There are ten of you, and you want to go to dinner. The other nine people include two vegetarians, a curry-addict and someone who likes anything ethnic. Some of the rest have very mild preferences; others aren’t really bothered where they eat. You have boiled down the choice to six restaurants.

How on earth do you decide where to go?

Currently, the way in which we try to resolve such problems can be highly unsatisfactory.  People may game the system by exaggerating the intensity of their feelings – noisily campaigning for their favourite option to the exclusion of anyone else. After all, you have nothing to lose by pretending to care a huge amount about going to a restaurant you only slightly prefer to all the others.

Or perhaps the opposite happens. In a kind of triumph of lazy compromise, you end up going to the restaurant the fewest people object to. This generally means that everyone ends up in a pizza restaurant, pizza being a kind of default food which offends no one. This is okay, so far as it goes, but it always seems a little sad to me to pay £35 or so for a meal which is not much better when prepared in a restaurant than when bought for £5.95 and cooked at home.

If only there were an app that could sort out this problem. Something which could help a group chart a safe course between the Scylla of minority tyranny and the Charybdis of lazy consensus.

When such an app appears, I suspect it will make use of the idea of Quadratic Voting, which was recently proposed by Glen Weyl and Stephen Lalley at the University of Chicago.

Their initial proposal was that people in, say, referenda would receive one vote each, but with the freedom to buy additional votes, where the cost of additional votes for any individual would be the square of the number of votes purchased. So that one vote might cost $1, 2 votes $4, three votes $9 and so on. (The money raised would be paid back equally to all voters). The system would thus be sensitive to intensity of feeling in a way that conventional elections aren’t, but not to the extent where a few people could overrule a large majority, at least not without paying them a fortune.

Like me, you probably instinctively recoil at this idea when applied to democratic politics. But the central principle behind this system has a widespread application in market research – and without real money being involved at all.

Let’s take that restaurant problem again. Suppose your app gave everybody 500 ‘voting dollars’ every year. When presented with a choice between restaurants, people who cared a great deal could spend 64 dollars and put it towards a vote for just one restaurant, which would count as eight votes. Others might put 16 dollars towards four restaurants, which would each count as four votes. The vegetarian might spend 9 dollars voting against the steakhouse. The indifferent could save up their dollars for another time, when they actually cared about the outcome.

This works a little like a truth serum. Because there is a trade-off between being totally inflexible and easy going, and between caring and not caring, this mechanism encourages people honestly to reveal the breadth and shape of their zone of tolerance. It hence uncovers, both individually and in aggregate, a level of nuance which no other system quite matches.

I would love to know what people think of this idea. Let me know via @rorysutherland.

Rory Sutherland is vice-chairman, Ogilvy & Mather

This article orginally appeared in Impact magazine Issue 9 April 2015.