For example, here in Australia the Unique Browser audience is just under 90m – not bad for a population of just over 22m. That sure makes those tired old sample based audience data look far more accurate. There is a HUGE difference between trackable and measurable. Online sure does a good job of tracking, but a lousy job of measuring – TV is the inverse, and guess what the clients prefer. And as for that hackneyed old argument that reach planning, GRPs etc are old media … puuhhhlllease! It’s called communication planning and it is media agnostic. Get some decent metrics and come and join the party!

By the way, I sure hope in your equations you took into account (i) bots, spiders, crawlers (around one-fifth of all traffic) (ii) click-fraud (generally pegged around 20%-25%). Why pay for something you’re not getting!

On the “logic used in this example” side of things: If the onus is on the publisher to produce a higher click through rate, then the math here would change — otherwise the publisher would simply go out of business. While I don’t believe CPMs are set to only eek out a profit, serving nearly 3x the inventory at the same cost could surely upset profitability. In short – that onus would push up the CTR and drive down that 100,000,000 impression number (and do away with getting more impressions for the same money).

Also, time is not taken into consideration here. If this model calls for roughly 3x the clicks in the same amount of time, then you run the risk of unqualified traffic. A publisher will know where to get traffic regardless of quality.Traffic is easy. Profitable traffic is more complex.

In the end, we can all talk about brand awareness and customer engagement — but I know that these things do not cross my mind as I look at my own shareholder dividend statements. These items are still only a means to an end. The model on which you choose to pay needs to work for your program — it simply needs to be (wait for it….) profitable. I don’t like CPM, but if I test properly (against other ways of buying) and find that it drives the best profitability and volume — the pig’s dead and I got the bacon.

No matter what all the customer loving, earth saving, baby kissing, Gulf cleaning, warm and fuzzy press releases say: the king never left the building — profit is here to stay.

You bring up an excellent point – when evaluating an cost per engagement program, you have to be clear on what is classfied as an engagement and the eCPM for the campaign.

Savvy buyers also know the quality of CPC buys across ad nets – rarely do they back out at lower CPAs.

Also need to consider the eCPM to a publisher when trading on non CPM metrics…the lower eCPM the lower the quality and so the cyclical cycle begins of running across poor quality inventory.

CPM can and does work for response.

That said, I like the idea of cost per engagement. As long as the engagement is something other than a click.

First off, thanks for stopping by and commenting. From my experience when doing media buys, frequently websites are filled with excess inventory and you can negotiate a CPC or cost per acquisition (CPA). There are tons of websites that are unable to monetize their inventory profitably so they make their inventory available through aggregators that can get a higher price for their ad space – sometimes these aggregators will offer the inventory on a CPC or CPA basis. You can also see websites that make the inventory available through Google Adwords bid platform.

If you were not running a brand awareness campaign, and conversions mattered most then I would highly recommend first trying to negotiate on a CPA basis where you are guaranteed a positive ROI. If you can present a CPA offer where it is profitable for both parties involved, you should be able to make this happen even if it requires the vendor testing it a trial basis before rolling it out completely.

I hope this helps.