determine if Gittip is viable for personal funding

[chadwhitacre]

I'm not interested in games here. To me, Flattr is a game. If it weren't, the people building it would be making their livelihood from Flattr itself, not from their 10% cut. Flattr is a game.

My proposal is that to make Gittip viable for personal funding we need to bump the tip amounts an order of magnitude, keeping a single "token amount" option.

---

Original: need to increase tip amounts an order of magnitude

I want people to make their living on Gittip, but the current tip amounts are too small for this.

$52,000 per year == $1,000 per week

If the average tip amount is 32¢ then you need 3,125 tippers to hit $1,000 per week. That's too many people. Think about the numbers you see in followers and friends on Twitter and Facebook. We need to optimize for Dunbar's number: 150.

$1,000 from 150 people is $6.67 each. We need to up our tips by an order of magnitude and then some:

$0.08 => $ 1.60
$0.16 => $ 3.20
$0.32 => $ 6.40
$0.64 => $12.80
$1.28 => $25.60

The current average tip is $0.47 to five other people. Let's surmise that options presented affect the average by setting people's expectations of what they're "supposed to" tip, that people basically distribute their tips evenly across the available spectrum of tip amounts. We have to assume also, though, that upping the amounts is going to reduce the number of people tipping. Let's pretend that the difference between the $0.47 that people currently average and the $0.32 that is the average of the options can buffer some of the drop we'll see. The other thing is that $52,000 is above the US average of $42,000. But there's health care, of course, and also that number is from 2005. Meh, I think this is a sensible next step, anyway.

I believe the $25.60 per week option would satisfy @jacobian's use case of tipping $100 a month to 10 people (#76).

[chadwhitacre]

Existing tips would be preserved, but if someone wanted to change a tip they would have to go to one of the new options.

[chadwhitacre]

The risk I see is that this would flatten our growth to the point of death.

[chadwhitacre]

Of course, if people can't make a living on Gittip then it's as good as dead to me, anyway. :-)

[synapticarbors]

I think it would be better, if you are going to increase the tip ranges, to either:

1. Add a box to take custom amounts to collect data about what people are willing to give, while still trying to skew the expectation.
2. Have a button that either increases or decreases the suggested amount on the buttons by some fixed factor. This maintains the increased expectation, but allows tippers an out that is discrete in nature (like the original idea of the fixed-amount tip buttons).

[chmullig]

(As always, feel free to ignore me. I'm not planning on putting much, if any, money into gittip anytime soon, nor can I imagine any circumstance where people would tip me).

I have two divergent feelings:

1. Yes, some people should definitely be able to make a living off gittip
2. Most people won't and shouldn't.

The reality is most people aren't contributing so much that they should be making any substantial amount of money off gittip. Some will get a beer or two every once in a while. Others might get an annual laptop upgrade paid for. Others may get a large chunk of living expenses, and some might find it their primary income source. I think gittip should honor these asymmetric situations. The list of people I'd be willing to tip $1 a month to is pretty long, the list of people I'd tip $25 a week to is quite short (namely, zero).

I like @synapticarbors suggestion of custom amounts. Let me tip $100/week to someone if I want to, but certainly don't disallow a quarter a week to some new developer that I want to give a token amount to.

[chadwhitacre]

@chmullig You're right that there's a range of possibilities that need to be considered, and you've described them well. It should be possible to make a living on Gittip but that needn't be the modal use case. Question: Is the low end, $1.60 per week, too high?

