In the print world, once you lower your page rate, it is very difficult to raise it. Contrary to that is broadcast where rates fluctuate quarterly based on supply and demand. Online advertising inventory should also be based on supply and demand.

Today many publishers quote advertisers a high rate and sell a small percentage of their ad space directly to advertisers and media buyers. The unsold impressions are then offloaded to Google Placement targeting and/or the 100+ ad networks. The problem with this scenario is it generates huge revenue for Google Site Placement targeting and the ad networks but very little revenue for the branded web sites. It also makes the ad performance of Google Site Placement and the ad networks perform better than direct buys with sites.

I recommend sites lower rates to the advertisers/media buyers with the goal of an 80% sell thru rate and adjust the rate based on supply and demand of impressions. The remaining 20% should be used for remnant insertion orders and held for last minute IO's where the CPM can then be raised.

Remant impression insertion orders are insertion orders that are way below "rate card" that the site may or may not run depending on their supply and demand of inventory. Instead of relying on sub $1 CPM's that Google and ad networks pay sites for remnant inventory, sites should accept remnant impression insertion orders from their direct advertisers and media buyers.

It is much better to get $5 CPM for 80% of your inventory than $10 CPM for 20% of your inventory. Sure, you may feel some pride telling your peers that your site CPM is $20 ... but when you sell only 20% of your inventory, you are not only decreasing your revenue potential, but also undermining your advertising relationships -- the kind that develop into profitable repeat business. A lower rate will result in repeat business as the media will perform better for the advertiser.

With a high CPM, you are also missing sales without even realizing it. When media planners are in the initial stages of developing a laundry list of sites, sites with high CPM's may not even make the list, or get lopped off the top because of price.

If you need to charge a high CPM because your financial model requires it, but you aren't selling a high percentage of your inventory, then you need to rethink your financial model.

So what should the CPM be? Web CPMs should be based on the market supply and demand. This is especially true with sites that receive spikes in traffic (election sites, Forbes 400, Fortune 500 issue). Those sites should lower prices to maximize their total revenue by lowering their CPMS and limit the excess unsold impressions.

For most sites, if you charge a $1 CPM you would sell 100% of your inventory and if you charge $100, you would probably sell 0% of your inventory. In between $1 and $100 lies the sweet spot of optimum revenue for you and a happy advertiser base. As you lower your CPM you raise your sell thru percentage. If your salespeople are doing their job and you aren't selling a majority of your impressions direct, then lowering the CPM is the way to go.

With this model, changing the CPM on a regular basis (daily if you have to) is perfectly acceptable. If your site is 80% sold at $8 and an advertiser emails for rates, then quote a $9 CPM. If they buy great, but if they want lower than $9 then tell them you can take a remnant IO for anything less than $9 but no guarantee you can run it.

The overall goal is profitability and that is achieved by satisfying your advertiser base. And satisfying your advertising base is achieved when the amount of money they spend on your site performs better than the amount of money they spend on your competitor's site.

This commentary is insightful. I recommend it to others.