Friday, November 16, 2012

Programmatic Ad Networks: Threat to Editorial Models?

The New York Times posted a nice summary on programmatic ad buying, a rapidly emerging breed of online display ad auctioning that targets the people visiting a site more specifically than typical ad-serving systems. Briefly, programmatic ad networks learn about what you've been clicking on and reading about and tries to match ads to your expressed interests wherever you go on the Web.

So, to take the example on the right, I was investigating free-install solar panel systems online recently, and sure enough when I go to all sorts of ad-supported Web sites I see ads from a supplier of these systems whose Web site I visited. The advertiser bids for my ad space in real-time, so as the ads are being inserted into a Web page the system is choosing which bidder will get to put their ads up in programmatic ad spaces on their site.

The key factor here is that the editorial content of the Web site using programmatic ads generally has little to do with the ad that's displayed. For example, I was visiting general news Web sites when I saw these solar panel ads - not some other site related to this topic. In other words, since my interest in this type of product is fresh, the advertiser wants to keep it on the top of my mind as much a possible in the time immediately after my visit to their site, regardless of what I choose to focus on at a given moment. I think of it a bit like the movie "Minority Report," the science fiction film which portrays personally targeted ads of the future following us around on digital displays wherever we go.

Apparently programmatic ads work very well, and understandably so. While the context of editorial content certainly can provide very important content that can qualify an audience based on their immediate interests, the combination of focus and intent relating to purchasing things may not relate to that content's editorial profile with the precision that a programmatic ad system can provide. In other words, my profile as a reader at a Web site may contain any number demographic features and metadata based on my use of its content, but none of those may indicate where my focus and intent is regarding purchasing things in a given moment. Programmatic ad auctioning enables the key element of time to be added to an analysis of buyers' focus and intent, perhaps using editorial content as a general filter for excluding the value of a bid in certain circumstances but in general going with the data that they have in hand already to follow a reader into a Web site.

On the surface this may seem to explode centuries of efforts by newspapers and magazines to tailor editorial content to appeal to specific demographics in order to convince advertisers that they have appeal to those demographics on a consistent basis. But if revenues are no longer tied in lockstep to how editorial content can attract ad content, then it does open up the potential for different typed of editorial filters. Just as today there are tools like Inbound Writer that help publishers to tune their editorial output to key words and trends found on the Web in a given moment, perhaps programmatic ad networks will spawn a generation of tools that will help publishers to match a day's given editorial content to real-time ad demand. In other words, if it's between writing about fracking for natural gas or going green with solar, perhaps solar wins on the day that I and others were poking about for more information in places monitored by programmatic ad auctioning networks.

The NYT article highlights some of the usual whinging from journalists about this type of ad program, but I am not sure that it's wholly warranted. The old models of matching and selling ads to mass-market or niche market editorial content via display ad services has been in decline for a long time anyway, in part because of the very problem that programmatic ad networks are addressing. Publishers may know about what readers do when they show up at their Web sites, but they really don't know what we really do with our lives on a moment by moment basis. That's the purview of platforms provided by the likes of Google, Facebook and Apple, where the lion's share of online time is spent by the average Web-centric content consumer.

Google focuses especially on this type of insight; in essence they give away most of what they create to gather signal on who we are, what we do and what our focus and intent is at any given time; hence, Google Now, the new Android-based service that kicks out content onto phones and tablets automatically based on their analysis of what we might be interested in next. You might say that Google Now is trying to out-guess the predictive ad network technologies and push the content to us that's most likely to engage us before we ever go to look at it.

While editorial content is still very important, publishers rarely lead the online content industry in developing services that make use of content and experiences beyond their own editorial output. They're like storekeepers who "know" people who come in to their store based on the interactions that they have with them there, but have little idea what their lives are really like once they leave the store. Smart shopkeepers make it a point to learn those dimensions of their clients' lives, but with many major media sites not even bothering to integrate effective commenting systems, many don't really get the concept very strongly. By contrast, programmatic ad networks are like a shopkeeper who stalks someone who was just browsing in their store all over town. A bit creepy in real life, perhaps, but done right with the subtleties of Web communications, it can work well.

Clearly this type of ad placement technology is in its early days and the matching needs to be improved with enhanced signals and semantic processing of those signals. and editorial content can be a big part of that effort. Given the chase for search engine optimization that many publishers have been addressing in recent years, this is a challenge that may not seem unfamiliar to many of them. They could learn a lot more about their readers' behavior in general through whatever signal they can have their hands on. Unfortunately much of their needed signal is locked up in platforms like Android and Apple phones and tablets and Chrome and Internet Explorer browsers. So perhaps a good part of the answer is just writing the very best content that you can and get links into it on as many search engines and social media platforms as they can - in other words, accept that you don't really know who you're writing for all that well, but say what needs to be said. It wouldn't be the end of the publishing world if ad networks did all of the heavy lifting and writers did what they do best, after all.
Post a Comment