The Friendly Shopkeeper Advertising Model

The dirty little secret that nobody wants to admit, amidst the Web 2.0 crowd’s rush to monetize any piece of content that isn’t nailed to the floor, is that advertising sucks. This isn’t a new problem. Prior to the Internet, you were lucky to get 2% response rates to advertising campaigns. The magnitude of the failure of traditional advertising only became apparent when two guys from Stanford realized that you could show ads based on people’s search criteria and provide much more targeted ads. While this observation only doubled response rates to a paltry 4%, that multiple is currently responsible for the billion dollars a quarter that Google generates in revenue.

But this isn’t really an achievement to celebrate, at least, not yet.

The problem with most advertising is that, on a macro-scale, it’s almost indistinguishable from spam. Think of all the advertising that gets thrown at you on a daily basis: the flyers that show up at your door, simply because you live in a certain neighbourhood; the brochures and pullouts stuffed in magazines and newspapers that pitch you products simply because that magazine’s readership fits a certain demographic profile; and let’s not forget the billboards, the radio ads, and, yes, even the urinal posters. These woefully unwanted come-ons inadvertently communicate one unavoidable message to a consumer: that the advertiser, while desperate for your business, knows absolutely nothing about you.

There is a better model of advertising that’s coming, something I like to call the “friendly shopkeeper” model of advertising. Allow me to elaborate…

There was a time when you would go into physical store and the shopkeeper behind the counter might know your name, but more importantly, remembered the kind of stuff you’d asked about in the past, and most important of all, the stuff you had bought in the past. This, combined with the shopkeeper’s own rich and deep understanding of new products and trends in the market, would allow the shopkeeper to provide a valuable service to you: the role of a trusted advisor. This person would helpfully suggest items that might interest you – for example: a new album from a band you’d never heard of, but that resembled something else the shopkeeper knew you liked.

Of course, the shopkeeper wasn’t doing this out of philanthropy, but out of a vested interest in providing a valuable service to his customers. The shopkeeper acted as an editor, weeding out the stuff that wouldn’t interest you, and suggesting things you might like. In exchange for this service, you bought the products suggested and, if the suggestions proved worthwhile, you came back and bought more stuff based on the shopkeeper’s advice. It was structured as a win-win scenario.

This is the model of the web that we should aspire to create. While many might, as seen in recent months, decry the idea of sharing information willy-nilly (cf: the Facebook Beacon fiasco), I’m not averse to sharing that information with vendors. But there’s a catch: I want some value out of giving you that information. I don’t want more advertisements, I want less ads, and only for items that I might actually be interested in.

In many cases, we’ve come close to this model, but only in specific product categories. For example: a number of services do a great job in suggesting music, and generating affiliate revenue when the listener buys a suggested song or CD. Unfortunately, none of that information ever makes it to other vendors where it might lead to better recommendations. If, for example, I end up buying a CD at Wal-Mart.com, that information never gets used to improve my Amazon.com recommendations. Instead, I end up with a number of different vendors pitching me the same stuff again and again because they have incomplete information about me. In the simplest case, they just don’t know that I already bought that CD. That doesn’t serve their interests (it’s highly unlikely I’m going to buy the same product again and again), and it doesn’t serve mine (because I’m not getting introduced to interesting new music).

In an ideal world, online vendors would know me better than I may even know myself. They wouldn’t suggest stuff I already own. They would know my friends and what they’ve purchased. They would even have noticed who among my friends are key tastemakers and influencers on my preferences to derive even better recommendations (after all, nobody chooses friends with whom they have nothing in common). They would amalgamate all of this information wherever it lives. It would do this in an aggregate fashion that, unlike Facebook’s Beacon, doesn’t reveal my purchases directly to others but instead uses that information to feed a recommendation engine with the fuel it needs to provide personalized, customized advertising.

In the end, what I want is not advertising, but something that achieves all the goals of advertising: matching people who have something to sell with people who want to buy it, but with 100% accuracy (or, at the very least, 0% annoyance and intrusion).

(This post was inspired by something I’ve been thinking about for a while, coupled with a desperate need to beat Ian Bell to the punch on claiming the “friendly shopkeeper” metaphor, something his company is working on achieving with their Pul.se Facebook application. Disclosure: Ian is a friend. )

Pandora: Good Enough for Miserly Bastards

I had a chance to meet Tom Conrad from Pandora last week at the dinner after Blog Business Summit, and then again at Bar Camp. Pandora’s service basically allows you to generate a personal set of “stations” based on a favorite song or artist – their backend does the rest!

Now, hear me as I speak with the authority that only a miserly cheap bastard can muster: this is actually something I would pay for.

Like any self-respecting music aficionado, I have a pretty big collection of music. But that amount of music has an unfortunate side effect – the more music you have, the less you actually listen to new music. Pandora solves this, feeding you a constant stream of only the good stuff. I’ve been listening to it the last three evenings, and it’s been a “set and forget” experience. Wow.

The part that’s really interesting about this service is that it’s a big of a legal hack – because of the way webcasting royalties work it costs Pandora almost nothing to provide the music itself (the cost is really in the bandwidth and the tech to provide accurate recommendations). The service, currently in invitation-only “preview”, is expected to be in the “couple bucks a month” price range. Totally acceptable.

The one thing I would like to see Pandora do with its technology is extend it to the music I already own. Why? Well, because of the aforementioned webcasting hack, there are certain restrictions on what Pandora can offer. It can’t, for example, just serve you the exact song you want – music on demand is verboten (which makes sense, given that webcasting rules are meant to be analogous to the rules for radio). Also, when streaming music to your machine, you can’t see what song is next, or play the previous song. A bit annoying. But imagine if you could had a Pandora plugin for iTunes that allowed you to autocreate playlists of stuff you owned that were similar/matched. You do this yourself currently, but what if Pandora could do it for you, and allow you to effectively take their service with you? Imagine how much music you currently own that you don’t listen to, simply because you inevitably end up playing the same stuff again and again.

Bottom line: Pandora is awesome – check it out. Uh, I mean, check it out when there’s a public beta. If you’re interested in additional details about Pandora, you might also want to check out this review at TechCrunch.