I just backed this project. I love the way they packaged the pledge levels.
Cutting to the chase
My final analysis on Diaspora’s hyper-successful crowdfunding strategy demonstrates that they sold 1,942 more $25 pledges than they “should have” because of their t-shirt strategy, resulting in additional $48,550 just for the $25 pledge category!
Take a look:
This was because to get a Diaspora t-shirt, a person had to pledge at least $25. Diaspora succeeded in nudging people who had a willingness to pay something, say $5 or $10, up to a $25 price point. In one word: Clever.
But why a t-shirt? Because a Diaspora t-shirt provides significant, visible and hard-to-fake social proof of geekness (being “cool” amongst geeks), which translates into social capital. (Clearly, this is nothing new. Rock bands have been doing exactly this for who-knows-how-long with concert tour t-shirts.) If you plan on crowdfunding a project, it would behoove you to draw the right lessons from Diaspora, although I am not sanguine about the possibility of replicating what is essentially a Black Swan.
First off, I don’t have access to all of the real data points. My data are gleaned directly from the fund-raising page itself. This means I don’t “see” the exact amount pledged in many cases, if someone pledged $645.21, for example.
Second, my major assumption is that each pledge bucket is not a different product, ergo each pledge does not have its own unique demand curve. What I think we are seeing is that Diaspora is selling one product at various prices (hence, one demand curve).
A pledge is a product that signals you are:
a) a geek
c) a trend-setter
while also providing:
e) the satisfaction of sticking it to The Man (Facebook)
f) the good feeling you get from being philanthropic
Nuts and bolts of it
I threw the publicly available data in an Excel file and given the non-linear nature of the data, I applied a natural log transformation to both price and quantity before running an OLS regression. The regression fits well and the log transformation preserves the $25 spike.
The regression predicts a quantity of 6.5 for a price of 3.22 — to make sense of that, we simply take the antilog of 6.5 to get 664 t-shirts. Plugging 664 into the original data set for the expected number of t-shirts to be sold at $25 pledge level generates the second image from the top above. The one with the smooth curve. Here, I’ll show it to to you again.
And of course, my new regression is even a better fit (no surprise given that I have plugged in one of its own outputs). :)
If you want to see the actual regression outputs, download this Excel file: Diaspora Regressions for Blog Post