• by Justin Evans , September 5, 2014

The television industry right now is very lucky. It is being challenged by a host of technology and business innovations shorthanded with the term “programmatic.” Unlike other business revolutions, there is a near-perfect precedent for programmatic TV: digital. (This is no accident; digital ad concepts — and people — are making this new TV revolution happen.) I have made this comparison in past columns; today I will focus on programmatic media measurement.

In 2009, when I worked in the Nielsen company’s online group, one of the issues we faced was what to do with online ratings. The company’s online offering at that time was a rough facsimile of TV audience measurement. We were helping Web publishers sell ads by providing a detailed count, and description, of their audiences. As in TV measurement, we were measuring the number of people who showed up to a publisher website (just like with a TV show) and their characteristics (such as their age and sex). The problem was, the industry cared less and less, and the data was valued accordingly. What was going on?

It turns out there was this whole thing happening in digital called “programmatic” (which we weren’t calling programmatic at the time), where the campaigns were targeted to certain audience types and splintered, by ad networks, across hundreds of websites. Anything meaningful about the campaigns needed to be technically stitched together not through site tagging, but ad tagging, and then condensed into a meaningful story via custom analytics.

At the time, we had an appalling acronym for these custom analytics (“CAT” or “TAP” or something — you had to have the A for “audience” in there) which later became the more elegant “OCR”—but the thinking was sound. You could use technology to gather data on even highly fragmented campaigns. And the data could be sliced a number of different ways, and connected to other databases, to yield a handful of meaningful metrics:

• Audience analysis: whom you reached, how many times.
• Brand lift: how much you influenced recall, opinion, purchase intent.
• Offline sales: how much you influenced in-store sales.
• Online conversion: how much you influenced online visits, leads or sales.

In 2014, as our excitement gears up about the promise of programmatic and addressable advertising for the massive TV ad market, it’s important to remember the lessons of online.

Programmatic is an abstraction — and this is YOUR problem. Audience-targeted media by definition strips away the value that advertisers associate with content. When you’re targeting based on the presence of a target audience, you no longer have the luxury of turning on “American Idol” or the NFL and seeing your ad run and feeling that swell of pride. You are trusting a series of tech vendors, some of whom you’ve never heard of, to run your ad on Spike in Pensacola Florida at 1 a.m., and ESPN in Terre Haute at 9 a.m. — and so forth, hundreds of times per campaign. It’s abstract. You can’t put your hands on it. In the absence of the “NFL Effect,” you must rely on dry, but still substantial, effectiveness measurement, to prove the worth of programmatic. The CMO will want proof that shoving her long-labored-over creative down the programmatic chute yielded something more valuable than a stack of affidavits.

Scale and standardization are critical. Online missed something in the heady programmatic era, and that was the opportunity for clients to believe, really believe, in effectiveness measurement, by, for instance, a) measuring and recording their own norms and trends, and b) managing their mix and their plans based on these metrics. I’m not saying no clients did so, but many didn’t (and don’t). The cause is threefold: first, the industry trained itself to rely on media vendors to supply traditional effectiveness metrics. So even if the results of each individual study were objective, the media partners were biased. So any metrics the vendor supplied were dismissed as mere marketing, and not something to invest in. (I know there are plenty of exceptions here, but I am confident that this was the rule.)

Second, diverse media partners means diverse research vendors, and thus no consistency. Metrics without consistency have value, but only in a limited way. Third, the most scalable and consistent metrics available were the “attribution” metrics of digital, which, as many have pointed out, are often weak and misleading in methodology — and are not portable to non-digital media.

Measurement creates markets. The value that media measurement provides goes right to the heart of economics. Measurement creates transparency — either about delivery, or effectiveness — and transparency creates trust. Trust creates an efficient marketplace. An efficient marketplace is what buyers and sellers crave — especially when they are entering new territory. New territory is exactly what online was 20 or even 10 years ago, as more and more budgets moved online and the medium’s ad spend caught up with the time consumers were spending on it. For programmatic to grow up and earn a vital, sizable place in the media mix, trust, transparency, and efficient markets will all be crucial. Help buyers understand what they are buying — and they will buy more.