change the tip structure

[chadwhitacre]

Tips started here:

   Per Week  Per Year (x 52)   Per Month (÷ 12)
   ---------------------------------------------------------
$    0.08          4.16              0.35
$    0.16          8.32              0.69
$    0.32         16.64              1.39
$    0.64         33.28              2.77
$    1.28         66.56              5.55

Then they went here (ticket is #95; here is the change announcement on the blog):

   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

Now the sentiment on this HN thread is that they should go even higher still. The specific asks are:

• raise the minimum tip from $0.25 to $1.00
• have more options in the $4 to $24 per month range

[chadwhitacre]

I count +4 on HN atm. Maybe +5.

[singpolyma]

I see no benefit to raising the minimum. People can always choose more.

Unless it's costing you money. Then, sure.

[mythz]

25c shouldn't give the impression that you're "supporting the project", if it's not going to be raised, the min should be configurable per user.

[ricardobeat]

@singpolyma those $0.25 add up to $13/year, and the next option is at $156/year. Maybe it shouldn't be removed, $1/month is a decent amount if you want to support multiple devs, but I think there is room for middle tiers.

There is the problem of perception too: $1/month sounds ok, but $0.25 feels like pissing in the wind.

[ghost]

At 25c the transaction fee for me is higher than the tip. $1 is pushing it, but better

[chadwhitacre]

The transaction fee problem can be solved separately; see #167.

[chadwhitacre]

+1
+1

[chadwhitacre]

Counter-position (HN):

Should weekly tips only be given to people whose work I use weekly? If so, then yes, I agree that $0.25/week is not a good tip.

But what about the person whose Emacs config files I read at GitHub and found somewhat informative? And what about the person whose personal wiki (e.g., Org Mode files) I read at GitHub and found somewhat informative? Should I give these people good one-time tips instead of small weekly tips? Maybe I should, but I think I'd like to have the option to give them small weekly tips. Or is there some reason I shouldn't have this option?

[chadwhitacre]

I'd like to see someone address the counter-position above.

I hear a lot of concern about the $0.25 minimum, without noticing the $24 maximum.

The "need more options between $4 and $24 a month" argument is separate and interesting.

[pbiggar]

I commented on HN, following up here. Basically, $0.25 allows people feel like they're supporting someone, without actually supporting someone. Fuck em. Make it a dollar. Hell, make it $2, or even $5.

Better yet, test it. Show a $0.25 minimum to 50% of people, and a $1 minimum to others. Pick the one which helps your goals the most. Repeat as needed. I'll wager you make 2x the revenue (and help 2x as many people) by increasing the minimum significantly.

[agravier]

What pbiggar said, but test it again later, in a few months, and again later, when your userbase will be larger and less biased by the proportion of HNers / early adopters.

[chrismdp]

I think a $1 option is definitely a great idea, whether or not we remove the $0.25 option.

Also a fan of testing it: perhaps 50% of newly created accounts get a slightly different version for a week or two?

[pbiggar]

I wouldn't jump to add $1 if you don't remove the 0.25 option. What's likely to happen? A portion of the $3 will drop down, and a portion of the $0.25 will move up. It's likely that more $3 people will drop down, but even if they were the same you'd be losing $2 to gain $0.75.

[jurre]

I'm sorry if this is answered somewhere else but why are there fixed amounts and why can't I decide to give someone $2.56 a week if I'd want to?

[chadwhitacre]

@jurre Simplicity. Here's the clearest articulation (#95):

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.

[chadwhitacre]

I'm considering running a test with these amounts:

   Per Week  Per Year (x 52)   Per Month (÷ 12)
   ---------------------------------------------------------
$     1             52                 4.33
$     2            104                 8.67
$     5            260                21.67
$    10            520                43.33
$    20           1040                86.67

That raises the bottom and lowers the rest. Remember that we're trying to maximize total throughput.

I'm also considering simply raising the floor:

   Per Week  Per Year (x 52)   Per Month (÷ 12)
   ---------------------------------------------------------
$     1             52                 4.33
$     3            156                13
$     6            312                26
$    12            624                52
$    24           1248               104

Current, for comparison:

   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

Running a simultaneous A/B test where we show 50% of users one thing and 50% another is too much for me to pull off right now. Instead my plan would be to do like we did with the first change, and change the amounts for everyone, and ask people to pick again. We have historical data on tip amounts so we should be able to make meaningful comparisons.

[ricardobeat]

@whit537 you could draw some inspiration from Humble Bundle, their stats display provides good guides on how much to spend. It's not live now, but it starts with an input filled for $9 (or something around the average) which you can adjust typing or using a slider.

I don't see why you would reduce the higher amounts, the gaps will remain roughly the same but you'll reduce totals.

Also, just a hunch, but in my mind $3/month is a sweet spot. I continue to pay that amount to Grooveshark just for support without even thinking, even though I rarely use it (since it's banned everywhere).

[chadwhitacre]

Hmmm ... why limit ourselves to six buttons? What if we just add a whole bunch more options? Like so:

   Per Week  Per Year (x 52)   Per Month (÷ 12)    New
   ---------------------------------------------------------------------------
$     0.25          13                 1.08
$     1             52                 4.33         ✓
$     2            104                 8.67         ✓
$     3            156                13
$     4            208                17.33         ✓
$     5            260                21.67         ✓
$     6            312                26
$     9            468                39            ✓
$    12            624                52
$    18            936                78            ✓
$    24           1248               104

Would be interesting to watch the tip distribution curve fill in.

