search

gratipay / inside.gratipay.com

Here lieth a pioneer in open source sustainability. RIP
https://gratipay.news/the-end-cbfba8f50981
57 stars 38 forks source link

get a loan #78

Closed chadwhitacre closed 10 years ago

chadwhitacre commented 10 years ago

Looking at venture capital in #68. How about a loan?

Went to a seminar with Bridgeway Capital today. Gonna try to meet with someone from there next week.

chadwhitacre commented 10 years ago

I'm putting a spreadsheet together:

https://docs.google.com/spreadsheets/d/1zvnAr59LImzeMCRYakhMAAg6MX6JFFcUqq4yM-RNCNw/edit?usp=sharing

Here's a query I used to determine how much Gittip gives out in payroll + gifts each week:

   select substring(timestamp::text from 0 for 8), sum(amount)
     from transfers
    where tippee='Gittip' group by substring(timestamp::text from 0 for 8)
order by substring(timestamp::text from 0 for 8);
chadwhitacre commented 10 years ago

I'm doing two tabs, one without a loan and one with.

dsernst commented 10 years ago

Is there a clear list of what the funds will be spent on?

On Thursday, August 7, 2014, Chad Whitacre notifications@github.com wrote:

I'm doing two tabs, one without a loan and one with.

— Reply to this email directly or view it on GitHub https://github.com/gittip/building.gittip.com/issues/78#issuecomment-51563566 .

chadwhitacre commented 10 years ago

On the "Projection: Loan" sheet I show the funds being put into payroll. I think our bottleneck is people to work on Gittip, and if we get capital we should use it to incentivize people to work on Gittip. What do you think?

dsernst commented 10 years ago

Yes, agree, developer bandwidth is of course super-valuable.

I just wonder in part what the loaner would ask when assessing the risk:

By payroll, do you mean that some of this money is meant to let certain developers completely leave their day jobs?

I do wonder what the effect might be for some contributors not invited to the Funding Party.

Probably helpful to learn how some other leaders approach full-time open-source devs, e.g. Mozilla, Automattic, Canonical, etc

My overall feeling on this matter is that deciding what to spend money on is markedly more important before deciding where to fundraise. (Cf the detailed history of tech startups floundering despite bountiful capital, sometimes even over-capitalization.) Particularly since Gittip seems more of a sustainability-type business, as described in Joel Spolsky's Strategy Letter I: Ben & Jerry's vs Amazon http://www.joelonsoftware.com/articles/fog0000000056.html

Perhaps BountySource could be particularly handy here?

So if there are some explicit "most important" issues, it could make this whole discussion more concrete and clear.

chadwhitacre commented 10 years ago

Do you imagine pumping loan money into the existing Gittip/Gratipay team distribution, then business as usual?

Yes, exactly. As I see it, we've got a chicken and egg problem:

The purpose of taking on debt would be to get out of this dilemma by growing Gratipay to the point where there are enough people deriving value from it that we can be sustained by their direct support. What do you think?

dsernst commented 10 years ago

Ok yes, the benefit for a loan is clearer when you frame it in the chicken-and-egg light.

dsernst commented 10 years ago

Still, I would ask if "growing Gratipay" could be unpacked into more specific & measurable goals?

Examples:

(I'm just making up these numbers for illustrative purposes)

chrisdev commented 10 years ago

@whit537 I found this analysis about Patreon http://research.gigaom.com/2014/02/analysis-patreon-seeing-strong-growth-in-creators-pledges-pageviews/ Which includes this growth chart patreon-grow-in-pledges

chrisdev commented 10 years ago

:+1: to @dsernst suggestion about including additional metrics Although they are implicit in the spreadsheet. You will probably have to do some sort of executive summary for the loan proposal

Changaco commented 10 years ago

Yes, exactly. As I see it, we've got a chicken and egg problem:

  • The Gratipay team's main source of funding will be people who are making a living on Gratipay.

