Jumo

The bet

Jumo bet that the friend-following mechanic that had just made Facebook a phenomenon, its own co-founder Chris Hughes had run Barack Obama's 2008 digital campaign, could do the same for activism: follow the causes and organizations you care about the way you follow people, and let your feed become a discovery engine for where to give your time or money. Hughes launched it in public beta on 30 November 2010 after a year spent traveling and thinking about how people find causes to support.

Peak

The launch was a genuine sensation by civic-tech standards. Within two days the site was overwhelmed with traffic and landed on the New York Times's most-emailed list; within weeks, roughly 15,000 nonprofits and NGOs had created a presence on the platform and more than a million people were following issues through it, at a peak of about 114,000 monthly unique US visitors in December 2010.

What worked

Jumo proved the discovery problem was real and that a well-known founder with a media moment behind him could solve the cold-start problem almost instantly, both nonprofits and individual users showed up in large numbers with essentially no independent marketing effort. The org-discovery layer itself, a clean, browsable index of thousands of causes and the ability to follow them, was a genuine, useful piece of infrastructure that outlived the company.

What failed — the mechanism

The mechanic that produced the launch spike was not built to produce return visits. Following an issue is a single click with no ongoing obligation attached to it, so once the initial novelty and press cycle passed there was little pulling people back, and traffic fell from that December peak of 114,000 to roughly 8,500 by the following June, a drop of well over 90% in six months. A feed of causes to follow, without a structured reason to come back, behaves like most engagement-optimized social products once the discovery phase ends: it either has to manufacture new stimulation to hold attention, or it empties out. Jumo did neither at scale, and in August 2011, less than a year after launch, it merged into GOOD, a media and magazine company, with Hughes staying on only as a senior advisor; the code was open-sourced two months later.

What Agora takes from it

Jumo shows how fast a following-based mechanic can manufacture an audience, and how little that audience is worth once the follow itself is the entire interaction. Agora deliberately does not build a feed you follow and forget: participation is structured around specific, acted-upon civic moments, a bill, a budget line, a response from an office, not an open-ended stream of causes competing for a single click. The lesson is not "add more engagement." It is that a following mechanic without a real obligation on either side, the office to respond, the citizen to act, decays to the same place Jumo did, however good the first two weeks look.