Tuesday, June 10, 2008

Recommendation: it's all in the mind - theirs not yours

When I started out in the world of work, as a journalist, we had high principles about the distinction between advertising and editorial.

Where those lines blurred we had a blurred word to describe it: advertorial. An advert dressed up as editorial. A subjective advertising message posing as editorial recommendation.

We thought editorial should entertain and inform, be useful to you. In consumer media it was important to us that our recommendation should not be tainted by commercial concerns (beyond trying to sell newspapers, magazines etc of course!).

Advertising was clearly and explicitly about selling something, nothing more.

The advertorial aimed to do a little of both. And this adulteration made it something of a dirty word to editorial people. Written through gritted teeth, we felt as if we were perpetrating some kind of con trick.

And we held steadfastly to the control of the metadata so we could raise alarm bells for our readers. Warning: You are being sold to. We are being paid to write this. Warning! Suspend your trust in us!

This metadata was expressed through a label 'Advertisement Feature', for example, and the use of clearly different design cues to distinguish it from the pure editorial content. We wanted to be clear this wasn't really our recommendation.

That carries through in most advertising mediums. There is consistent metadata which makes it easy for my 3-year-old daughter to distinguish the difference between a TV advert and a TV programme within seconds. Ads online comply too. We feel guilty if we make google adsense indistinct from the editorial it sits within.

Roll forward. The digital disruption has changed who gets to publish content (recommendations) and who gets to distribute that content - through peer-to-peer networks of trust. We've moved from a centralised broadcast model to an edge-in networked one.

The need for a distinction between advertising and editorial content is an outcome of mass broadcast models. Where control of production and distribution of the message was in the hands of the few, to get your message to the eyeballs meant you either had to own the means of production/distribution or pay for placement.

If you were paying for placement we should be warned that the message was likely to be biased.
So we created a set of rules, both implicit (metadata) and explicit (legislation) to regulate those placements and warn us of the risk of bias.

But it's no longer the case that you need to either own the means of production or distribution in order to get your content OR advert published and distributed.

In the networked world we are all creators of what would once have been called editorial messages AND advertising messages.

And where the payment distinction dissolves, perhaps so does the need for distinction between advertising and recommendation.

Where they meet we find services. What advertising message would you forward (recommend) to a friend? One you think they will find useful. How did you decide to turn that advert into a recommendation?

Well, that's entirely up to you. First it has to be something you personally found useful (a service to you). Being able to change it to better suit your needs (a widget with a personal outcome, for example) makes that more likely. Moulding it to the needs of your own network of trust makes it more likely you will decide it's worthy of your recommendation. Where the content becomes a service, the ad becomes more likely to morph into a recommendation. But there's no certainty.

As JP Rangaswami was suggesting here, recommendation doesn't happen in the mind of the forwarding party, or the person who would distribute. It happens in the mind of the receiver. It's pull. The forwarder, the wannabe distributor, is pushing - is advertising.

I guess that means that one man's advertising is another man's recommendation.

The fact is, peer-to-peer distribution models are likely to make a better fist of selecting the right 'adverts' to forward to the people most likely to regard that service as a 'recommendation'.


  • Your service must be adaptable (so the user can make it a better fit for those he would advertise it to).
  • Your service must be peer-to-peer distributed (it's the best fit with the networked world and the only way it can morph from advert to recommendation).
Remember, this model is not likely to fit the broadcast world. And not everything digital is a pure networked model.

Models where audiences consume content (eg everything from straightforward broadcast models to youtube - where 1% of creators forge a ugc-powered broadcast model for 99% of consumers) more traditional ad models continue to create value.

But in Total Communities - where 99% are creating and distributing (eg myspace, twitter, facebook, social networks in general) where to take part you have to create part, making it ever easier to turn advert into recommendation is the new alchemy.

I'm going to be talking more about this at WidgetWebExpo in New York next week. In the meantime, please join in the journey by commenting below.


The rate of change is so rapid it's difficult for one person to keep up to speed. Let's pool our thoughts, share our reactions and, who knows, even reach some shared conclusions worth arriving at?