@synapticarbors Well, I'm reluctant to make it too complicated, and I think the distribution I just put up on the stats page gives us some interesting data about what people are willing to give (see below):

Amount ($)   %   No.
--------------------
   0.08    27.9  212
   0.16    15.3  116
   0.32    15.0  114
   0.64    12.2   93
   1.28    29.6  225

What I'm seeing is a bimodal distribution, where we've got a bunch of tips that could probably be even higher than $1.28 a week, and a second bunch of tips that want to hover down in the range of the "token amount." For IHasAMoney.com I figured the "token amount" was $2.99 a month (I tried to price it so it felt free), which is about 75¢ a week. @chmullig suggests 25¢, or $1 per month. Is $1.60 per week, $6.40 per month, too high for a "token amount"? I think so. Maybe we drop the bottom down? How about this:

   Per Week  Per Year (x 52)   Per Month (÷ 12)
   ---------------------------------------------------------
$     0.25          13                 1.08
$     3            156                13
$     6            312                26
$    12            624                52
$    24           1248               104

This keeps us well under the $13,000 per year where we start to get into tax implications in the US (#96). It gives people an option who are looking to tip a token amount, and a range of options to people interested in more significant support, thereby steering people towards that option and counteracting the natural gravity towards zero.

For the record, I like the multiples in the higher amounts because it helps with anonymity in the lower reaches. If I receive $24 a week, do I have one, two, three, or four supporters? Or 96? :-)

[chadwhitacre]

I'm rescheduling #22 for next week to make room for this ticket this week. I'm starting to see this ticket as a critical moment for Gittip: I don't have bandwidth for or interest in Gittip as an eternal side project. I need it to look like it can provide an income for me in the foreseeable future, or I'll have to hand it off or wind it down.

I'm not interested in games here. To me, Flattr is a game. If it weren't, the people building it would be making their livelihood from Flattr itself, not from their 10% cut. Flattr is a game.

I'm interested in input from @jtauber, @alex, @jezdez, and anyone else interested in whether Gittip is viable for personal funding.

[chadwhitacre]

Copied some of the previous comment to the top and renamed the ticket. Going to call for input on Twitter.

[alex]

I want gittip to be viable for personal funding, and I agree that higher amounts can help towards that. But I also think you need to look at better marketting and conversion. ATM there are under 1000 users and only 52 with a credit card. Getting those two numbers up is the number one thing to focus on. Github has over a million users, stay focused on developers and target that market aggressively.

[chadwhitacre]

@alex What are we marketing? A fun game, or something more?

[alex]

Gittip should be two things:

1) An ability to give back, to say thanks.

2) An ability to help make the next great open source project.

We're all users of dozens of great projects, and if you ask the creators of any of them what they're missing to make it better, or to create the next thing, the answer is always time. Gittip is a way to try to give them that.

[jtauber]

• I want gittip to be viable for personal funding too and am starting to explicitly ask people to support me via it
• I think it's very early days and initial growth has been good so I think it's a little too early to worry it's just an eternal side project
• I agree with @alex that the focus should remain on developers on Github
• I also think the ability to tip an order of magnitude more is useful (and necessary)

[chadwhitacre]

The strongest marketing for Gittip is when Alex Gaynor and James Tauber and Jannis Leidel say at conferences and on the Internet, "I make my living on Gittip." If I hear that, I'm like, "Oh, wow! That must be serious."

So it's chicken and egg.

@alex and @jtauber: You each have ~4,300 followers on Twitter. The math needs to be more like:

(300 * $5) + (2,000 * $0.25) = $2,000 / wk

And less like:

4,300 * $0.47 = $2,000 / wk

Tippers will be some fraction of Twitter followers. The data above shows that a) a good number of people are willing to give more, and b) we still need a "token" option.

[chadwhitacre]

I think it's very early days and initial growth has been good so I think it's a little too early to worry it's just an eternal side project

Fair enough. I'm at a point where I need to a) spend less time on Gittip, or b) see more promise in it. Trying to navigate that. I've removed some of the "sky is falling" from the description.

[jtauber]

I'm at a point where I need to a) spend less time on Gittip, or b) receive more money from it. Trying to navigate that.

Oh I hear you on that.

[therabidbanana]

Definitely think allowing bigger tips is necessary to make it a viable way to make a living, and I definitely think that making it a viable way to make a living is an awesome goal. I'd love to see developers making a full-time income working on their open source projects without having to be sponsored by big companies that can afford to do that.

