TRANQUIL and picturesque, Friesland in northern Germany might seem an unlikely spot for an experiment in digital politics. Its population is ageing, its broadband patchy. Yet for five months local people have developed and submitted recommendations to their councillors using a website called LiquidFeedback, best known as the policy-setting platform favoured by the Pirate party, a tech-loving outfit that has won seats in four regional parliaments. Some citizens have voted for the council to stop publishing the location of speed cameras (on the ground that this helps drivers avoid them); others demand more say in how schools spend their money. Sonke Klug of the local council says he gets telephone calls every week, from big cities and small towns, asking for advice.

Digital politics carries high hopes. Many think it will help citizens govern themselves more effectively than via a professional caste of politicians. But just as putting cameras in parliaments did not usher in an era of teledemocracy, so digital politics has failed so far to displace the baby-kissers.

Pressure to reform once came from below but innovators now have powerful patrons. Some governments, impressed by the success of online campaigning sites such as Avaaz and wired parties such as Italy’s Five Star movement, are embarking on fresh experiments.

Sites which allow citizens to draft e-petitions for their pet causes are mushrooming. Politicians in Germany created one of the earliest, in 2005, and dozens of countries have followed, most of them within the past three years. America’s “We the People” is perhaps the most successful. Since its launch in 2011 citizens have created nearly 180,000 petitions; since November the total number of signatures has tripled to almost 12m.

This is not just for show. In Finland successful e-petitions bring about a parliamentary vote. American petitions that pass a threshold—raised in January from 25,000 to 100,000 signatures—are guaranteed only a response from the White House. Between November and January 34,000 people implored the government to start building a Death Star, the planet-sized spaceship in “Star Wars”. Paul Shawcross, Barack Obama’s space wonk, dryly responded that it would cost a deficit-swelling $850 quadrillion and pointed out that, in any case, it had a “fundamental flaw” (the cinematic version is blown up by a one-man starship).

More serious proposals have flourished, too, especially those favoured by the digerati. On March 4th the White House endorsed a petition to abolish rules that prevent mobile-phone users from moving their handset between operators, even when their contracts have expired.

Macon Phillips, the White House digital director, says citizens appreciate conversation with their governments, even when their petitions do not change policy. His team surveyed some of the 2m or so people who had received responses through “We the People”: two-thirds found the dialogue “helpful”, and 86% said they would use it again.

But critics find e-petitions too unambitious. Platforms rarely enable users to discuss issues or fine-tune their demands. “Successful politics is not about finding people who agree with you,” says Anthony Zacharzewski of Demsoc, a British think-tank. “It is about making difficult decisions without killing each other.”

Crowds of citizens are good at tabling proposals and then voting on them; less successful at the horse-trading required to produce consensus—let alone a law that actually works. The idea of writing legislation collaboratively using “wiki” software (which anyone can edit) has generated headlines but few results. Madison, one such platform released in 2012 by campaigners battling anti-piracy laws, helped American politicians weed out errors in a draft bill but has not been much used since.

Focus on the small stuff

In one year Congress passes just a few hundred laws (when not hobbled by partisan mudslinging), whereas American government agencies pump out up to 8,000 new regulations, says Beth Noveck from New York University’s governance lab. In 2007 she founded Peer-to-Patent, an attempt to enroll experts in regulatory decisions—in this case, to speed up America’s tedious patent process. Pilot studies followed in Britain, Australia and Japan.

About 1,500 cities, including Chicago and, last year, New York, have also enlisted the public in setting budgeting priorities. In 2012 around a million citizens took part in the annual budgeting process in Rio Grande do Sul, the Brazilian state which also hosted the first such event, in the town of Porto Alegre, in 1989.

In 2011 the state governor collected 1,300 ideas for improving local health care, and then let citizens vote for their 50 favourites; 120,000 people took part. The voting software presented ideas in pairs; users could pick the one they preferred.

Consultations are one thing. Acting on the results is another. Tiago Peixoto, an expert on government innovation, worries that poorly planned participatory exercises could risk creating “a new generation of disillusioned citizens”. In Iceland in 2011 the citizen council elected to write the country’s new constitution adopted an open approach that included releasing drafts online and inviting advice and criticism through social media. Voters approved its work in a referendum in October, but lawmakers have still to ratify the document. An election on April 27th will prolong the wrangling.

Enthusiasts must also compete with other democratic innovators. Through the Open Government Partnership, an informal international talking-shop founded in 2011, 48 countries have promised to make governance more transparent and participatory. So far they have put openness first—unsurprisingly, as projects involving public data are easier to conceive, to measure and to crow about. But handing citizens oodles of data without providing better means to respond is like offering “a dashboard without a steering wheel”, says Clay Shirky, a technology writer.

The data deluge should help political debate though, by educating citizens and giving experts more numbers to crunch. And governments may detect and solve problems that electors experience but cannot articulate, says David Eaves, a consultant. Websites and apps that make it easy for citizens to report potholes, faulty traffic lights and other inconveniences already make local councils more responsive. So inviting citizens to share more data about their daily lives—including data produced by a proliferation of smart devices—could help solve broader policy questions, too.

That may not be the direct democracy that Utopians covet, but more modest ambitions are in order. Cool-seeming digital tools can narrow participation by excluding poor, old or disabled people. Researchers in Germany report that e-petitions are mostly created by the same well-educated males who create and sign paper ones. “Move fast and break things” may be a good motto in Silicon Valley, but it is a poor prescription for politics.