It seems so simple - you charge an advertiser x amount per # of impressions, bung a javascript ad tag they give you into your website et voila, you are making money serving an ad. However, there are many nuances to measuring the success of how that ad is targetted, and if you have control over variables such as exclusivity or impression spreads through a time period, things can get tricky.
Given how accesible ad serving is today, many people are avoiding Google's Adsense placements and instead looking for their own bespoke implementations of an ad server. One we work with exclusively @ Design Guru is called OpenX - a great piece of Open Source Software that gives industry leading functionality to any of our clients who need to manage ad campaigns, rather than, say just placing one or two dedicated spots on their site(s).
I got thinking about how confusing the notion of ad serving is today after bouncing some concept notes over to a client and thought it may be useful to repost them here. This got penned in response to the question: "How come I can't see a particular ad in a spot live on my site even though I've reloaded the same page a few times?"
Well, there is a fine balance between how many ads share a space, the weighting of those ads individually, the periods those ads share the same space and so on. This all adds up to the fact that a banner advertisement may need many more reloads than you'd assume without working the numbers.
For example:
If there are 10 ads sharing the same space that yeilds 100,000 impressions per month, you can assume that each ad gets shown 10,000 times per month. However, if they are weighted differently such that some are forced to display more often and thus reach higher than 10,000 say, impression targets (or even less than 10,000 - as is the case with our current Lacoste campaigns), the other ads in the mix will be affected accordingly.
Now, if one of those 10 ads has been set to display with a low weight and effectively captures 2,000 impressions/mo, one would have to reload the same page to see the ad 50 times! [100,000/5,000]
This is where using the stats as accurate guidelines makes sense; the probability of seeing a single ad is isn't as high as you'd expect.
For anyone running their own ad servers who are trying to understand this stuff, I recommend looking through the stats for past campaigns you've run with a calculator - for revising those already running, and better placing future ads, think about how many impressions you want them to achieve per month and whether those should be capped (ie. once reached, the ad auto turns off) or spread through the month by appropriate weighting, or exclusive (as when you want a campaign to gain maximum geo-targetting potential, for example.)
Ad serving can get complex and it helps to have a clear grasp of what goals a campaign should have when it is initiated so we can setup each one appropriately.
Check out some features of OpenX in the video below - if you'd like to place ads on your website, drop us a line - we can help make sense of all this :)
