Anatomy of a crowdfund, week 1
It was a whirlwind first week of campaigning for the Crowdsourcing Discovery (@Crowd4Discovery) project!
Out of the starting gate we benefited from tailwind generated by a post on Sam Arbesman’s wired.com blog, Social Dimension. That amplification, plus my own volley of tweets and Facebook sharing, netted us in 24 hours 35% of the total contributions donated in the first week of the campaign. Then in the 24-48 hour interval, more coincident blogging by the science blogosphere, a post on Derek Lowe’s In the Pipeline, followed by a post on Ash Jogalekar’s Scientific American blog Curious Wavefunction, sustained our momentum going into the first weekend.
If you’re never heard of or seen the “Lowe Bump,” behold it in full bloom (orange arrow) in this plot of site traffic to perlsteinlab.com (courtesy of Google Analytics) from October 1 to October 11. (Chemists, please note that Ash Catalyst was added to the reaction):
Here’s a closer look at the first 72 hours post-launch, spanning October 4-6:
As I blogged two months ago in a post called Quantified Self Publishing, those blips that peak at 30 visits correspond to retweet (RT) events. Several RT blips occurred around Noon EDT, which is when I usually send out my first substantive tweet of the day. Each RT blip is short-lived, lasting approximately one hour. The Ash-catalyzed Lowe Bump was an entirely different beast. First, it was an order of magnitude above baseline traffic levels. Second, it faded gradually over the course of 12 hours.
Okay, traffic data to my lab website is nice, but what about traffic to the Crowdsourcing Discovery project page on RocketHub? As a proxy for pageviews, which are not directly available from the RocketHub website, I’ll use data from the Crowdsourcing Discovery project video page, which I scraped from Vimeo. A similar picture emerges:
So how much money did we raise? How many people contributed? Who’s contributing?
As of one week + one day, we’ve raised $5,990 (24%) from 75 contributors, so an average contribution of around $80. The median contribution = $50. The daily totals are shown here:
It’s pretty clear that getting more than 10 donations/per day requires popular press amplification. Coordinated tweeting by me and a few project supporters, including the setting of small, realistic calls to action, resulted in the 3-fold bump on October 11th. Going forward, the only way we’ll top 25 contributors in one day will be if we’d be fortunate to get a RT or mention from a massively followed tastemaker or fresh press coverage garnering 5,000 – 10,000 pageviews. What happens at the 100,000-pageview level I can only imagine…
Here’s the distribution of donations at the 8-day mark:
Who are these brave souls? 60% are friends and family. Of the 30 people not directly connected to me, at least 5 are directly connected to one of the other team members. Less than 25% are scientists, defined as a grad student, postdoc, tenure-track professor or private-sector researcher.
For the next crowdfund drive update in a week, I’ll start to visualize the data as actual networks in order to see if any interesting connectivity patterns emerge. And in case you missed it, I blogged about the first 96 hours after launch here.
In the meantime, your feedback is welcome in the comment thread below. Thanks!