On averages and risk or skewed business models 3


Last week after a great (even Magnificent) visit to the Magnificent Maps exhibition at the British Library with Gary Gale, Steve Kennedy and Giuseppe Sollazzo we sat out in the courtyard rabbiting about maps, mobiles and stuff.

GeoRabbiting at Magnificent Maps by Gary Gale

Steve was explaining to us how the broadband ISP’s who purchase capacity from BT and buy in bulk but on a volume used basis (GB or similar) and then resell to us consumers on an all you can eat basis for a monthly fee. Old style guys (like me) might imagine that you could lose money when you buy on a pay per use basis and sell on a flat monthly fee but the business model depends on the fact that most people do not hammer their connections so on average you make a margin on their business. But there is risk – what happens when there is a sudden change in the behaviour of your customers or at least some of them? Along comes BBC iPlayer and suddenly bandwidth is being hammered by a lot more users and all of a sudden the numbers don’t look so good for ISP’s let alone when you add in the other on demand services that are coming to market.

Another example of the law of averages was the early SMS based version of Twitter, they were actually paying the mobile operators a tiny fee for sending SMS messages to subscribers. You can see how that model was not going to scale well as their user numbers rocketed and usage also exploded even at tiny rates intensive users would become financial burdens.

My old company once made a deal with a customer whereby the customer got unlimited access to our service for an annual fee and we were in turn buying transactions from a SaaS supplier/provider in bulk. The deal was profitable at about 4 times their realistic volumes but it still left a chink of risk if usage suddenly exploded way beyond our expectations. I never felt completely comfortable with that exposure even though the deal was highly profitable.

In business we have to take risks but it helps to understand the scale of the risk and its potential to do your business serious harm.


3 thoughts on “On averages and risk or skewed business models

  • Giuseppe Sollazzo

    To go a step further, I reckon this is why all the ISPs add a “fair usage policy” in the small print [1] but then don’t enforce strictly these policies: they’re still outselling what’s available to them thanks to the law of averages, but I’m sure they would start enforcing the limit the day they realise users are massively getting over that limit (BBC iPlayer being possibly the first case in which we risked such a scenario).
    The difference I see in here is that for a small company or a start up is even more difficult to make calculations over fair usage policies, and to apply them in a way that doesn’t make potential users run away.

    [1] as far as I know, giffgaff still offers a ‘real’ unlimited data package, but I’m not sure they’ll keep on doing that

  • Ian Painter

    Interesting post, I think there’s going to be a lot less ‘eat all the can services’ and eventually a move towards acceptance of per transaction payment (even down at the micro payment level). That said I can’t see government taking this up model as they cannot fix expenditure. For now, I think the banded model will be used more and more. Salesforce takes a nice approach with its customer portal and lets you go 3 times over your allotted use. Do that 4 months in a row and they charge you 1.5 times for the additional use. I like the fact that you can ‘mis-behave’ a little 😉

  • Thierry_G

    If a risk feels comfortable it’s not a risk 😉

    Interesting post, it confirmed my suspicions and probably explains a thing or two about customer service… Where good customer service is important is still justifies a healthy premium in my view.

Comments are closed.