I've been reading Malcolm Gladwell's The Tipping Point, and came across this rather interesting interpretation of Pareto's Principle;

Twenty percent of beer drinkers drink 80 percent of all beer.

Can this be correct? I've searched around online but can't find any study that backs this up, despite the fact that I found roughly the same thing quoted elsewhere, for similarly difficult to believe "facts", but I never see references to any studies.


2 Answers 2


No it isn't, at least in UK in 2010.

It's the best data I could find and it will require some calculations. I will try to err in favor of the claim in order to disprove it convincingly.


According to the General Lifestyle Survey Overview and in particular the General Lifestyle Survey - Drinking tables 2010 (Excel sheet 609Kb) found here, in particular Table 2.16.1

The Units and unit% columns are mine and are so calculated:

  • using the Rural units
  • assuming 0 units for nothing, 2.5 units for up to 4/3, 6 units for up to 8/6, 10 units for up to 12/9, and 15 units for more than 12/9. I believe I've exceeded in favor of the claim.

As you can see, the top 20% (more or less) consumes about 50% of the alchool (more or less). It's very far away from 80%.

Furthermore, there is no guarantee that the top 20% is the same people every week: if enough people have a sporadic wild night out once a year, they would appear as heavy drinkers in this table. This makes it either realistic or an over estimation in favor of the claim.

Based on this I can conclude that the evidence points against the claim.

1The double numbers "4/3", "8/6", etc. refer to male vs. female numbers of units.

  • 2
    Worth noting that this is alcohol consumption, not beer consumption. (Ah, I see this is because the title doesn't match the cited claim.)
    – Oddthinking
    Dec 2, 2013 at 4:09
  • 1
    True. Beer counts for a large part of it, I did check, but the question title speaks abot alcohol
    – Sklivvz
    Dec 2, 2013 at 10:04
  • @Sklivvz But we analyze the notable claim, not the title. I don't doubt that this answer applies to the notable claim as well.
    – user5582
    Dec 2, 2013 at 19:43
  • @Articuno I couldn't find beer-specific sources, maybe you can do better and disprove me, though. I do not think that the numbers for beer will be significantly different. We are not talking about a situation where minor differences here would change the answer qualitatively: the actual measured units for the top 20% drinkers are about half of what they should be, and they are not even the regular top drinkers.
    – Sklivvz
    Dec 2, 2013 at 19:56
  • 1
    @Sklivvz looking at the question we are only looking at people consuming alcohol. Many people may drink either excessively, or 0 depending on what they did that weekend. In this survey, someone who drinks every weekend comes up the same as someone who drinks 2 times a year, once you only count drinkers. As alcohol consumption is a social event for many, it feels like a whole year should be examined to allow counting for different social events
    – Andrey
    Dec 5, 2013 at 15:55

A long time has passed since this question popped up, but I have found evidence that yes, the Pareto Principle is valid for alcohol/ beer consumption.

Quoting this paper (BMJ 2016;353:i1860):

Harmful and extreme drinkers comprise a tiny minority, 4.4% of the population, but consume one third of all alcohol sold; the combination of hazardous, harmful, and extreme drinkers provides almost 70% of drinks industry sales by volume

I have also provided the table 1 from this paper. If you add up the percentages (in parenthesis) of the Hazardous and up drinkers for "strong cider/lager/beer) it is equal to 82%. Other categories and total alcohol comsumption are a little below 80%, but close enough I guess.

enter image description here

  • I don't understand your math. 38.8+14.8+13.2 = 66.8 Jun 11, 2018 at 16:07
  • @Accumulation you appear to be reading from the "Normal cider/lager/beer" line.
    – CactusCake
    Jun 11, 2018 at 19:31

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .