LiquidFeedback

The bet

In 2009 a group of Berlin developers, prompted by frustrated Pirate Party members, published LiquidFeedback. It let a party decide its positions through delegative voting. On any given question you could vote yourself or delegate your vote to someone you trusted, who could in turn delegate it onward.

The promise was a fluid democracy without hierarchy, where expertise and trust routed power as needed instead of freezing it into a party establishment. Germany's Pirate Party adopted it, and for a moment it looked like the future of how parties, and maybe governments, might decide. It remains the most serious attempt anyone has made to run liquid democracy at scale.

Peak

It became the largest experiment of its kind in the world. By January 2015 the Pirate Party's installation had 13,836 registered users, the largest platform of its kind in the study. Active use surged with the party's electoral breakthroughs: first when the Berlin Pirates won seats in 2011, and again in the run-up to the 2013 federal election.

Internationally, LiquidFeedback became the reference implementation, the thing every discussion of liquid democracy pointed to. And it ran actual internal decisions. It never stayed a thought experiment.

What worked

LiquidFeedback was well engineered and open-source, and the party used it for consequential decisions. It was no demo. It carried liquid democracy further than anyone had.

Even its failure was a contribution. Because the experiment ran in the open, it produced hard evidence about how delegative democracy behaves in practice, including rigorous academic study. The field learned more from LiquidFeedback than from a dozen tools that merely proposed the idea.

What failed — the mechanism

Every member was handed power, and the power flowed uphill. Delegations stacked transitively, so a small number of super-voters built up large blocs of votes. A system designed to dissolve hierarchy ended up with super-voters: the study counts 38 people holding more than 100 delegations. The same study found they often voted with the majority and did not fully use their power to change outcomes. After a debate over super-voters, it reports that members questioned the system's democratic nature and many became inactive.

Meanwhile the broad participation the model assumed never held. Active use spiked around the elections and then fell off sharply, leaving a small, unrepresentative core to make the decisions. Votes were public, and an article in the Liquid Democracy Journal reports that strong concerns about voting by roll call were prevalent among the Berlin Pirates.

When the Pirate Party itself splintered and lost its electoral relevance, the platform that had been its proudest innovation fell silent with it.

What Agora takes from it

Who ends up holding power in an open participation system? Left to itself, a hyperactive few. LiquidFeedback showed it inside the biggest experiment of its kind.

Agora doesn't let that minority stand in for the constituency. It weights signal by constituency, in-district against out-of-district, and not by raw activity, and it refuses gamified mechanics that reward the most active users. The same lesson is part of why Agora anchors participation to the substance of officials' work instead of open member-to-member delegation. Agora doesn't assume broad participation will sustain itself without an anchor like that.