Over the past two weeks, we’ve been looking at ad pricing models, and how they work for the site publishers. Whether you’re a publisher choosing to offer CPA (cost per action) pricing, are responding reluctantly to advertiser demands, or selling on any sort of CPA program, you need some guidelines about how to charge and how to allocate the risk.
In the simplest possible calculation, a publisher whose site generally produces 1.5 percent clicks per ad campaign can take the intended impression size of the buy. Multiply it by 1.5 percent to get the expected click yield, then use that number to determine the equivalent price per click.
Where X = number of impressions to be sold,
Figure X/1000 * CPM = Z (pricing benchmark),
X * 1.5% = Y (expected click yield),
Z/Y = target cost per click
But what does this formula say about risk? Any decrease in the average click result is a risk for the publisher, even though evidence indicates that the ad creative is the largest determinant of click results.
So, publishers should protect themselves by calculating for risk (just as bankers and investors increase their share as the risk of the investment goes up). If your average click yield is 1.5 percent, consider doing the calculation at 1 percent, or .75 percent, or whatever level you consider likely for the advertiser in question. This is where accurate and detailed records of site performance really help. The more you know about your site’s yields by type of ad, product advertised, length or size of ad campaign, degree of targeting or any other metric, the closer you’ll come to setting a cost per click you can feel comfortable with.
What about when the advertiser wants to pay only for completed forms, or for sales actually transacted? The same basic calculation would work, except it’s the rare publisher who has detailed historic information for transactions in each advertiser category, and you can’t assume that the numbers provided by the advertiser are representative of your audience’s behavior. Let’s face it, advertisers and agencies do the same math, and they know that inflating the promise of units sold could lower their cost per transaction cost. Until you have a relationship with an agency or advertiser, don’t assume the expectations shared with you will match any reality for your site.
So, how to set a price? Be conservative about performance expectations and keep the duration of the buy short to allow you to gather data and adjust. We’ve seen far too many publishers brag about a six month advertiser commitment, only to realize in the third week that they will be giving away the store for the next 21 weeks. And make sure you have a way of tracking actual performance to your satisfaction – no CPA deal makes sense if you are taking the word of a dishonorable partner as to what actions took place.
Advertisers who are committed to a longer-term relationship will be willing to work with you, to adjust agreements as data is gathered. Lasting business partners know that both sides of the partnership need to win for a deal to make sense.
Some sites we know routinely add a premium for any CPA deal; from 15 percent to 50 percent of the calculated cost per action price, depending upon the size of the buy and the expected likelihood of successful return on the inventory committed to the buy. The rationale for the premium is that it is the cost of freeing the advertiser of the performance risk. Of course, the ability to charge assumes a site in demand by advertisers, and a knowledge base that allows the publisher and the sales people reliable, accessible data about past performance at a high degree of specificity.
The underlying message here is that you can’t set appropriate pricing without a real commitment to tracking and understanding all kinds of performance data about your site. Without continually updated data to rely upon and a deep understanding about how your audience behaves in a wide range of situations, you will never be able to take real control of your pricing or realize the maximum value on your site’s advertising potential.