As I mentioned in the earlier post – quite a few of us at Concise are interested in probability distributions – i.e. how to calculate the probability of a certain event happening in all sorts of different markets.

Now, it may sound a bit dry but it’s absolutely fundamental to the way we make predictions about the future. And, as I hope you’ll see, it has profound commercial implications for commercial strategy.

[Plus, a lot of people get it wrong so it’s one more situation where being a bit of a geek gives you a little bit of an advantage.]

I can’t remember how I was taught probability, but nowadays it feels like the first, and most consequential, thing you should learn is that normal distributions behave very differently to power law distributions. I say most consequential as, according to Nicolas Nassim Taleb (http://www NULL.fooledbyrandomness NULL.com/), economists making this mistake led directly to the credit crunch. If you want to read about that get his book. It’s very good. I’ll try to write a blog post on it soon.

Anyway, the second, area of debate is not about the shape of a curve, it’s about the degree of inequality, the steepness of its curve. And that’s what we’ve been talking about a lot recently.

For example, everyone reading this has probably heard of the Pareto principle i.e. that 80% of your sales come from 20% of your clients.

But what if in some markets those numbers were 90% and 10%? I guess you’d really want to focus on the top clients. Similarly, if 60% of revenue came from 40% of clients, then you would probably distribute your efforts more widely.

Essentially, this is what Chris Anderson wrote about in “The Long Tail – why the future of business is in selling less of more (http://www NULL.thelongtail NULL.com/)”.

He argued that online media retailers like Amazon and Neflix – who are able to offer a massively increased range of products compared to a traditional retailer – would enable consumers to buy more obscure titles in the “long tail”. This availability, together with the abundance of information and reviews, would mean that the retailers would make more money from the obscure titles and have a more even revenue ratio.

Not everyone agrees with his analysis though. Looked at another way, better information means consumers will use reviews and personal recommendations to zero in on the best products more quickly than ever before. Success snowballs and hits would become even bigger. Rather than the “long tail” this would be described as a superstar market.

Essentially this is a debate between two shapes of curve, the superstar one steeper than the other.

Evidence actually suggests that both interpretations get something right. Anita Elberse, in an excellent article (http://hbr NULL.org/2008/07/should-you-invest-in-the-long-tail/ar/1), analyses data on DVD rentals and music purchases and finds that, yes, obscure titles are doing better but this is at the expense of mid-weight successes rather than the hits which are doing better than ever before.

So, although Chris Anderson is right to predict that obscure titles will do better, the revenue ratios are actually getting more extreme.

What might this mean for businesses?

A long-tail strategy – in which new products are produced for niche audiences – starts to look pretty risky. Indeed, as Anitia Elberse says, it is far less risky to widen the appeal of an existing hit.

This is especially true in industries in businesses like cars, computers, TV’s, and mobile phones, where consumers cannot try both the hit and the obscure choice. Invariably they will choose the hit.

The question, of course, is whether new technologies offer any new ways of creating that broader appeal?

Let’s remind ourselves why Chris Anderson started talking about the long tail in the first place…

He is certainly right to say that consumers have better information about products. As we have seen though, this can lead a lot of them to pick the hits.

Interestingly though, they also have more information about each other. They can form communities of shared interest rather than being limited to people that live near them. Communities that would not have been viable in the past are now thriving, communities that were thriving have grown genuinely powerful, for example Mumsnet.

Many people within these groups may be considering buying one of the hits; maybe all they need is an extra reason why it’s right for them.

A canny manufacturer could reach out to a big enough community and find out what that reason might be. And then, once they have built it, they know exactly where to find the audience.

In this too, a hybrid strategy is the right solution – focus on the hit but take advantage of new technology to reach out to the bigger niche groups.

What else might be becoming less equal?

Hopefully by now it’s clear why we’re so interested in the way these distributions are changing.

The great thing is that once you’ve got used to the idea, you start spotting similar situations everywhere.

For example – what effect might social media have on the distribution of advertising success…?

In the past brands bought TV, press, or outdoor and they knew roughly how many people would see their ad.

Nowadays, a lot of brands would like to earn either a part or all of their media by having their ads passed from one person to another.

A nice idea in theory. But if the ad doesn’t get passed, people don’t see the ad, the production budget is wasted, and the brand is worse off than they would have been buying TV.

This would be ok if the ratio of success to failure was pretty shallow – a large number of campaign could be confident of success.

But if the ratio is steep and getting steeper then we’re going to have a very small number of massive successes and a very large number of that get very little return at all.

Maybe that’s why the same one or two Burger King campaigns keep getting brought up in strategy meetings…

[We’re doing some analysis into other less well known campaigns at the moment so watch this space.]