I studied maths until the age of 19. It’s no great academic achievement (my A levels were like a night in a 90s nightclub… lots of Es), but it did mean that I did more than my share of Statistics. I don’t remember much about it, more than anything because in my subsequent academic endeavours I never used anything more complex than means, medians and modes, and in my working life nothing in advance of a standard deviation (and that only to understand the sigma bit of Six Sigma).
The fact seems to be that, with modern life being dominated by numbers, most people, most of the time, don’t use Statistics. A consequence of this is that the confusion between correlation and causality is being fallen into with frightening regularity.
In part, maybe this doesn’t matter… does the A/B testing world that we live in need to understand cause?
In that A/B world, two versions of a way of doing something on a, for example, website are run concurrently. The one that is most successful is implemented, and so functionality evolves in a quite Darwinian way. At no point does anyone need to worry about why one is better than the other… it works, so deploy it.
The challenge comes, it seems to me, if you attempt to make people do things based on assuming causality in a correlation. Take customer satisfaction, for example. There is correlation between recorded customer satisfaction and profitability of a company. There are outliers… many firms have gone bust for providing too expensive a good service, and Ryanair at the other end seen to thrive on upsetting people. (Total aside from a conversation earlier today… if Ryanair made milk, you’d have to milk the cow yourself but the cow would actually be a cat).
However, if you attempt to improve your profitability by improving customer service, it’s probably going to end in tears because my hunch would be that both are merely indicators of other factors such as product quality, what competitors are up to, how customer interactions are handled at both macro and micro levels, and so the list goes on…
I think that the stem of all this is the fundamentally poor understanding that most people have of statistics. I know that I would need to look up how to verify statistical relevance of a sample, and I am sure I’m not alone on that. If I were, there would be far fewer people doing the lottery…
mmitII 
On the customer service front, I’m reminded of one of the few things I remember from my own studies. This was from the Harvard Business Review and the legendary Ted Levitt, speaking about services:
“The most important thing about intangible products [services] is that customers usually don’t know what they’re getting until they don’t get it. Only then do they become aware of what they bargained for; only on dissatisfaction do they dwell. Satisfaction is, as it should be, mute. Its existence is affirmed only by its absence.”
People put up with low frills/questionable service from Easyjet (to extend your Ryanair point) as long as its cheap and reliable; when they started experiencing huge amounts of delays, that’s when people got the hump.
You get superb service in Apple stores (they recently replaced, FOC, a Macbook battery even though it was 3+ years old) which reinforces the premium positioning of the product and brand. But we’d still buy iAnything if that wasn’t the case for other intangible or product reasons.
Might also be worth re-looking at the Client Experience Model that I talked about a while ago here; https://mmitii.wordpress.com/2010/06/30/managing-the-client-experience/. That’s a whole way of thinking about making sure the intangible is tangible before it’s too late…