I\’m sure there\’s something wrong with this idea but I don\’t know what it is

Essentially, a new method of retail banking.

The banks have vast amounts of information about their customers. Who they pay, how much they pay them etc. So, use that information to run comparison engines. Split the savings between he bank and the customer.

I would assume that the problem is somewhere in the Data Protection Act.

As to whether it would work: well, imagine that you were in fact running a comparison engine. Say, Go Compare. Then someone said you can mine the retail customer information at Lloyds to t3ell people how and where they could save. You\’d do it like a shot, wouldn\’t you? Zero customer acquisition cost and millions upon million of new customers.

There is therefore a business in there.

So, what is it I\’ve missed?

Maybe the banks simply aren\’t competent to do this?

12 thoughts on “I\’m sure there\’s something wrong with this idea but I don\’t know what it is”

  1. The banks can have only the vaguest idea what’s bought (unless the customer actively tells them). E.g. how many kWh have I bought for my £x eletricity bill?

  2. This already happens, to greater and lesser extents. The problems are mainly to do with collecting enough data to do really good profiling. But yes, Google Bank is on its way.

  3. Neil>

    I’d wager that Google’s profile of you is already good enough to make a decent guess at your electricity usage.

  4. Taken to its logical end point, a bank should recommend that its customers move all their savings to any other bank with a better interest rate.

    Anyway, why would I trust a bank to tell me which is the best utility supplier to use or what. I pay my bills with as low a direct debit as possible (to benefit from direct debit discounts) then use credit cards to earn points on the remainder.

  5. This seems to fall foul of the pretty standard fallacy that data==knowledge, and oodles of people still believe this (including big people like Google) either because they’re a bit dim, or haven’t studied artificial intelligence.

    The AI project so far has had only one really useful result, which is that intelligence isn’t data processing, at least not algorithms and syllogisms. People often think it’s just an insufficient data problem, but we’ve gone from kilos to megas to teras and the dream of algorithmic intelligence is still no closer. And that is the wheel on which every data mining dream- including the bullshit emanating from Google at the moment- is broken. Data doesn’t supply understanding.

    Or to put it another way, I’m currently seeing adverts for brassieres everywhere i go on the web. It took me a while to remember why; reading a discussion about old 1960s pointy bras, and somebody linked to the Playtex website.

    So now the advertisers have the information that my IP looks at bras, so the algorithm says I might be in the market for a bra. Because it has no knowledge of me, or of the world.

    People always think you can improve this by adding in more variables (“ah, but if it knew you were male…”) but that doesn’t fix the problem.

    So, having waffled a lot already, what this actually tells us is that mathematics cannot model society. Which is why all central planning fails. Hence, having the banks take on central planning via algorithmic data mining won’t work either.

    Sergey Brin started off with data mining, btw. That’s why the more sophisticated Google gets, the shittier it gets. Broken paradigm.

  6. Ian>

    That’s just daft. You might as well argue that, I don’t know, coal wasn’t a key factor in the Industrial Revolution because people still had to do something with it.

    Data mining and analysis is still at a stage where saying it’s in its infancy would be a vast overstatement of its maturity.

  7. Males can be in the market for buying a bra, for the significant other. Plenty of marketing designed for males to buy females stuff.

  8. Data mining and analysis is still at a stage where saying it’s in its infancy would be a vast overstatement of its maturity.

    No, it’s not that. It actually doesn’t work. That’s the basic lesson of AI. Or one of them, anyway. And it’s fundamental. It’s why we keep hearing fantastical ideas for future data systems, and they never come to fruition; think, for instance The Semantic Web. It can never work.

    Doesn’t matter how many more orders of magnitude of data you acquire, it still won’t work. There isn’t any understanding in the data, and there never can be.

  9. Ian>

    No-one suggests there is understanding in the data. The understanding and analysis is creative work. It is very possible to draw all kinds of conclusions about someone if you have enough data about their life. It’s hard to generalise those methods such that a computer can do the same thing from limited data, but the principle is sound.

    To give a silly example, ‘enough data’ might extend to following someone around their entire life – but when you see a light-bulb blow in their living room, followed by them getting a new one from the cupboard, you can very reasonably predict that they’ll change the dead lightbulb for the working one they just took out.

    That example, silly as it is, demonstrates the principle.

    To give a slightly less silly example, where the prediction is less accurate – but one that actually exists in the real world – google can gather enough data to say whether or not people who search for ‘travel agents’ are more likely than the general population to buy a holiday in the relatively near future. As it happens, they are. It’s significant but not overwhelming, so holiday-sellers are willing to pay slightly more to advertise holidays to those people.

    Am I talking about something different to what you’re talking about? It seems to me that this is no different in principle to such well-established concepts as actuarial tables for life expectancy or car-accident risk.

  10. The reason I don’t think this will fly is that you’d need to trust the bank not to cut corners by colluding with the providers.

    Which I don’t, for a moment. The reason I “trust” banks with my savings is that I’m reasonably sure that most of the obvious and mildly subtle scams have been worked out of the system and that the courts will come down like a ton of bricks on a bank that tries outright stealing. No such mechanism exists for this – try to prove that collusion legally, will you?

    Besides… supposed I am so frugal as to appreciate the 0.05

Leave a Reply

Your email address will not be published. Required fields are marked *