Tag Archives: expenses

Do 99% of Amway distributors lose money? Part 1

It’s a common claim among critics of multilevel marketing and Amway that “most people lose money”, some even go so far as to give specific figures. Jon Taylor of Consumer Awareness Institute claims loss rates exceeding 99.9%. Robert FitzPatrick of Pyramid Scheme Alert essentially just repeats Taylor’s analysis and claims the “loss rate” exceeds 99%. Former Amway Emerald Eric Scheibeler claimed that a UK court case (BERR vs Amway UK) found a 99.7% “loss rate for investors” and this was reported in a news article and is currently included in the wikipedia article on Multilevel Marketing. The truth is that no such finding was made. Prolific Amway critics like Joecool, Shyam Sundar, and David Brear repeat these myths, and unfortunately so do members of the media.

Is there any truth to these accusations?

An example of a Jon Taylor & Robert FitzPatrick analysis

Robert FitzPatrick and Jon Taylor come up with their figure by analyzing various companies Income Disclosure Statements. Using the averages and number of distributors qualifying at a particular level (a frequency distribution), they work out the total income each level earned and use this to calculate the average income of the “bottom 1%”. Taking their Nu Skin example, using 1998 average income data, they calculated that -

The mean average payment to the bottom 99% of Nuskin distributors was $7.43 per week

and go on to add “before expenses and taxes are deducted – resulting in a significant loss.”

Mathematically their calculations are roughly correct (though averaging from a frequency distribution doesn’t give the exact mean). Statistically however, their analysis is completely bogus. Why?

There’s several flaws. First, when calculating statistics like “mean” or “average”, a measure of central tendency, you need to consider differences between groups included in your sample. For example, when statisticians calculate and present average heights, it’s typically broken down by age and sex. It simply makes no sense to average the heights of, say, 5 year old girls and 30 year old men together. You can do it and get a figure, but what does it tell you? Pretty much the only thing it tells you is you need a better statistician! This however is exactly what Taylor and FitzPatrick do. They pile together people who have been registered for a few months and mix them together with people who have been actively building a business for 30 years and more! They include people working 30 hours a week and people working one hour a week. They include people with a goal to generate a full time income with people whose goal is to buy some products cheaply. It simply makes no sense. The only thing it tells you is need a better statistician!

Still, that’s possibly not the worst thing they do. Jon Taylor claims to have a PhD in Applied Psychology. I too have qualifications in Psychology (and postgrad in Sociology) and I can assure readers that you do not get these qualification without quite extensive training in statistics. Here’s one of the things you’ll typically be taught about statistics like “mean” or “average” -

The important disadvantage of mean is that it is sensitive to extreme values/outliers, especially when the sample size is small. Therefore, it is not an appropriate measure of central tendency for skewed distributions.

What is a skewed distribution? It’s helpful first to look at what’s called a “normal distribution”. Here’s a graph of a sample of men’s heights.

NormalDistribution
source: www.cuclasses.com/stat1001

You’ll note how the graph peaks in the middle and tails off to either side in roughly symmetrical fashion. The more symmetrical it is in this “bell curve”, the more useful a statistic like “mean” is in describing the population or sample you’re interested in.

Now let’s take a look at another distribution.

NuSkin1998 income distribution

You’ll note this distribution is very heavily skewed to the left, then a bump, then a tail to the right.

Remember what I said above about “skewed distributions” and “mean”? It is not an appropriate measure.

But that’s exactly what Taylor and FitzPatrick have done. The second graph above is a graph of the actual data they used to calculate that “The mean average payment to the bottom 99% of Nuskin distributors was $7.43 per week”.

So what is that big group on the left? It’s Nu Skin distributors that earned no bonus at all. According to the 1998 Nu Skin income disclosure statement, fully 86% of distributors earned no bonuses at all. That’s no surprise. This is what it says on Nu Skins’ 2004 income disclosure (I was unable to find a copy of the 1998 one Taylor & FitzPatrick quote) -

As with any other sales opportunity, the compensation earned by distributors varies significantly. The cost to become a distributor is very low. People become distributors for various reasons. Many people become distributors simply to enjoy the Company’s products at wholesale prices. Some join the business to improve their skills or to experience the management of their own business. Others become distributors but for various reasons never purchase products from the Company. Consequently, many distributors never qualify to receive commissions.

The FTC, in their Business Opportunity Rule – Revised Notice For Proposed Rulemaking note comments from, for example Shaklee -

85% of individuals who sign up with Shaklee do so as “wholesale buyers” rather than distributors

Primerica, Quixtar, Melaleuca and others all reported similar statistics to the FTC. Quite simply these people are not operating a business, and their predictable lack of income from not operating a business should obviously not be used in determining whether it’s possible to earn an income through the business. It’s as absurd as judging whether a particular medicine works by including all the people who didn’t take the medicine! It might tell you something, like the pill is too big so people don’t want to take it, but it won’t tell you whether the medicine itself worked or not.

The obsession that anti-mlm zealots have with the low income of people who don’t actually try to make money makes you wonder if their disappointment with MLM comes from the fact it’s not some kind of “get rich quick scheme” and requires work to succeed, just like any other business. It seems they wanted fast riches and were disappointed.

FitzPatrick & Taylor don’t stop with their bogus analyses there though. In part 2 I’ll look at the other side of the profit equation – expenses.