all out – because there can’t be only larger pledges. Crowd based matching just turns this upside down
- Undoubtedly the math adds up in the end and both approaches have an overall identical effect. So, concerning the matching we face the question of which kind of pledge to “favor” (ever so slightly). In that regard I’m actually interested in making smaller pledges more appealing in the moment; it resonates more with our mission to motivate many people with small participation, as opposed to motivating fewer people to give more.
Have you seen the update to the original post with the video? My hope was to make it somewhat more approachable, as getting an authentic feeling for the dynamics isn’t trivial.
It is a rough video that should not consist of frames that show “half” patrons, and nicely snap between them (actually impossible as the video shows both mechanism version at the same time). Also the number and scaling of patrons isn’t correct – but the general amounts do match good enough.
1 Appreciation
- By “existing crowd” I meant the crowd as it exists before I join it, and on that basis it is entirely true.
This can be expressed as this:
-
“If we had $X of reliable funding per month we estimate that we could do Y.”
-
“If we had N people (Ø $1.5) funding us we estimate that we could do Y.”
Ok I agree that’s still more fuzzy, but not more complicated than it needs to be 
. I can see how it is still a benefit, though. In the light of it being the mere wish - or hopeful prediction - where you want to land, what’s the “worth” of a precise but inaccessible number? 
In my eyes this downside does not outweigh the benefits we get by literally counting in humans.
2 Appreciations
Maybe I’m getting you wrong, but I don’t see a reason to expect any impact in or on a crowd before joining.
Extra dollars are only matched in $-based goals
While it’s not a question of before-and-after strictly, that’s just a normal framing of pledge-or-not. The matching question is: How would things go if I pledge vs if I didn’t pledge (or if I pledge more vs pledge the minimum)?
If we assume that a goal that is ambitious and not inevitable:
- With dollar-goal, every dollar I pledge results in extra matching from others
- With crowd-goal, only my participation gets matching and extra dollars do not get matched (though I still am doing extra matching of others)
We can speculate about the effects of this difference on everyone’s motivations. We can discuss whether we want to keep the extra from being matched in order to prefer crowdsize as a goal, but the pure math is pretty clear.
Getting outside feedback, testing our hypotheses
At this point, I think both approaches have some chance of success. But while I might like the more-radical crowd-based idea, I think it comes down to which will actually get more buy-in from projects and patrons. I suggest some intentionally-done outreach to get insights from potential users.
It seems pretty likely that patrons and projects will find it much easier to accept the $-based model as just a logical step beyond things they already know. I suspect that the subset of people who will really be more inspired and enthusiastic about crowd-based goal are those who already have more activist views and experience (but with good framing and messaging about our politics and mission, they are likely to be happy with dollar-goal too).
We need some way to get more evidence to use in this decision. It’s possible that one approach or the other really would be dramatically easier to sell, to explain, to succeed with. If we can figure out that both are truly viable, then we could choose based on which fits our political mission.
FWIW, if I had to decide unilaterally right now with only the current understanding, I would go with dollar-based goals but frame much of the presentation around emphasizing the crowd size. I think there are ways to do that which would not cause confusion at all.
1 Appreciation
I mean the donations by people who were in the crowd before I joined will change by the same amount after I join regardless of whether I pledge $1 or $100.
1 Appreciation
That is exactly my view and what i tried to explain. Well said!
That is the point here!
Is that our main audience to focus on or do they donate already? I think we should clearly define our audience we want to reach before creating the framing.
I think that could be: Users of FLO projects that don’t donate yet. They might not even value FLO but use the program just because it’s good (think about Krita) and might even pay if it’s proprietary, but don’t because it’s gratis. Why pay if you don’t have to?
With such an audience (average people), we can reach our goal to get a big crowd.
The core message should be easy to understand and convince those to contribute.
But our aspiration should be, that the detailed information (mission, values, …) also please FLO activists.
I fully support this! We could implement a demo with both and do A/B testing, where visitors get site A or B at random.
We can then announce, “we are working on a new mechanism to fund FLO, please help to test it” (not saying that it’s actually 2). Not sure if Liberapay would post a blog post or at least share our social media post. I would.
2 Appreciations
OK I get it now, yes that’s true!
It is pushing forward by the amount of an fixed average pledge, instead of the variable own pledge. No approach is favored here, since the average pledge is average 
– (I think it is even likely that most of the pledges are below average). We would enable more people with increased pledging – and “richer donors” have an extra incentive to create the “fertile ground” for others that join (at the cost of having a smaller impact themselves at the moment they join).
1 Appreciation