Quakers have been taking decisions by consensus, from the bottom-up, for 360 years or so. They in turn had a great effect upon all protest/social movements since.
Today "democracy" is really 'mob-rule by ballot'. Those who shout loudest, or scare more people with negative campaigns, win. More importantly, if you want to rest control from people you have to keep the agenda running quickly before people wise-up to what's really going on (err, Peak Oil??) and change their mind -- think of all those political/economic U-turns over the years after the true facts of the case came to light (e.g. Iraq). Once you stoop to representation you must instantly simplify or "frame" the issue to make it easy to communicate, and that inevitably involves "spin" in order to try and secure a mandate from the largest possible number of people.
Consensus decision making takes longer for the simple reason that it has to consult everyone involved. You also, to keep the scale of the decision-making process manageable, have to take it at the lowest level/amongst the smallest "critical mass" possible. That is of course the antithesis of the type of decision-making that favours minority control, and the forms of hierarchy that this perpetuates.
And what about those who "don't" vote? E.g., I vote at elections (it's too hard-fought a right to ignore), but as no party (not even the Greens) represents my views I usually write something erudite on the paper instead. As far as I am aware, the counting of spoilt ballots is up to the decision of the returning officer in the UK, and the centrally collected figures for turnout do not include the figures for spoilt ballots, only valid votes --
i.e., a large proportion of the "no show" in the turnout figure could actually be people soiling their ballots, but the system doesn't record this. If you want to demonstrate that democracy works then you have to include the proportion of those not voting -- and has turnout has gone down for the last 40 years it means that parties today are only elected on a fraction of popular consent.
The ONLY time that a government in Britain was elected on a full (50%+) majority of the population was in 1931. If you calculate the result of the winning party on the total number eligible to vote, rather than those voting, the situation looks pretty dire. Dumping the data from some figures I keep up to date for situations such as this, as a proportion of the whole electorate the winning party's proportion of those eligible to vote was: 2010 Con-Dem (38.5%): 2005 Lab (21.7%): 2001 Lab (24.2%): 1997 Lab (30.8%): 1992 Con (32.3%): 1987 Con (31.8%): 1983 Con (30.8%): 1979 Con (33.4%): 1974 Oct Lab (28.6%): 1974 Feb Lab (29.3%): 1970 Con (33.4%): 1966 Lab (36.3%): 1964 Lab (34.0%): 1959 Con (38.9%): 1955 Con (38.1%): 1951 Con (39.6%): 1950 Lab (38.7%): 1945 Lab (34.7%).
So, Bliar never even had a third of the the UK electorate behind him, and on his last victory he scraped home with just over a fifth! Even the present Con-Dem government has just less than two-fifths support (although "support" is an interesting choice of words since a lot of Lib-Dems I know "didn't vote for that").
Let's face it, in Britain only big parties can win elections, and that's the root of the problem of representation of public will --
it's the parties, not simply the voting system! I think, secretly, that's why both Con and Lab oppose proportional representation -- under a proportional system both their parties would fracture down their traditional internal fault lines; Labour into at least two parts, and the Cons into two or possibly three.
In short "anarchy", where there is no "central order" but everyone accepts the discipline of management and contributes to the evaluation of strategy and decisions, is far preferable to "democracy" where the minority always rule.