PIGEON: An Interest-Based, Targeted Email Service for Large Organizations

Large organizations, like colleges and companies, send a lot of internal advertising emails. These mass advertising emails are necessary, but they cause two pain points for recipients and senders. Recipients end up with cluttered inboxes with irrelevant emails they need to sift through, decreasing productivity; senders suffer from ineffective emails, since mass emailing reduces the chance of any one person reading the email. Pigeon is a two-pronged solution tackling both problems by allowing recipients to subscribe to tags they’re interested in receiving emails about and senders to send emails with lists of tags that will only end up in the inboxes of the recipients who are likely to be interested in the email. Experiments suggest that Pigeon increases the relevance and effectiveness of emails that users send/receive, since emails sent through Pigeon yielded a 70% open rate, compared to a 20% open rate when sent through a traditional email list. Pigeon can also suggest tags to senders through applying natural language processing techniques to the email content. Preliminary results show that 32% of tags suggested by Pigeon are highly relevant but Pigeon often misses relevant tags altogether. While I’ll describe the product as a whole, I will focus my implementation details and results on my half of the product, the Chrome Extension that solves the sender-side problem.