Maybe if you need to verify that there's promise in the idea, use a more established crowd-sourcing platform like Kickstarter that allows a set goal with deadline to try to get enough money to focus on building it full time for a few months? I'd certainly be willing to donate a bigger amount to try and kickstart the gittip economy. :D

[chadwhitacre]

I changed the tip distribution query to only include backed tips, and the distribution is flatter, actually with a bump in the middle. Old:

Amount ($)   %   No.
--------------------
   0.08    27.9  212
   0.16    15.3  116
   0.32    15.0  114
   0.64    12.2   93
   1.28    29.6  225

New:

Amount ($)   %   No.
--------------------
   0.08    23.6   97
   0.16    17.3   71
   0.32    19.2   79
   0.64    16.8   69
   1.28    23.1   95

The data doesn't seem as encouraging as I had thought.

[alex]

Another idea: stop allowing unbacked tips. When you try to use that part of the UI, redirect to the add a credit card.

[chadwhitacre]

I'd love to see developers making a full-time income working on their open source projects without having to be sponsored by big companies that can afford to do that.

+1

Maybe if you need to verify that there's promise in the idea, use a more established crowd-sourcing platform like Kickstarter that allows a set goal with deadline to try to get enough money to focus on building it full time for a few months?

Hmmm, will have to chew on that one ... is it going to be easier to raise $12,000 ($4,000 x 3 months) on Kickstarter, or find 80 tippers at $12 a week on Gittip? Got to market it one way or another. If the (time) cost of marketing is the same in both cases, then having the sustainable income already in hand seems better than having to subsequently raise the 80 x $12 again anyway. The ultimate goal is the 80 x $12. Not clear to me yet that Kickstarter adds enough value as a proximate goal, but maybe?

[chadwhitacre]

Another idea: stop allowing unbacked tips. When you try to use that part of the UI, redirect to the add a credit card.

Right now you get a highlighted "Back your tip with a credit card!" link inline when you try to tip unbacked.

I can see not recording the unbacked tip. That could simplify several queries. The downside would be that the user wouldn't have those tips waiting for them on their profile page to remind them who they wanted to tip.

Reticketed as #97.

[therabidbanana]

The advantage to the kickstarter method is that everyone pays or nobody does - it makes you feel more confident about contributing. If only ten people pay $12/week, and that is not actually enough to sustain gittip, so the project dies out, then the people that went in have the feeling they have thrown away their money. I think a possible reason that more people haven't backed their tips is that they want to wait and see if gittip actually becomes reasonably viable before actually kicking in - which keeps it from being viable... a painful cycle to break out of, and exactly the kind of thing Kickstarter was made for.

Another open source project on kickstarter that was essentially "give me money to work on this full time for a few months" and was reasonably successful: http://www.kickstarter.com/projects/joeyh/git-annex-assistant-like-dropbox-but-with-your-own/

[matin]

As an MVP, you could add a field for a custom amount. You can modify fixed amounts based on the custom amounts that you're seeing.

[chadwhitacre]

@mtamizi An MVP is someone receiving a lot through Gittip? Not a bad idea.

[alex]

I think he meant MVP as Minimum Viable Product.

[chadwhitacre]

@therabidbanana The thing is, if I do the Kickstarter, I still need to then go and raise the 80 x $12 in order to keep going, or I'm right back where I started, just with even more momentum behind the site (= pressure to make it work "somehow"). On the other hand, Kickstarters are known quantity by now and are kind of hot, even. They might lend some credence to Gittip by association. Also, one can look at it the other way: if I were to get 500 backers on Kickstarter it sure seems like I should be able to convert 200 of them into gittippers.

@mtamizi, @alex Got it. I'm still reluctant to complicate the app with a custom amount box, but I've heard that suggestion a few times now (also from @synapticarbors, e.g.). Especially if I hear it a couple more times I'll probably go ahead with it.

[matin]

The value of gittip to me is that it's recurring. I'd rather use gittip to pay people who are actively maintaining projects instead of creating a campaign for new projects.

If you're trying to supplement their income, weekly payments instead of one time payments makes the most sense.