I'm not sure that's true. [Removed non-public information. —@whit537]

chrisdev commented 10 years ago

@whit537 I just noticed the FTE column, is this Full Time Employee? I think the spreadsheet makes a very convincing case for the ability to repay and future viability of Gratipay. However, I think the growth projections past year 3 should be adjusted downwards. Is income for august 2020 projected to be $4,004,137,698,384

chadwhitacre commented 10 years ago

I have a meeting with Bridgeway Capital tomorrow morning at 10am. Turns out a friend of mine is an analyst with them, and he provided this feedback when I linked him to the spreadsheet:

Two issues that you may face - how does an increase in payroll monetarily benefit you? I understand that you will be able to employ another developer, but how soon will Gratipay realize an increase in volume due to the addition? Also, I see that you imply a 35% month over month, is this realistic, or a best case scenario? The best way to win over Bridgeway is to fulfill the mission - employing 1-2 individuals, in a low-income neighborhood. Let me know if you need more advice, but overall, the projections were actually much more thorough than most I've reviewed.

chadwhitacre commented 10 years ago

I just noticed the FTE column, is this Full Time Employee?

@chrisdev Yeah, that divides payroll by the estimate at the top of the column to get a rough estimate of the number of people we can sustainably support.

chadwhitacre commented 10 years ago

However, I think the growth projections past year 3 should be adjusted downwards.

I tried to use a Gompertz curve (you'll see this on the "Assumptions" sheet, and I wrote a gompertz macro that is attached to the workbook somehow), but I couldn't get it to work so I kept the projections to a simple exponential model. How do we achieve this downward adjustment?

chadwhitacre commented 10 years ago

I'm not sure that's true.

Fair enough. We should answer this question with aggregate data, probably in a way that we can easily track over time. An addition to the stats/charts? We should also talk about how to incentivize giving back to Gittip. Both are probably out of scope here, though?

chadwhitacre commented 10 years ago

I guess what's in scope here is understanding what the purpose of taking a loan would be for us. The purpose is to get over the hump identified in Gittip, Year Two, where Gittip clearly works as a platform for making a living, but not yet for us. Our goal is to get Gittip to the point where it is itself sustainably funded. How do we do that? By working on Gittip, continuing to write PRs that close issues, and getting the word out to people who want to use Gittip. Right?

chadwhitacre commented 10 years ago

Whatever the details of our source of revenue (whether our money is coming from receivers that regive to us or from givers that also give to us), the fact is that we haven't yet reached the economy of scale we need to be able to adequately pay the people building Gittip. Like, I'm not making enough money from Gittip. Me, personally. I don't make enough from Gittip to say that Gittip is working for me yet. @Changaco and @seanlinsley are the only others that are remotely trying right now (based on the current team takes). I believe @seanlinsley has a day job, and I think @Changaco also has other work.

I guess what we're faced with is a revenue-per-effort differential. Gittip requires proportionally more effort relative to our revenue at this stage than it will once the business is further along. Right?

chadwhitacre commented 10 years ago

So if we're able to fund the extra effort now, then we'll be able to grow the business to the point where we are taking in more money per unit of effort, and it will be enough more money to comfortably pay back the loan.

dsernst commented 10 years ago

I hear your thinking, but some of the growth may also limited by time more than anything else.

One question: who's legally obligated to the debt? You in a personal fashion? Gittip LLC— which in name is you, but if the debt is defaulted on, someone else would try to seize assets?

On Sunday, August 10, 2014, Chad Whitacre notifications@github.com wrote:

So if we're able to fund the extra effort now, then we'll be able to grow the business to the point where we are taking in more money per unit of effort, and it will be enough more money to comfortably pay back the loan.

— Reply to this email directly or view it on GitHub https://github.com/gittip/building.gittip.com/issues/78#issuecomment-51724965 .

chadwhitacre commented 10 years ago

some of the growth may also limited by time more than anything else.

Right. Money buys us time.

