• by David L. Smith , January 14, 2002

The argument over whether sites should take CPA deals, and the sites that come down on the side of not taking CPA deals misses the point.

Flash bulletin: There is a lot of unsold inventory out there. I know that may be shocking to some of you, but it is true. Further, folks with the unsold inventory could do something to get some income for it. Probably substantial income. So why don’t they do it? Part of it is stubbornness, part of it is pride, and part of it is a lack of understanding as to how other media have worked in the past. Yes, it’s true, we can learn from traditional media. We were going to start with a clean slate with the Web, ignoring the pitfalls of traditional media. But what about the solutions that offline media has. Can we borrow some of them? We certainly can - from an inventory management standpoint.

It is going to take some adjustment on the part of all parties involved. This includes agencies, clients, sites, networks, reps and even third party servers. But if we work together, we can solve this and ALL do more business and return more to the bottom line. Because, after all, if you are getting nothing for some inventory, isn’t it better to get something?

Let’s start with the agencies and clients. In traditional media, the model is called PI or per inquiry advertising. It works like this: We have all seen the 800 phone number advertising on various Turner networks, especially in January when broadcast and cable and are not sold out. How do they make these deals? Turner has long had a policy and a group to deal with this. First, an advertiser or agency must produce a commercial that they think will work. Then, they must test in successfully with a paid schedule. Now, this does not have to involve the expense to be a big national schedule. It could be tested in Cleveland or Denver. But it must be successful. And the advertiser and/or agency must be willing to share their results with the Turner organization. There is a significant proof process that one must go through to get their commercial accepted on a PI basis, but once it is done, it results in an agreement between Turner and the advertiser/agency that they will be charged based on the parameter agreed to, e.g., so much per sale of the item in question.

Major companies have built their business on this concept with Turner, other broadcasters and with print. An example is Charles Schwab. Prior to 1991, with the exception of the Wall Street Journal, MOST of the Schwab broadcast advertising was PI. I know that Web advertisers don’t think that CPA/PI advertising works relative to awareness and branding. But the fact of life is that, according to a Schwab annual report, their awareness in the early 90’s was an impressive 23% with this strategy. Sure, they eventually did turn to significant paid advertising to take them to where they are today, but they built their business on PI, the offline equivalent of CPA.

Web advertisers and sites can learn from this. Too much of the Web CPA market today is predatory, putting all of the risk on the site side. (Full disclosure, Mediasmith pleads guilty of this too, but we are doing our best to educate clients on the new model).

What if an advertiser established a paid track record prior to asking sites for a CPA deal and shared the confidential results with the site? Isn’t it possible that the major sites that currently reject CPA might consider a deal where they had evidence that their efforts would pay out? Think of the sales that could be made by the advertiser, the new volume that could be done by an agency and the unsold inventory that could be put to use by the site or network. Disclosure of information results in an “everybody wins” situation.

But wait, there’s more. The third party ad servers suffer when CPA deals are deployed. Reason: the agencies and advertisers do not want to pay the third party server on a CPM basis if the number of impressions cannot be defined in advance. There is too much risk involved. If a site runs tens of millions of additional impressions (of unsold inventory) in order to make the advertiser’s deal pay out, there is a risk that the cost of ad serving could be astronomical in relation to the value of the media originally contracted for. We have seen this happen and it is not pretty. So the agencies use “click command” or embedding of the “Web bug” on the banner and have the site serve the banner. The agency and advertiser are financially protected in that they are only paying for clicks. But there are two losers in this scenario: The third party ad server gets less income. And, the agency and advertiser do not get the metrics from the ad server on post-impression or “view through” tracking which could be half or more of the campaign, thus decreasing the chance that they can document that the buy is paying out.

The ad server community needs to recognize that the market has been changed from the time that their model was established, circa 1996-98 which was CPM based. In a market where CPA is used by many advertisers, the ad servers need to get creative on their pricing and figure out a way to get their fair piece of this income stream. The result would be the ad serving companies getting more money and the campaign having a better chance of paying out with both post-click and post-impression volume attributed to the campaign payout. The net effect produces more volume and revenue for everyone.

Wait! you say, not all advertisers can stick with their tested creative. Web creative is fluid and dynamic and needs to be changed often to maximize the chance for success. Every option cannot be tested. True. That’s where the hybrid deal comes in. The hybrid deal guarantees the site some income based on a lower CPM that the site and agency/advertiser can live with. Then as a plateau in actions is achieved, the site gets incremental income, based on performance. Everybody wins.

A combination of the PI model and the hybrid deal is not out of the question. We all need to rethink the way that we are approaching the media buy when CPA is a factor. Create a situation where all can win, and you just might win too.

David L. Smith is President of Mediasmith, Inc., the Integrated Solutions Media Agency based in San Francisco and New York.