[chadwhitacre]

Kickstarter preview:

http://www.kickstarter.com/projects/whit537/1818838801?token=59ef74a4

[chadwhitacre]

@mtamizi Yeah, Gittip is focused on recurring payments. There was an early ticket (#5) for one-off payments but it was tabled.

[bruceadams]

I agree that the maximum tip amount needs to be higher than $1.28. (All of my tips are given at that maximum.) I really like the existing small set of tip amount buttons. I think keeping the amount choices limited makes it much easier to choose one. Making people type in an amount is a hurtle. An empty type-in box does not give any indication of what is typical within the Gittip world.

I'd like to suggest the same number of buttons, with each amount being four times as much as the previous (instead of the current doubling).

This sequence gives round dollar numbers: (0.00) (0.06) (0.25) (1.00) (4.00) (16.00)

This sequence continues with the geeky power of two theme of the current tip amount options: (0.00) (0.16) (0.64) (2.56) (10.24) (40.96)

I like the second of these, the power of two sequence. I haven't figured out how to articulate why.

This allows people to participate for very small amounts, increases the upper amounts and still presents an easily comprehended short list of choices. (On a related note, by having a maximum tip amount, no single giver can have an overly large impact on a recipient).

[matin]

On a related note, by having a maximum tip amount, no single giver can have an overly large impact on a recipient

This is a shockingly good point. If the point is to supplement the tipee's income, then you don't want it all to be driven by one person. It's much more reliable when it comes from many sources.

[chadwhitacre]

Options:

1. (0.00) (0.06) (0.25) (1.00) (4.00) (16.00)
2. (0.00) (0.16) (0.64) (2.56) (10.24) (40.96)
3. (0.00) (0.25) (3.00) (6.00) (12.00) (24.00)

I was just implementing the third option. @bruceadams Did you notice the comment way above where I floated it? Were you intentionally suggesting alternatives to it, or do you think that would accomplish our goals?

The reason I like option three is that it clearly differentiates between those who want to give a token amount (0.25) and those who are more serious. That is, I like the gap between 0.25 and 3.00. I think it will be a nice statistic to give tippees: number of tippers at $0.25, and then the N and average tip from the higher four options. Like this:

1,244 people tip you $0.25. 95 people tip you $3 or more (averaging $5.23).

Starting the doubling at $3 yields a higher maximum, $24, than starting the doubling under a dollar. The $40.96 maximum feels too high to me.

Something to keep in mind, too, is that non-programmers aren't going to get the powers of two. 3, 6, 12, 24 will be more accessible beyond GitHub.

[bruceadams]

@whit537 I had missed your proposal. I was not attempting to propose something only slightly different. I just wanted to be concrete about what I thought was good and would be good. I think (0.00) (0.25) (3.00) (6.00) (12.00) (24.00) looks great!

[chadwhitacre]

Cool, thanks @bruceadams. :-)

[chadwhitacre]

Related IRC snippet:

[12:39pm] whit537: You're good with those amounts, then?
[12:39pm] whit537: 0.25, 3, 6, 12, 24?
[12:39pm] whit537: Tops out at ~$100/mo.
[12:40pm] whit537: Or did you really want the $40.96 option? :-)
[12:41pm] bruceadams: I wasn't sure what the top should be. Having one slightly over the top 
                      number might encourage people to go with the second highest more.
[12:41pm] whit537: Good point.
[12:41pm] bruceadams: I'm just a little concerned, in your set, with the gap between 0.25 and
                      3.00. that's more than a factor of ten.
[12:42pm] whit537: Maybe I should make it 0.50.
[12:42pm] whit537: Well, no.
[12:42pm] whit537: The average today is about $0.47.
[12:42pm] bruceadams: I do think it's important to have a very low end, like 0.25.
[12:42pm] whit537: I like the gap between 0.25 and 3.00, because it makes it clear who's who.
[12:43pm] whit537: It leans into the long tail.
[12:43pm] whit537: On both sides of the equation.
[12:43pm] whit537: "Here's the people I'm really committed to, and then here's the whole mess
                   of people I tip a quarter."
[12:44pm] whit537: Streamlines the decision to tip someone just a little bit.
[12:44pm] bruceadams: that makes sense
[12:44pm] whit537: Otherwise you're like, "Well, I want to tip them a little bit, but do I want 
                   to tip them 16 cents or 32?"
[12:45pm] bruceadams: right, 0.16 or 0.32 is a pretty useless choice.
[12:45pm] whit537: That sort of extra decision makes sense for the people you really care about.
[12:45pm] whit537: 3 vs. 6 vs. 12 vs. 24
[12:45pm] whit537: That's worth an extra few seconds or so of decision.
[12:45pm] whit537: But if you've already decided this person is part of your own long tail then,
                   yeah, let's make it easy.

