Inventory and Risk

Google made some waves by expanding their PPA program. Here’s the way I think about online ads:

  • PPV (pay per view). This is the traditional way ads are sold in other media. Mostly because it’s the only choice. You charge for every thousand people who ‘see’ an ad, where see is a pretty vague term that means circulation or estimated eyeballs.
  • PPC (pay per click). This is one of the core elements of Google’s success. The publisher charges not for views but for someone actually responding to the ad by clicking on it. (We helped create this notion in 1998 at Yoyodyne).
  • PPA (pay per action). This is the new new thing which is actually not so new. In this case, the marketer only pays when someone actually does something valuable, like apply for a mortgage or buy a book from Amazon.

PPA isn’t new of course. Ever since Amazon has had an affiliate program there’s been widespread PPA. Put an Amazon link on your site (there are plenty to the left if you want an example) and if the book gets sold, the site makes money.

There are two forces at work here: risk and inventory.

If you’ve got a ton of unsold inventory, you’re willing to take more risk with your ads. Hey, if no one buys an ad, you’ve wasted that moment, wasted that impression. So you might as well take what you can get.

Google’s network has a ton of inventory. Vast number of page views, perhaps in the bazillions every day. With that much supply, PPV (banner ads) won’t work. I remember when banner ads cost $60 per thousand. Now they’re often down to a penny per thousand. All because they don’t work so well and because the supply is so high.

PPC was a brilliant compromise, because both sides take a risk. The publisher is risking saleable ad space on the hope that the ad will actually lead to a click (and then they get paid). The marketer is risking the fact that many clicks don’t actually turn into something valuable, in which case they pay for clicks that don’t pan out.

PPC rewards publishers that create content that attracts action-oriented readers. PPC rewards marketers that create ads that actually match the content on their landing pages, and rewards marketers that convert surfers with attractive offers of permission.

Of course, this sometimes leads to click fraud, where people click on ads with no intention of taking action. I think click fraud is overrated as a problem, certainly compared with circulation inflation at magazines or bored surfers or couch potatoes.

In a market filled with supply, though, it’s easy for marketers to demand a better deal. So some buy PPA ads. Amazon, for example, has nearly a million affiliates. From a marketer’s point of view, no need to be picky. If someone wants to sign up, let them. There will be no click fraud, of course, because you only pay if someone buys something.

Which leads to the biggest reason that this doesn’t scale. There’s a huge incentive for marketers to cheat. Marketers have an incentive to run PPA ads that don’t work so well… unless they work great. For example, a car dealer should do a PPA ad that pays when a car is sold, not when someone calls the dealership. Why not?

There are very few sites built on Amazon affiliate revenue alone, mostly because the affiliate deals don’t convert often enough to make them worth what’s paid.

Over time, the interests of the publisher and the marketer diverge. And the user, the surfer, the person that all this is directed to, is the ultimate arbiter. If the ads are boring or dishonest or ineffective, we just ignore them. And then everyone loses.

I’m not ‘against’ PPA. Squidoo is filled with affiliate links, after all. I think, though, it’s premature to expect that many ads are going to switch to this model. And I guarantee that the last publishers to switch will be the ones you most want…selected, focused, high-demand media that actually works.

One last thought: Gil is waiting for Google to go Buddhist and offer PPN: Pay per no-action. Breathe in, breathe out.