The goal of this post, and our follow-up later this week, is to help you understand the AdSense revenue model so you can diagnose and treat revenue fluctuations like an experienced MD.
Study up
The first step is knowing how the figures reported in your account (such as eCPM, CTR, and page impressions) interact to describe your total revenue. Think of each number as a variable in the revenue formula for your site. At the highest level, you can calculate revenue by multiplying your page impressions by the effective cost-per-thousand impressions (eCPM) and dividing by 1000.
Revenue = Page Impressions * eCPM / 1000The eCPM metric provides an estimate of how much revenue you can expect to earn for every 1000 page impressions. For example, if you serve 10,000 page impressions and earn $40, your eCPM is $4. If page impressions increase to 30,000, you can predict that you'll earn $120 given the $4 eCPM.
eCPM = Revenue / Page Impressions * 1000
Most AdSense ads pay on a cost-per-click (CPC) basis, so eCPM is really a measure of your average ad performance. Breaking eCPM into the click-through-rate (CTR) and the average cost that advertisers pay per click (CPC) gives you a more accurate measure of performance.
Revenue = Page Impressions * CTR * average CPCOnce you know your average CTR and your average CPC, you can predict how much revenue you'll earn for a given amount of page views. You can also analyze your revenue by looking at placement-targeted ads versus contextually-targeted ads.
Total Revenue = Revenue (contextual) + Revenue (placement-targeted)While contextually targeted ads always pay per click, advertisers can pay for placement-targeted ads by impression (CPM) or by click (CPC). To account for both of these bid types, you should look at the average eCPM for placement-targeted ads. More simply, you can just add placement-targeted revenue to your contextually targeted revenue.
Revenue = (Page Impressions (contextual) * CTR * average CPC) + (Page Impressions (placement-targeted) * eCPM (placement targeted) / 1000)Even though we're looking at contextual and placement-targeted revenue separately, don't forget that these two types of ads compete against each other in the auction. We'll always show the best performing ad, regardless of targeting type, so more competition creates higher winning bids.
Revenue = (Page Impressions (contextual) * CTR * average CPC ) + Revenue (placement-targeted)
Identify the symptoms
Now you're ready to diagnose any revenue fluctuation. Just like the revenue formulas above, let's start simple and gradually get more complex.
The first question to ask is: Did either your page impressions or your eCPM change? You can compare trends in both page impressions and eCPM using the Advanced Reports in your account.
If your AdSense page impressions have declined, you should determine if traffic to your entire site is declining as well. A web analytics tool such as Google Analytics can provide you with this information. In addition, you should check your pages for unpaid public service ads (PSAs).
If your eCPM is down, you'll need to dig one level deeper and find out if your contextual or placement targeted ad performance has dropped. You can also find this data in the Advanced Reports tab using the options shown below.
Let's consider your contextual ads first. The two key metrics to investigate are CTR and average CPC. CTR is given in your reports, but you'll need to calculate your average CPC using your favorite spreadsheet. (My favorite goes without saying). Please keep in mind that this is still an average CPC for your account and doesn't necessarily correspond with the price paid by any specific advertiser. Once you've narrowed the change to CTR or average CPC you're ready to start treatment.
For placement-targeted ads, you should analyze how much total placement-targeted revenue you are receiving and the average eCPM. Changes in either of these metrics usually indicate that advertisers are beginning or ending campaigns targeted to your site. Again, placement-targeted campaigns are more likely to be short-term than contextual campaigns.
That's all we have time for today -- now that you have a better understanding of what factors can affect revenue, don't forget to check back later this week for the second part of this series. We'll be discussing ways to treat revenue fluctuations based on the symptoms you've discovered.
Updated to fix typos