Last week AdExchanger put out a research report on “The State of Programmatic Selling, 2015,” which MediaPost covered
here. What interested me in particular were the pain points that have emerged in
the programmatic space, and the extent to which measurement can remedy this pain (or, conversely, the extent to which absence of measurement may be a source of pain).
I wrote in this space last month about the nature of currency audience measurement data in the 21st
century. But in the programmatic space, the question of currency grows even more complex. Certainly we understand that whether you’re talking about a major agency holding company making its
annual upfront commitment to a broadcast network, or an advertiser buying a few million impressions in a programmatic environment, there needs to be some way for buyer and seller to value the
impressions transacted.
The complexity in digital arises from the fact that essentially these impressions are being bought and sold one at a time. So a digital version of a program rating (or
even a commercial rating) is largely useless, especially in a programmatic context. Program and commercial ratings are born of a broadcast construct, where commercials are placed in programs and
beamed out into the ether. In digital, the impressions are served one at a time, and you may not even be able to tie these impressions back to specific content environments.
According to
AdExchanger, inventory quality was a pain point for 27% of agencies and 25% of marketers, but only 10% of publishers. For publishers, the biggest problem was vendor complexity. (The second publisher
pain point, a lack of understanding about programmatic technologies, might reasonably be seen as another flavor of complexity).
The disparity between buyer and seller perceptions of inventory
quality, coupled with publisher frustrations over vendor complexity, suggest that we have yet to solve the “currency” challenge in the programmatic space — and that we need to.
To me, conversations about inventory quality lead directly to questions about valuation: What is an impression worth, and how it may be valued. From valuation, it’s a short hop to currency,
which in a media-buying context refers to the (typically third-party) data that buyers and sellers agree to use in inventory valuation. Historically, we’ve called such currency data “the
ratings.”
So, then, what should “programmatic ratings” look like?
Ideally, in a programmatic environment, buyers and sellers should have equal access to
impression-specific or placement-specific third-party data that shines a light on the potential value of that impression or placement. The data must be available in real time. And it should be
provided in as granular and transparent a fashion as possible.
At the impression level, audience size is irrelevant: the audience size is one (issues of co-viewership notwithstanding). So what
data helps in the valuation of the impression? Certainly, targetability is essential. What are the demographics of the person exposed to that impression? What are the relevant behavioral drivers (i.e.
auto-intender, etc.)? Then there are metrics about the placement itself: How likely is that impression to be viewable? How likely is that impression to be fraudulent? What can we know about the
context of the placement: the content of the page, with respect to IAB or other page content taxonomies?
When buyers and sellers are able to unite around credible third-party currency data
sources that help shine a light on impression-specific value, demand for that inventory increases — and both sides win. When TV currency moved from household ratings to demographic ratings in
the ‘60s, both buyers and sellers profited. I have no doubt that the pain points associated with programmatic transactions — and the disparity between programmatic share of impressions
versus share of spend — will heal as we coalesce around currency “programmatic ratings."