The loan would be to Gittip, LLC.

chadwhitacre commented 10 years ago

I was calculating our growth wrong. Now it looks like we've averaged 12% growth over the past 23 months (with a standard deviation of 10%). During our best run, from January, 2013 through February, 2014 (14 months), we averaged 17% growth (σ = 6%).

chadwhitacre commented 10 years ago

Still, I would ask if "growing Gratipay" could be unpacked into more specific & measurable goals? :+1: to @dsernst suggestion about including additional metrics

My goal is basically for me to make a living via my take from the Gratipay team + my personal receipts - my own giving. I want to focus on growing the income to the Gratipay team so I can take more from there; I don't care if my personal receipts go up (though if they were to go down significantly that'd make a difference to my calculations).

I receive about $520/wk right now and I give $120, so call it $400 for personal. I need about $1,000 per week to live modestly in my town (cf. the U.S. median of $51,017 mentioned in Year Two), so let's say I need to be taking $600 per week from the Gratipay team. I'm not comfortable taking more than 20% of the total income because we need 15-20% for operations/savings and I want to leave plenty of room in the payroll for others. $600 is 20% of $3,000, or $13,000 per (normalized) month.

In Gittip, Year Two, I set a goal of reaching "our first $100,000 week by June 1, 2015." That's $433,333 per (normalized) month. If we maintain the 3% income to volume ratio then our payroll would be ... $13,000! This suggests that reaching our target of moving $100,000/week coincides with my personal target of taking $600/wk, and that the $100,000 volume goal is a suitable proxy for "reaching sustainability" for the Gratipay team as concretized in my own personal ability to be sustainably funded on Gratipay.

To hit the "first $100,000 week" goal we will need to see 25% average monthly growth (AMG) for the next nine months. Our current best nine-month run was March, 2013 through November, 2013, and we only achieved 19% AMG during that run, so we would need to beat our current best run by six percentage points AMG in order to hit the "first $100,000 week" target. Which is to say, it looks like we're not going to achieve this goal.

I'm going to use 17% as our best-case scenario because that was our longest sustained run. I'm using 12% for our likely scenario since that encompasses the most history. I'm using 2% for our worst case since that is one standard deviation below our long-term average.

Scenario AMG Months to Sustainability
Worst-Case 2% 106
Likely 12% 18
Best-Case 17% 13

Assumptions:

chrisdev commented 10 years ago

No your original calculations are about correct. Your total growth from month 10 to month 29 is (55311/6498 -1)* 100 = 751.2% So the average monthly growth rate is 39.537 %

chadwhitacre commented 10 years ago

The purpose of a loan would be to artificially sustain the Gratipay team until we reach sustainability. I would start taking $600/wk immediately. @Changaco @seanlinsley Are you able to estimate what you'd expect to up your take to if we receive a loan? Really it'd be good to hear from all team members on this point.

chadwhitacre commented 10 years ago

No your original calculations are about correct.

Then why am I only seeing from -9 to 25 percent growth month to month in the new "Growth" column I've added? What am I doing wrong?

seanlinsley commented 10 years ago

Normally my limitation isn't money, but time. It'd be easier to estimate how much I'd up my take (and my time dedication) if I was a contractor instead of working as a full time employee elsewhere.

If I were to work for Gratipay more-or-less full time, I'd imagine I'd base my take on some fraction of a normal contracting rate, multiplied by the hours I actually put in.

chrisdev commented 10 years ago

Ok the average growth rate is misleading. Based on the Compound Monthly Growth Rate we have this formula ((End Value/Start Value)^(1/(Periods - 1)) -1 ((55311/5439)^(1/(24 - 1)) -1 = 10.15%

chadwhitacre commented 10 years ago

@chrisdev Hah! That's a third number again! :D

chrisdev commented 10 years ago

Well it only takes the the start and end into account but it's close to your average of monthly growth rates so we go with your method :)

chrisdev commented 10 years ago

What is your break even for the loan repayment?

chadwhitacre commented 10 years ago

Our sustainability target is $13,000 per month in income for Gratipay. We spend ~18% on operations (includes savings), leaving $10,660 for payroll at our sustainability threshold. Our goal with this loan is to reach that threshold.

Let's determine our payroll cap by looking at the effect on our balance relative to our income. Let's say that we never want to let our balance dip below 3x our month income—really we want more like a year's savings, but reaching sustainability comes first, so let's allow ourselves to get as low as 3x income.

