Not too long ago, a certain retailer had a single store front in New York City. The CMO would stand out front and count the number of people who looked in the windows, and the number of those that came inside. Later he would tally the sales and revenue.

He noticed that when he placed some print advertisements these metrics would increase by greater margins than when he placed others.

This is an early case of using macro measurement to understand the effects of media on total sales. Notice, he wasn't surveying the customers as they walked through the door - "Excuse me Miss, can you tell me why you came in today?" - and he wasn't collecting coupons at the point of sale.

Instead, he relied on a simple macro method that is MORE reliable and MORE accurate even than trying to discretely track the motivations behind each purchase.

And what's even better is it's these methods that have built some of the largest most respected brands today. That single shop in the story? Coach.

Isolate a geography, measure marginal changes on a macro scale: Wouldn't it be great if we more brands incorporated this tried and true method when attributing online media?

I have to say that I loved this article discussing - and dissing - online marketing measurement. It echoes my sentiments (and frequent statements) perfectly by stating that "Yes, we [in the online marketing space] have more data than ever before, but we're measuring whatever we can rather than what we should."

The fact is that this is not an exact science - at least not in the way 'science' is being defined here (as based on numbers alone) and we should stop pretending that it is. For me this comes down to attribution - how discrete attribution models largely based on cookie tracking can mislead strategy and work counter to overall business goals.

They speak more in detail with regard to creative and social - how sentiment and engagement cannot be measured in the way we think. I say they can be measured but must be done so through a framework for testing and measuring marginal performance on a MACRO level. There is a science to Impression-to-Profit because it is repeatable and predictable.

For more about this, see my posts here and here.

But great work, Jim Sterne, I couldn't have said it better myself.

In a recent release by a marketing analytics company, buried towards the end, was a statement I thought hit the nail right on the head. Paraphrased, "The ability to quantify the impact of media is absolutely critical. Not being able to measure [it] diminishes [its'] importance."

Well, duh. But I think we in the online sphere just get so caught up in the idea that we need to track every single part and piece in order to call it measured. And the trouble is, we can't do it. Not entirely.

This means that discrete tracking alone will always undervalue the impact on the earlier phase media-- like owned and earned media, or other Awareness media. Because it's more susceptible to being missed by tracking. No matter how sophisticated the solution, the multiple device, the multiple user and the cookie deletion issues will still be there.

So if discrete tracking doesn't- can't - tell us the whole story, then isn't that media still being de-valued? And isn't that exactly what we're trying to avoid?