Here it is! Our submission for the DonorsChoose Hacking Education contest:
Hacking Education: A Contest for Developers and Data Crunchers
Ten years ago, a teacher in the Bronx launched DonorsChoose.org. Since then, more than 165,000 teachers at 43,000 public schools have posted over 300,000 classroom project requests, inspiring $80,000,000 in giving from 400,000 donors who performed over a million search queries and made more than a million donations.
We’ve opened up that data, and invite you to make discoveries and build apps that improve education in America.
Exciting stuff, right? (I’m a nerd, just go with it)
One of the issues DonorsChoose identified is the fact that only about 2% of site visitors actually donate to a project. The challenge? How could DonorsChoose convert more visitors (potential donors) to donors? We’ve submitted an admittedly simple (but potentially powerful) idea:
Increase conversion rate by showing potential donors only the projects they have the strongest connection to on DonorsChoose.
There are thousands of great projects on DonorsChoose– that’s clearly not why people aren’t converting from visitor to donor. So why do only 2% of site visitors make a donation?
The plethora of projects to choose from might actually be a deterrent to giving- out of thousands of projects, how can you narrow them down to the perfect project for you? We brainstormed about the motivations behind why people give- and we think a strong motivation is relevancy; if you feel connected to a project, you’re more likely to give to the project (and the DonorsChoose data we analyzed told us it wasn’t a crazy idea).
The challenge to ourselves: How could we introduce relevancy to DonorsChoose projects?
Simple enough, right? Advertisers have been targeting for years– geographic targeting, behavioral targeting, demographic targeting… all in an effort to reach the right type of consumer. Why not apply this same concept to the non-profit world?
All we needed was data. The DonorsChoose donation data we looked through suggested that people were likely to give money to a project in their state. People also searched for projects based on subject matter and by keywords.
[Searching for and by relevancy to pick the perfect project]
But how to get more information on donors? We chose Facebook as our data source. Why? For one, ~135.6M people in the US have a Facebook account. And two, people list a ton of information on their profiles. Large sample + information that could be used to make connections with DonorsChoose projects= A+! From Facebook profile information we could get info like home states, current states, college majors, interests, likes, activities. And all the information would be available with a single login.
Our suggestion? For DonorsChoose to add a Facebook Connect option on their homepage. DonorsChoose actually already had a Facebook Connect option, but it was buried behind the homepage and only gathered basic information like name and email from visitors. Featuring the Facebook Connect option prominently on the homepage would make it easier for visitors to immediately sign in with Facebook– and instantly see projects only related to their interests and their geographic location. Visitors would no longer have to go through a multi-step, manual search process to find ‘their‘ perfect project- projects would be automatically tailored for them.
People who connected to DonorsChoose through Facebook would be matched to projects based on their current and/or home state, their college major, and their interests. The state information would be used to match the visitor to DonorsChoose projects by their relevant states. Then the visitor’s college major and interests would be matched (weighted by strength of connection) to DonorsChoose projects with similar subject matter.
For example, if you live in New York, were a computer science major, and list technology as an interest, you’ll get a project whose classroom is in New York and whose subject matter is related to math/science and technology. Pretty cool, right? We think so!
Here’s our submission for the contest, we’d love to hear your thoughts.
And just for kicks, here’s a funny story about why I’m really attached to this idea. I’ve been ruminating on it every since I participated in The Great Twitter Secret Santa. What is The Great Twitter Secret Santa? It’s pretty self-explanatory- a traditional Secret Santa game, only via twitter. Mutual friends/followers were matched up, and I got matched with @jderoner (who coincidentally also made the app). We were 100% strangers at the time, and his request for a gift was that a donation be made to DonorsChoose. I had no idea what project to select for him, and I spent quite a bit of time trying to figure out a project he would like– only knowing that he lived in NY and was a web designer/liked technology. As I searched and searched for projects, I kept thinking that it would be so much easier to select the ‘right‘ one (for him- all of the projects are ‘right’ for someone) if I knew a bit more about him. Fast-forward to spring (and the fact that we’re no longer strangers) and the announcement of the Hacking Education project- what a perfect opportunity to share the idea!
Our original idea was to turn the analysis into a working app, (since he’s a computer whiz and all, it was the perfect combination) but we ran into a snag when we couldn’t get access to the Facebook Ads API. Fingers crossed we’ll get access soon since our ultimate goal is to one day build the app… so stay tuned!