On our bootstrapping projection, payroll starts next month at $1,185. I'm saying that with a loan I would plan to immediately bump up my take so that payroll totals $3,525 per month ($1,185 - ($59.99 * 4.33333) + ($600 * 4.33333)). That establishes a lower bound for a payroll cap. The question for each scenario is, can we reach our 10,660 sustainability threshold for payroll without dipping below our 3x floor on our balance?

The worst-case scenario is pretty dire. Let's say that we cap payroll at $3,525 per month for that scenario. My take is $2,600/mo so there's not much room for anyone else. It looks like we hit 3x at 27 months in that scenario, which is 79 months before we would've reached sustainability on that path.

In the likely scenario, we reach sustainability after 18 months, which means we can cap payroll at about $6,100. This puts me at about 43% of payroll for the next year and a half.

In the best-case scenario, we reach sustainability after 13 months, which means we could cap payroll at about $7,600, and I'd be at 34% for the next year-ish.

If the above analysis is sound, I believe it indicates that a 72-month loan of $90,000 at 8% interest could in fact help Gratipay reach sustainability by enabling me personally to continue working full-time on it, in addition to providing another $3,500 per month in money that we can safely make available for payroll.

chadwhitacre commented 10 years ago

but it's close to your average of monthly growth rates so we go with your method

You just want to go with the higher number. :-P

What is your break even for the loan repayment?

Does my last big comment answer your question? If not, can you help me understand the question better?

chadwhitacre commented 10 years ago

I guess I've focused so much on reaching sustainability that I haven't actually thought about repaying the loan. The projections go out 72 months which is the likely term of the loan. It would take more work to solidify our projections beyond the point where we reach sustainability. I guess I've assumed that if we reach sustainability then we'll be able to continue growing and finish repaying the loan.

colindean commented 10 years ago

Also, what happens if you take out the loan, but income doesn't increase enough to cover the loan payments?

chadwhitacre commented 10 years ago

@colindean Does that differ materially from the worst-case scenario outlined above? I suppose the answer is that we default on the loan and Gittip, LLC goes bankrupt. Eh?

chrisdev commented 10 years ago

Does my last big comment answer your question? If not, can you help me understand the question better?

Yes +1

You just want to go with the higher number. :-P

I like your approach since you have also provided a measure of the riskiness i.e the stddiv

Sent from my iPad

On Aug 10, 2014, at 10:32 PM, Chad Whitacre notifications@github.com wrote:

but it's close to your average of monthly growth rates so we go with your method

You just want to go with the higher number. :-P

What is your break even for the loan repayment?

Does my last big comment answer your question? If not, can you help me understand the question better?

— Reply to this email directly or view it on GitHub https://github.com/gittip/building.gittip.com/issues/78#issuecomment-51736896 .

chadwhitacre commented 10 years ago

I like your approach since you have also provided a measure of the riskiness i.e the stddiv

Okay, I buy it. :-) I appreciated seeing the standard deviation as well.

dsernst commented 10 years ago

What happens to Gittip the platform, & everyone else's payments if the loan defaults?

On Sunday, August 10, 2014, Chad Whitacre notifications@github.com wrote:

@colindean https://github.com/colindean Does that differ materially from the worst-case scenario outlined above? I suppose the answer is that we default on the loan and Gittip, LLC goes bankrupt. Eh?

— Reply to this email directly or view it on GitHub https://github.com/gittip/building.gittip.com/issues/78#issuecomment-51737323 .

chadwhitacre commented 10 years ago

@dsernst Dunno. Good question. Anyone here have experience with bankruptcy?

Also, @dsernst, did urBin take funding? Do you have direct experience with all this?

colindean commented 10 years ago

What happens to Gittip the platform, & everyone else's payments if the loan defaults?

I was more concerned about the wellbeing of Gittip as an organization, but I think this is the better thing to consider.

