April 18th 2011 | Andrew Edwards

Until evidence was presented to me this year at certain digital marketing events, I had long understood the crossroads where agency meets analytics to be not unlike the crossroads where guitar picker meets the devil: legendary, but doubtful and not wholly convincing in its result (except for rare genius).

Now the signs begin to show more significantly that at last — probably because of customer demands– agencies of every stripe are abandoning prior blindness and taking a dangerous peek at what web analytics is and is not, what its value really is, and how that helps the customer but not necessarily the agency (which has always been the catch).

Even last year we saw agencies doing their best to discount accurate measurement of campaigns by calling the whole process "difficult," "inaccurate," "not that meaningful" and–as a catch-all dismissal of the entire field of data collection about actual usership — "we use comScore."

This last conflation has always seemed especially self-serving and unhelpful to the customer.

Nothing is the matter with comScore, or course. They provide valuable benchmarking of a very few top-line measures based on data extrapolated from a sample audience. This is helpful for a "how am I doing compared to my competitor?" exercise, but only for the broadest, least insightful and least valuable measures such as total page views and total unique visitors.

The types of measurements comScore is typically cited for are eerily reminiscent of the types of measures that have long been inaccurate liabilities for ad venues like print, TV and radio. And therefore it is peculiar that we have ever had to listen to any "industry expert panel" try to shoehorn massively measurable digital media into the massively unmeasurable old brogans worn for decades by media often faulted for its inefficiency.

Have agencies been clinging almost with desperation to a self-professed inability to truly measure campaign success? Did it suit their agenda that one really never could know which half of one's advertising was a waste?

Of course it did! Efficient data in the hands of the customer would — by the math I would deploy against that old Wanamaker complaint — result in an immediate reduction in revenue and headcount at agencies by about 50%! Therefore, no surprise that accurate measurement has been adopted by agencies the same way a small village at the edge of a jungle adopts a baby tiger: in other words, not at all.

But we now see agencies begin to adopt at least a recognition that analytics can be helpful, even if they still seem to regard the matter as a theoretical exercise through which we all must pass before getting back to the fun creative stuff and message-spinning at which agencies so excel.

For instance, we are now hearing about "improper tagging" as if it were but a small pebble on the path to analytics success. We are also hearing about "proper campaign attribution" as if it were never possible until just now; and as if we could hardly know, without superhuman effort, which set of clicks actually resulted in a purchase or desired action; and which set of clicks led to a dead-end and waste of time and bandwidth. It is as if some folks haven't heard the news even from 2007: yes, that can be done as long as you know what you are doing.

I would love to see agencies take the lead on web analytics. After all, they remain in conference with their customers in a way that few other vendors have the privilege. If agencies can communicate the importance of deep, accurate, meaningful analytics, it would help those customers immensely (not to say help the analytics community which in fact exists only to serve its customer base).

Then, if those agencies would come clean about their own general lack of technical and professional analytics expertise, we would probably clear the air even further. We might get to a place where analytics technical experts could sit with analytics "analysts" and with "practitioners" inside the organization and come up with reasonable measurement plans based on specific business requirements; such that tags might be created and managed in such a way as to yield accurate, meaningful results for the customer; such that evidence of campaign success (or failure!) might be communicated back to the creatives so they can add value rather than winging it once again on the customer's dime.

All of the abovementioned possibilities have been available for years. For years they have been largely ignored by all but a few of the great message-makers that server major marketers globally. We know it's at least partly because of a "fear of the unknown": and because of a fear that accurate measurement might possibly point to a lack of success on the part of the agency itself.

Earlier I have written that for this last reason, it may make sense for customers to decommission any analytics work being done by an agency that also has creative input: because it's a fox-in-the-henhouse dynamic. But that is only a related subject and we have little room for that discussion right here.

For now, let us praise those mortals in advertising who attempt to rise above the quarterly concerns about billing and headcount, and who at least seem to be recognizing that their customers are being put at a serious disadvantage by the non-deployment of robust analytics — even if it is "difficult."

As for "not that important overall"? The person who says that, simply doesn't understand analytics—or its power to help companies make and save money as they spend enormously on measurable media. But it is heard often; and often it takes the form of "we use Dart already" (to measure campaign clickthroughs) or "that goes outside the domain" (and therefore, falsely, "cannot be measured"); most remarkably it has been said that "we asked a panel" which, while delivering its own value, is no substitute for machine-based data collection. I chalk these up as excuses for not digging deeper into the possibilities of measurement, and I count it as a detriment to whomever is footing the bill.

Remember why "the web" and now "mobile" and "social" have become such advertising darlings? It's because they are many times more measurable than print or TV or radio. And when customers start hearing from their agencies that you really can know more than hits and uniques and what comScore gets you, then we will need to celebrate a victory for fact-based decision-making such as we have never before seen.

Real analytics are only just being whispered of this year in the halls of advertising — even at those places that "have embraced Omniture" or who "do all our own analytics" or seem to be quite certain they have got the tagging right even when a cursory audit might reveal they have not — but it is a start. I won't say it's a "healthy" start, as it remains to be seen what direction this takes once an agency loses an account over an analytics report.

But it is getting harder and harder for content-makers to hide from measurement. And if you are paying for content creation, and would like to know which half — or which nine-tenths — is wasted, then this new openness can be nothing but good news for you and your wallet.