Monthly Archives: June 2006

Kevin Lee of Did-It.com analyzes for Clickz what “ad servers” need to do to win in today’s day and age of online advertising.  By “ad servers” he’s referring to the search-based advertising networks run by Google, Yahoo, MSN, and display ad networks such as Valueclick, Burst, and others.

His answers are:

• Critical mass of serious advertisers
• Critical mass of quality ad impressions
• Ability to detect click, impression, and action fraud
• Ability to accept medium-quality publishers without value dilution
• Ease of use for advertisers (absence of friction)
• Targeting efficiency, yield maximizer, and relevance engine

That’s a tall order, and I think he’s mostly right about all those items.  There isn’t a single network right now that’s executing really well in all those areas.

However I did want to point out though how this is an example of why an advertising exchange makes more sense than a closed ad network.

Critical mass of serious advertisers

This is a key problem for most networks.  Very few networks have the problem of too many advertisers, and this is especially important if you’re doing contextual targeting as you need more advertisers to cover all the possible different content areas that exist on the web.

An advertising exchange helps solve this by bringing together ad networks and advertisers onto one platform.  As an example the Right Media Exchange has over 50 ad networks participating which equates to over 2900 advertisers.  If those networks were all fragmented closed networks, publishers working with each individually would  have a much smaller amount of advertisers to fulfill the inventory.  It’s that many more ad sales people selling ads onto one platform, providing advertiser critical mass.

Critical mass of quality ad impressions 

Same as above, networks bring their publishers into the exchange and there’s now a critical mass of 10,000 publishers and growing.  The exchange on it’s own also becomes a draw to publishers due to the number of advertisers and ability to fulfill normally unsold inventory.

Ability to detect click, impression, and action fraud

An exchange allows the participants to share information and data that can identify publishers who are causing such issues.  It also increases communication between networks where because ad deals are often passing from network to network hurting visibility, it’s easier for fraud to occur.  Depending on what type of exchange it is, if it’s not entirely CPC pricing there is less chance that click fraud will occur.  For example, the Right Media Exchange actually uses four pricing types, and publishers don’t know necessarily if that individiual ad creative is CPM, dynamic CPM, CPA, or CPC.  Making it hard to know which type of fraud to commit.

Ability to accept medium-quality publishers without value dilution

This is probably exchange-specific, but I know in the case of the Right Media Exchange because each ad impression is auctioned in real time, impressions are paid what they are worth to the winning advertiser at that point in time.  This helps value dilution because lower quality inventory doesn’t have the same sucking of dollars that could occur in other types of optimization.  Essentially, a site with low quality inventory isn’t getting an inaccurate amount of the ad dollars in the system at the expense of quality publishers.

Ease of use for advertisers (absence of friction)

As someone who has advertised on numerous platforms, this is one near and dear to me.  There are certain ad systems that I just couldn’t spend the time to learn and master, and those systems that were easiest to use got the majority of my dollars.  An exchange doesn’t really have an inherent huge value over a closed network in interface design, although one could argue that an exchange being open means it’s more likely to provide APIs for easy integration with other tools advertisers may be comfortable with.  It’s also possible that entrepreneurs will see the opportunity to create better interfaces with an API to these platforms.

Targeting efficiency, yield maximizer, and relevance engine

This is another one in which there is a big opportunity for exchanges with APIs to not only provide their own targeting/yield/relevance algorithms, but for other companies to build their own solutions on top of it as well.  Why is this important?  Well, one company’s algorithm may work great for a certain type of advertiser or publisher, and not so well for others.  As an example Google’s contextual algorithm can work nicely for a web site about a niche topic, but not work very well for a company like MySpace where the profile pages are hard to target contextually relevant ads to because of the diverse content on each profile.  So if an exchange could theoretically allow publishers and advertisers to select from various algorithms, they might have a winning formula right there.

Kevin’s summary is right on the money:

We can expect to see major improvements in non-search relevance and targeting in networks over the next year. The consumer will win as ad relevance increases, but the real driver will be money. The networks understand that in the end, all other things being approximately equal, the best ad server wins.

It’ll be fun to watch and participate in.