I would imagine that come kind of bankruptcy plan would enable Gittip to continue operating and reorganize its debt. It has no material assets, and its intellectual property is open source (not without value, but has there been a case wherein a company with an open source product, or really, a platform, gone into bankruptcy?), so there’s nothing that debtors could go after, other than Gittip’s leadership…

dsernst commented 10 years ago

Apologies if there was just a double post, I have weak service. (I'm @ a campground in the middle of South Dakota).

Yes I've been through funding for 3 different businesses, each pretty different though:

On Sunday, August 10, 2014, David Ernst david@dsernst.com wrote:

What happens to Gittip the platform, & everyone else's payments if the loan defaults?

On Sunday, August 10, 2014, Chad Whitacre <notifications@github.com javascript:_e(%7B%7D,'cvml','notifications@github.com');> wrote:

@colindean https://github.com/colindean Does that differ materially from the worst-case scenario outlined above? I suppose the answer is that we default on the loan and Gittip, LLC goes bankrupt. Eh?

— Reply to this email directly or view it on GitHub https://github.com/gittip/building.gittip.com/issues/78#issuecomment-51737323 .

chadwhitacre commented 10 years ago

Of course, the worst-case is actually worse than described above, because we're not going to start by capping payroll at $3,525, we're going to cap it at $6,200, so we'll burn down to 3x in 10 months, not 27. If our AMG for the next 10 months is 2%, I will be dismayed and probably ready to throw in the towel. I've reworked the worst-case scenario to reflect a drop in the payroll cap from $6,200 to $3,525 in month 11, which buys us an extra five months to hopefully bring Gratipay in for a landing (we run out of money in month 19 instead of month 14).

chadwhitacre commented 10 years ago

What happens to Gittip the platform, & everyone else's payments if the loan defaults?

In our worst-case scenario, we sell and/or turn off the service by the end of 2015.

dsernst commented 10 years ago

I agree the source code would be fine, but I'm concerned about the db: users, their tips, & credit card info. As far as I can tell, that's an asset of Gittip LLC's. So the bankruptcy courts would try to liquidate it if it came to that, in order to pay back the debt.

On Sunday, August 10, 2014, Colin Dean notifications@github.com wrote:

What happens to Gittip the platform, & everyone else's payments if the loan defaults?

I was more concerned about the wellbeing of Gittip as an organization, but I think this is the better thing to consider.

I would imagine that come kind of bankruptcy plan would enable Gittip to continue operating and reorganize its debt. It has no material assets, and its intellectual property is open source (not without value, but has there been a case wherein a company with an open source product, or really, a platform, gone into bankruptcy?), so there’s nothing that debtors could go after, other than Gittip’s leadership…

— Reply to this email directly or view it on GitHub https://github.com/gittip/building.gittip.com/issues/78#issuecomment-51738077 .

chadwhitacre commented 10 years ago

@dsernst You could be right. Credit card info is vaulted with Balanced, but of course we have the access tokens to make charges etc.

dsernst commented 10 years ago

At risk of changing the subject prematurely...

In what manner would the loan be repaid? Will you personally handle it from your take, Chad? Will it be automated as a percentage of the Gittip Team?

On Sunday, August 10, 2014, Chad Whitacre notifications@github.com wrote:

@dsernst https://github.com/dsernst You could be right. Credit card info is vaulted with Balanced, but of course we have the access tokens to make charges etc.

— Reply to this email directly or view it on GitHub https://github.com/gittip/building.gittip.com/issues/78#issuecomment-51738394 .

chadwhitacre commented 10 years ago

I think the real point here is that we just simply can't continue bootstrapping. Here's months to various payroll thresholds for our scenarios without outside capital:

2,600 3,525 10,660
Worst Case 48 63 119
Likely 9 11 21
Best Case 6 8 15
chadwhitacre commented 10 years ago

In what manner would the loan be repaid?

I would expect it to be an additional expense out of our operations bank account. I.e., paid in the same way we currently pay Heroku, GitHub, etc.

chadwhitacre commented 10 years ago

What the comment before last shows is that, even in our best-case scenario with no-one else taking anything, we are still six months away from having enough payroll to cover my minimum requirements. I can't afford another six months. What this shows is that if we don't get a loan, it's ... time for me to go back to part-time on Gratipay.