A Few Scary-Good Ads: the Cure for Advertising Blindness

I would have to agree: "we [the general population] are getting to a point in our internet usage where we've developed advertising blindness." ( Aaron Hoos) My Jellybattle was exactly that: we don't see ads anymore.

Still working on yesterday's Facebook example…

Let's say Facebook got really, really good at targeting. Scary good: like Pandora (link not available outside US) does for music. You fire up Pandora, give it a few songs and, through the magic of the Music Genome Project it gives you suggestion for music. Many other online music services do this (Yahoo Launchcast, Last.fm, etc.). But Pandora is different from those other ones: it's good, scary good (to be fair, I haven't tried Last.fm - it could be good too).

When I started with Pandora, I punched in a few of my favorite songs that I thought were similar (there are many different pools I sip from musically, so I grabbed an "easy one" and gave a sampling to Pandora). It started me off on a favorite song - not one of the ones that I gave it, but a different one by the same artist: an easy win for Pandora (and yes, I'm purposefully not giving any specifics here for fear of persecution for my musical tastes). But then it moved to other artists and songs. Artists that I had never heard of.

One artist in particular, wasn't new (they have 5-6 albums available from Amazon), but never got airplay on any radio stations that I have heard. I loved them. They were somewhat similar to one of the artists that I had entered, but better. Better? The algorithm found something better and that artist is now part of my "list of bands I like."

This story, by the way, happened over a year ago. I still remember it because it was so good. In fact, even though I can't listen to Pandora (over 6 months now without Pandora) - here I am talking about it. (And I'm not the only one: this week I was talking to someone about how we miss having Pandora - I have my entire CD library ripped onto my computer, but I miss Pandora for the new fresh music.) And that is just one example: it showed me a whole new artist who I now greatly enjoy, but there were many songs during my time with Pandora that I also enjoyed.

My point is just that you can have algorithms be scary good (although, to be fair, Pandora is more than algorithm, it uses people to categorize the music - but I'd argue that their taxonomy of music and their process for adding people to rate is the algorithm if a little more "lo-tech").

Meanwhile, back at the Facebook ranch So, let's say Facebook starts the "Ad Genome Project" and makes an algorithm that gets good results. It's scary good. You visit a friend's profile page and they show you an ad for a Frogpad. That product is interesting to you: it's already on your Amazon wishlist. And, not only is the ad for the Frogpad but the RadioShack down the street from your work has it on sale and they have that information too. And here's the phone number.

That's scary good. (I think it can get even better than that.)

Currently, the advertising headspace runs like this (from my limited understanding):

  • "Ads on TV or in magazines will get seen by a lot of people, but we don't know who and we don't know how much impact they have. We do know they make a difference, even if less than 1% of people who see the ads do anything - it doesn't matter because there are so many people seeing the ads that it makes a difference."
  • "Ads online are better: we can target them so that we get a higher response rate. It's still pretty low [when was the last time you clicked on a Google Ad in their Search? Yeah, thought so.], but better than traditional media. And we know when it is working and for whom."

So, let's say Facebook, being online, is in camp 2 but they have a better pitches (equal or better algorithm + more information). If they can get dramatically better targeting and therefore more likely to get clicks, why wouldn't they server fewer ads.

Again: serve fewer ads.

Fewer Ads: More Power Not only would this help with the ad blindness that we have - but, by making ads uncommon on Facebook I think we might be more attentive to those few ads. If, of course, they are more applicable.

In IdeaVirus, Seth Godin talks about different kinds of people who spread an idea: promiscuous sneezers and powerful sneezers (he uses a cold as the metaphor for how ideas spread: hence the "sneezer" is a spreader of the ideavirus).

He states that promiscuous are motivated by money (e.g. pay for your ad and I'll run it) and "are rarely held in high esteem." Powerful sneezers, on the other hand, can't be bought but has more influence: people listen.

I'm not suggesting that Facebook can swing all the way over on the pendulum: from their current location at promiscuous to powerful. But, I do think that by running fewer good ads that are relevant (seriously, I am not going to go to the University of Phoenix online: stop telling me about it) they can increase their power. It's not an either/or: it's a continuum.

So, if you run fewer ads but they are generally targeted what happens? I start paying attention because I have learned through experience that the odds are I am going to be interested in an ad from Facebook. In fact, if you are really good and only show me the really good stuff, I'm like to trust you. I'll see an ad, won't really know what it is for but will think "they are usually interesting, I'll click to see what it is." That's power. Pandora had that with me: even if I had never heard of the song, I'd give it a chance - because Pandora usually suggested good stuff.

Experience is how we got into this mess in the first place: we've learned that, most of the time, ads are useless and not interesting. Ads on TV, in magazines, on street-side billboards, online... Ads are what someone wants me to do: not what I am interested in. We've learned that lesson well and are adept at tuning them out.

So, when I can look forward to fewer ads and good ads in my online experience?

Update: More discussion on news & advertising on Facebook from GigaOm

Wednesday, October 24, 2007, 12:00 AM

tagged: aaronhoos, algorithms, pandora, socialweb, webadvertising