Cheap milk on my mobile...

Over on the Oxford Uni. next generation mobile apps forum, the question posed by Russell Buckely is what would it take to get users to sign-up for location-based ads. Some commentators think that "cheap milk here" ads from the local Tesco Metro won't ever make it. The future of loc. ads is special promos to the audience of a pop concert etc. Personally, I disagree - and if it's cheap milk versus Madonna wallpaper at her latest concert, I'll take the cheap milk....and here's why...

Firstly, a general point to make about location-enabling is that none of us has really experienced it. The reason LE apps haven't taken off, or aren't likely to, is because they're simply unusable in most networks - the accuracy is poor. I have argued since its inception that only GPS accuracy is going to work along with proximity-based services indoors. Once people become exposed to hyper-accurate services, they will not want to give them up.

By the way, I remember sitting through one of those boring operator developer forums and watching a presentation about how accurate LBS needs to be. Quite unbelievably, they did a survey and asked what kind of accuracy would you like....etc. They concluded with the answer they actually wanted, which is not to go spend money on GPS-accurate systems because no one has asked for them. That's taking paper studies too far! Sometimes you've got to give the service to the user and then ask for feedback. I am convinced that highly accurate services of all kinds will be compelling. Many LE services, including many ad-based ones, only come alive with highly accurate proximity sensing outdoors and indoors. I believe there are many compelling use cases, including "cheap milk".

So, now for the "cheap milk" argument. I believe 100% that this is the future of LE because I believe that the long-tail model fits here. It doesn't matter if there aren't 100,000 Madonna fans who want sell-off milk, just as long as there's enough. What's enough? Enough means that the price of advertising is less than the return. For example, if Tesco Metro can spend 5 pounds per store advertising day-end sell-offs and make more than 5 pounds - they're definitely going to do it because a lot of more-than-5-pounds adds up. Those of us who think "cheap milk" is spam, won't sign up. But to Tesco stores, that doesn't matter, as long as the economics still add up, which in the world of "pay per click" ads, it surely will.

That's perhaps not long-tail enough. What about the local plumber? If he can advertise that he's available in your area today (and God knows that plumbers are hard to find in any sense of the word) and he pays ("bids") 5 pounds for that hit that makes him 100, then he's going to do it too.

The problem with all this - and I've been writing reports for various customers about this for years - is that the cost and complexity of placing localised ads is currently prohibitive. There are all kinds of "messy" IT problems to solve . Take the Tesco example again. They will need to plumb their in-store inventory/EPOS system into the "ad service" and pipe that through the operator's network, including - probably - a revenue share back to the operator and so on. Average cost of such a project might be thousands of times the cheap milk sell-off margin over one year (or whatever ROI period accountants use these days).

Operator networks are by-and-large cumbersome telco networks, not nimble edge-of-network web servers and the like. However, that revolution is coming, albeit slowly. Operators are moving towards a Services Orientated Architecture (SOA) where eventually their whole business will be one big "software application" accessible via the net (web services etc) - as will Tesco and others. It is difficult to appreciate the benefits that SOA promises - and even operators don't really understand the transformative possibilities - but one of them is how previously very complex business rules and the expensive IT required to implement them will be as easy and cheap to implement as clicking a few mouse buttons. So, if the local BMW garage want to beam you an ad for a car that they already "know" you've been looking at on Froogle then they can implement that app/ad for next to nothing.

Many interesting mobile apps in the future can only come to fruition with this sort of complex multiple operation/business logic underneath. I believe that what we are headed for is a different question than the one posed by Russell. People will ask "why wouldn't I want location-based ads?" "Hello?" I suspect that Google already know this.

Buy my book (Amazon US/UK)
Join my email list
Subscribe to my "100 Mobile Product Ideas" free e-book