[pbiggar]

Choice forces a decision onto people. By restricting choices and making people fit into a box, you're freeing them from decision making. (I know this sounds stupid to developers, but it's pretty common in the SAAS world, where you typically see 3-4 plans, not 100.)

People most often choose the lowest or second lowest price. By bringing up the cost of those plans, you increase total revenue significantly. If you add a $1 plan between $0.25 and $3, lots of people who would choose $3 will choose $1 instead.

[chadwhitacre]

@pbiggar You're right.

[chadwhitacre]

This ticket plays into the bigger question of how to model the relationship between individuals and groups on Gittip (#27). If individuals are pulling residual, long-tail gifts, let that be because they are idenfitied with larger projects / groups / brands. Let's let direct gifts to individuals be for true supporters.

Thanks for your feedback, everyone. :-)

[alecperkins]

A little late, but some data relating to presenting choices of tip amounts and backing up exactly what @pbiggar said: The $144,146,165 Button

[chadwhitacre]

@alecperkins Good look.

[chmullig]

FWIW, I was and am in opposition to the $1 minimum (and I was and am in opposition to the $.08 minimum).

There are a ton of people I don't know anything about, but they seem to be doing good work. I'm happy to buy them a couple drinks a year for their seeming support of some software I use.

Like @jcheng5 on https://github.com/rstudio/rstudio/commits/master. He seems to commit to the project occasionally and be involved. I'd like to give him a token tip to help him out. However I don't know anything about him, nor do I care to. For me, giving him $13/year is totally reasonable. He's not important to me, but seems to be doing good work. @jjallaire on that project seems crucial though, so he might be worth $1 or $3. But jcheng5 seems like a minor contributor to a project that isn't that important to me.

So I'm looking at it from the POV of a donor. I'd rather give small tips to four times as many minor contributors, than have to pick just a quarter that number to give larger tips. That's more likely to result in me saying "ya know, they're just not worth it." The core handful who I'd heavily contribute to aren't impacted by the minimum anyway.

[lyndsysimon]

I agree with @pbiggar - this isn't really a decision that's best made by instinct. We need data. :)

[chadwhitacre]

@chmullig Tips to projects (#27) would address your use case, I think. Give $1 to a project with 10 people, and let Gittip distribute it "equitably."

@lyndsysimon I think we don't have enough users yet to get good data. Once we grow another 10x or two we'll be able to run much more meaningful tests.

[bitdancer]

There are people I'm tipping at 0.25. I'm not willing to up that to 1.00 (and no I won't go into the reasons for that), so if forced to change my tip they'd not get anything. I'm happy for you to make this data driven, but do consider the point that by raising the minimum you might increase the total throughput while eliminating the money going to people who otherwise would get the smaller tips, for whatever reason.

[jbuedel]

I'd like to weigh in the on upping the minimum to $1. I literally logged on to gittip in order to start tipping one of my favorite projects but decided against it when I realized I would have to tip $52 a year minimum. $13/yr was a better starting point for me. I believe that a $1 minimum will drive away tippers.

To the folks that say $.25 does not qualify as "supporting" the project: About a year ago I bought some commercial software (LinqPad and RubyMine) and paid ~$40 and $34.50 respectively. Next year this time I will have used them long enough that the cost to me will approach $.25 a week. Am I not supporting those projects? I think that I am.

Lastly I'd like to point out that I do some unrelated giving to various organizations. My pattern is that I tend to start with small amounts, and increase over time as I become more comfortable with it I guess. For example, I was giving $50 a year to an organization at my former university. I upped that to $10 a month at some point, then $20, and now give about $1200 a year. I think it is reasonable to expect that my gittips would follow a similar pattern. I appreciate that the $.25 option allows me to start small.

[chadwhitacre]

There are a bunch of people I still have at the $0.25 level as well. By contrast, I didn't think twice going from $0.08 to $0.25.

[chrismdp]

I think the $1 minimum is probably hurting Gittip. I feel much better about tipping 4 people $0.25 than I do about tipping 1 person $1. I think that $0.25 tips help behind the sceners, too (see #217).

[bitdancer]

I agree that I had no resistance going from 0.8 to 0.25.

[dusty-phillips]

I wonder if a better plan, wouldn't be to have each user say how much they want to tip in total per week, and then allow them to divide that as they see fit between members. I'm inclined to tip about the same amount per week, so at $1 I'm tipping fewer people, although gittip is moving the same amount of money.

At a $1 minimum, we are supporting the idea of complete funding for fewer superstars. at $0.25 we are supporting more people, but fewer of them can quit their day jobs. I think the interface should strive to make people more generous (increase the total amount they put into gittip), not to make certain people more funded.

I think an ideal interface would be:

• in my profile some way to set how much I want to tip per week (in total).
• a slider on other member's profiles allowing me to allocate a percentage of my total weekly contributions to a specific user.

In addition, my profile view should indicate the difference between my incoming and outgoing tips.

[chadwhitacre]

I'm turning the 25¢ button back on for January:

http://blog.gittip.com/post/40015064388/25-growth-special