The State Of Programmatic Out Of Home (or The Problem With “Impression Multipliers”)

Out-of-Home advertising is undergoing a wave of digital expansion, but has yet to enter the programmatic arena. Ari Buchalter, CEO of Intersection, explores the state of programmatic Out-of-Home and how current offerings fall short of what programmatic marketers need

The Out-of-home (OOH) industry has a lot to celebrate. It’s the second-fastest growing advertising channel after digital, and the only traditional form of advertising that is gaining share. It offers advertisers massive national reach, but can also be tailored to have local relevance. It enables life-sized and larger-than-life creative experiences unlike any other channel. It reaches consumers in the real-world, influencing thoughts and decisions on their journeys between home, work, shopping and entertainment. It’s been shown to be not only effective at driving foot traffic, but more effective than any other traditional medium at driving consumers online as well. Studies also show OOH to be successful at all points of the funnel, from driving brand awareness to driving conversion, generating higher ROI than channels like digital display, radio and print. And it is more trusted by consumers than any form of online advertising. With all of that, one would think digital dollars would be flocking to OOH, but that hasn’t happened because one thing OOH has not been able to celebrate is successfully entering the programmatic arena.

**OOH companies have been anxiously pacing on the sidelines of programmatic for some time now, watching digital ad spend grow by double digits annually.** They understand that growth has been driven in part by programmatic technology, as 75-85% of online, mobile, and digital video advertising now transacts programmatically. With OOH accounting for a mere 4% of US ad spend, OOH media companies are eyeing that massive and growing pool of digital ad spend, and wanting to get in on the action. After all, the growth in OOH has been largely driven by the digitization of advertising displays, and digital displays should be able to attract some of those digital dollars, or so the logic goes.

As a result, there has been no shortage of press releases around the launch of programmatic OOH solutions. A veritable mini-Lumascape could be drawn up just around this would-be cottage industry, which by some accounts has been around for 4 or 5 years now. But if you talk to the actual buyers and sellers of OOH media, they will tell you that all of those solutions have had disappointingly little impact, and that “programmatic” OOH is yet to crack the 1% mark in terms of share of OOH revenue. As is sometimes the case when it comes to programmatic, the ratio of vapor to substance, of talk to walk, has gotten out of hand (hence the use of quotes referring to existing “programmatic” OOH solutions below).  

The lack of traction has been true both with respect to traditional buyers of OOH media, as well as with programmatic buyers of digital media. From the standpoint of traditional OOH players, one challenge is that many “programmatic” OOH solutions are taking a page from the old ad network model, bundling unsold inventory and marking it up – often on both the buy and sell sides – in opaque fashion. For the sell side, that means OOH media companies have little ability to distinguish their inventory, and a huge share of revenue ends up in the middleman’s pockets instead of theirs. That is why many OOH media companies have either pulled their inventory out of these platforms entirely, or limited it to their less desirable remnant inventory. For the buy side, the specialist agencies and buyers who have been buying this medium effectively for decades, already have their own processes and technologies (and are innovating new ones as we speak). For many of them, there simply isn’t enough added value from today’s “programmatic” OOH offerings to offset the additional workflow and lack of transparency around the inventory and economics, not to mention the prospect of getting less media for every working dollar. As a result, many traditional OOH buyers have eschewed these solutions. (Tip: if you do work with an existing “programmatic” OOH technology platform – either as a buyer or seller – ask them what percentage fee they take from the buy side and from the sell side, as well as exactly what OOH inventory is in their system, and ask for it in writing).

From the perspective of programmatic buyers, however, there is another set of challenges arising from the fact that every “programmatic” OOH solution thus far has approached the opportunity by simply trying to automate the traditional process for buying OOH media. To be sure, there are some important differences between OOH and online, mobile, or social media, a key one being the one-to-many nature of the medium: a single out-of-home ad will be seen by multiple individuals, not just the user of a single device.

That one-to-many aspect has given rise to a curious feature of “programmatic” OOH solutions: the impression multiplier. **Impressions are the currency of digital ads, and implicitly there is an assumption that an ad delivered on a laptop, desktop, smartphone, or tablet is seen by one person, and hence is one impression**. But OOH ads are seen by many people, just like TV and other broadcast media. And just as with TV, there are companies in the OOH space that use statistical methods to estimate the number of people that see a given ad. Those estimates, typically using data from the census, department of transportation statistics, surveys, and other data sources, are the tried-and-true currency of measurement in the OOH space. These approaches, well-grounded in traditional media channels, are the standard for measuring reach-oriented buys.

What existing “programmatic” OOH platforms have done is ask DSPs to accommodate accepting an impression multiplier field so that for every ad served you can multiply it by this number to come up with the number of “impressions” delivered. So if an ad is shown, say on a particular billboard, during a particular hour, the pertinent model might say 10 people saw it (and that would be the same answer at any point during the hour). An impression multiplier of “10” would, therefore, be applied to that ad play, resulting in 10 impressions. But are they really impressions to a programmatic buyer?

In the programmatic world, a true impression is atomic; it is the record of a single event. It has its own log-level data, which importantly associates an identifier to that event that can be used as a key to perform attribution i.e., to map each impression event to subsequent business outcome events like purchases, app downloads, site visits, etc. The impression multiplier does none of that. Imagine if in the programmatic world, impressions could not be tied to outcomes and could not power attribution; if they were just counts that could be tied to nothing at all. The traditional TV and print industries have already found that statistical or panel-based approaches to impression measurement don’t translate for machine-driven programmatic buying. Neither does the impression multiplier for OOH. All it tells you is that your ad reached 10 people. Or did it…?

**Trying to adapt statistical, reach-based measures of impressions to programmatic could have the unintended consequence of leading, in the worst case, to discussions around fraud**. OOH companies proudly proclaim that there are no bots walking around the real world and seeing ads (well, at least not yet). But let’s define fraud, in a programmatic context, as “paying for impressions not seen by humans.” The impression multiplier for a given ad might say that 10 people saw it, and that is how many “impressions” buyers pay for. Remember, that is a periodic estimate derived from statistical models, which has served as the standard in the OOH world for decades. But the programmatic world is different. Imagine if big walled-garden media companies (which are often the systems of record for many digital ad buyers), or telcos, or mobile location data providers, or computer vision companies, etc. were able to determine that there were actually 6 real people who were exposed to the ad. What would a programmatic buyer call the 4 impressions they paid for that could not be definitively be tied to actual people? Perhaps “paying for impressions not seen by humans?” Programmatic buyers have a word for that.

Now, on top of all of this, let’s say you’re a programmatic buyer and your DSP does commit their engineering team to the non-trivial effort of supporting impression multipliers and other unique requirements around these “programmatic” OOH integrations. All you end up with is a painful new silo. Since they have focused on merely automating the traditional process for buying OOH, existing “programmatic” OOH solutions require completely different workflows, different creatives, and different reporting. You have to set up and manage a multitude of direct deals (essentially one for each campaign buy against each OOH media provider), go back to a creative agency and spend money and time generating new creatives against different specs, and (if you want more than the impression multiplier sums), settle for emailed Excel reports of metrics like aggregated delivery by screen. All of that to activate a channel that remains detached from the rest of programmatic and without any ability to perform attribution and measure impact in the ways normally done for digital media.  

It’s not hard to understand why “programmatic” OOH hasn’t taken off. It’s clear the industry needs a different approach before OOH can enter the programmatic mainstream. As mentioned above, this state of affairs has arisen because solutions to this point have all tried to automate the existing process for buying OOH. What no one has done yet, is look at the true benefits of programmatic and understand how OOH can be re-conceived to deliver those.