• by David L. Smith , October 14, 2002

We have written before about the value of online media optimization. For most media, planning has the most clout, closely followed by buying (the operable assumption being that it doesn’t do any good to get a great price if it is on the wrong inventory).

While we have been evaluating optimizers for what seems like years, it all came to a head for me a few weeks ago at a Gaspedal event (worthwhile although the food could have been better). I had a conversation with a guy who said his company did optimization. He gave me his boilerplate elevator speech and I asked if they optimized on post event or post impression in addition to post-click. He said that they definitely optimized on post click and rewound into his elevator speech. I tried to get him to comment on post event tracking but I was wasting my time. This was the second time in recent months where somebody talked to me about their optimization capabilities and did not know what post event or post impression metrics were all about. Frankly, I thought we had washed out the know nothings from the business. Oh, well.

When I talk about optimization today, I am limiting my comments to optimization of banner campaigns. We do recognize the value of and use search engine optimization. There seem to be five kinds of banner campaign optimization:

1. technology included with an ad server
2. standalone technologies which do not provide ad serving but work with third party ad servers
3. standalone technologies which work with third party ad servers or can act as a server and optimizer solution
4. companies who tout optimization but really want you to turn your media buying over to them and
5. hiring some really smart college grads to watch your third party ad server data and make decisions accordingly.

All of the ad servers now have some sort of optimization included with them. This includes BlueStreak, DoubleClick’s DART, Atlas/DMT, MediaPlex, etc. These tend to be rudimentary at best. The Atlas method does not optimize but makes recommendations to the human who optimizes. Using your ad server as an assist seems to be the only affordable technology available.

Paramark and Poindexter each have standalone optimization systems that work with third party ad servers. However, they are both more expensive than ad serving. This makes it extremely high out of pocket (hundreds of thousands of dollars per year with any kind of volume) and thus, hard to pass through to all but the most highly educated client. Keep in mind that with the low CPMs that are occurring these days, even third party ad serving fees are under attack from some clients. We’ve had buys so efficient that the third party ad serving fees alone were greater than our commission. Poindexter can also serve the ads, but the combined serving and optimization charge makes me wonder whether they aren’t really just trying to get your ad serving business at a premium.

Advertising.com would love to do your optimization. IF you turn your buying over to them. This is so that they can do the buying, ad serving and optimization. This is a sales ploy as old as ad sales itself. If it really worked, there would be no need for media departments. With all due respect to advertising.com, do you really trust them to do the right thing when they will make more money buying their sites than others?

As usual, this all comes down to money. When you do the math, you realize that you can hire a number of smart people to watch campaigns and do optimization for the price of these technologies. Some of the technologies like Poindexter are truly neat. But I have yet to find an agency that can justify the cost of using a service like this on an ongoing basis. If you can prove me wrong, please do so. In the meantime, we are going to stick with keeping the human in the loop. Technology is wonderful, just not always better or more efficient.