[chadwhitacre]

Okay, folks. Thanks for all the input! Let's see if this dog can hunt!

P.S. @therabidbanana Kickstarter reticketed as #98.

[DamosDaze]

First comment on this project (looking forward to Hangout w/ Chad on Monday). Gittip has so much potential it's crazy. :)

I think the tip structure could be simplified greatly while increasing the average tip AND incorporating the one-off tip option. Do this by making the default tip $1 (a super-low number everyone is used to) and provide a variable in frequency of giving, ie. one-time, daily, weekly, biweekly (every paycheck), or monthly. This makes it easy to give $1, $2, $4, or ~$30 per month and you can always add the custom gift option for larger gifts (both one-off and recurring).

(btw +1 to @bruceadams for "by having a maximum tip amount, no single giver can have an overly large impact on a recipient." This is so important.)

One significant related issue is the anonymous giving thing. IMO anonymous giving should be AN option, not the ONLY option. This is huge for several reasons. With public giving you can add a share option to post on fb, twitter, etc. so we can leverage our networks to tell everyone how awesome the people we're giving to are. The ripple effect will increase the number of givers by enabling the viral factor. It's also important b/c we're all in theory giving and receiving through gittip, so it's not just about the awesome things we're doing that makes us deserving of tips but also the awesome who's and what's we're giving to that makes us worthy.

With public giving we get more data to inform our own giving decisions, and can start building a much needed "social credit score."

I'm tempted to get into the option to define specific item/project goals (which I think is soooo important) but I'll hold my tongue for now.

[chadwhitacre]

After a year, I'm reopening this ticket, because we still haven't actually proven that Gittip is viable for personal funding yet. But we're close! @alexpott and I are both up over $400/wk and are on our way to $500/wk, which is some sort of watershed. I think we might have enough juice to go for it:

Hey @alexpott, I'll race you to $475. :-) I just updated my @Gittip goal to match: https://www.gittip.com/whit537/ .

https://twitter.com/whit537/status/370707205743337473

Here's Alex: https://www.gittip.com/alexpott/

[rummik]

The poverty level for the US is $23,550/yr for a family of four, $455/wk passes that. And for an individual it's $11,490/yr, so about $225/wk would reach that. I'm not sure what should be considered 'viable'... Maybe twice what poverty level is? Seems a bit lower than what most people might expect, but it is a baseline of sorts.

[chadwhitacre]

Yeah, @alexpott's goal of $475/wk is actually lower than what he needs to live on, per his blog post:

My short-term aim is to make it to January working 100% on Drupal core. To find out why I think this is necessary read http://xjm.drupalgardens.com/blog/triaging-drupal-8. So, I'm going to reduce some of my outgoings temporarily and invest another £6000 of my own money. However, based on this I will still have a shortfall of £5350 which is equivalent to about $475 a week. So that's my target with gittip.

I think we want to target mid-range salary equivalents, rather than the low end.

[rummik]

Hm. Probably. :P I might be thinking a little too much in terms of Middle of Nowhere, South Carolina -- viable funding in LA, or any other city, would be much higher.

[chadwhitacre]

@eric-s-raymond is interested in running an experiment on this. IRC

[chadwhitacre]

We now have three people depending on Gittip for primary income.

https://twitter.com/Gittip/status/396351785524158464

[chadwhitacre]

In light of that, I'm about ready to close this.

[chadwhitacre]

:horse: