Here comes the next wave of disruption: hyper-targeting
This may be the most important thing I have to say this year:
The next wave of disruption for newspaper companies is bearing down on us right now, and it will be a doozy.
I’m not talking about mobile, which others have often cited as the next disruption. True, mobile is coming on fast, and it will soon become a new wave of disruption for both print newspapers and desktop/laptop-based audience and advertising plays. But mobile isn’t yet taking away a lot of revenue from print and established web advertising.
I’m talking about the hypertargeting of Web advertising through exchanges and machine-based buying.
This is not new; it’s been around for a while. But it has lately reached such levels of power and sophistication that it represents an immediate and fast-increasing threat to all traditional models of advertising. For newspaper companies and others using traditional sales forces to sell Web advertising, it’s a powerful double-edged sword — both a threat and an opportunity.
And, viewed from a local market perspective, it casts a light of uncertainty on the strategy of using local content on local websites to create audiences, and selling display advertising on those sites. That could spell big trouble for newspapers and TV stations, because they rely mainly on that model to generate most of their digital revenue.
This disruption results from the confluence of several parallel developments:
- The rise of exchanges using real-time bidding, through which advertisers and agencies can specify their desired price points and buy web display advertising on thousands of websites.
- The rise of ad platforms (often called “demand-side platforms” — DSPs — or “supply-side platforms” — SSPs — through which ad buys can be placed on multiple exchanges simultaneously.
- The development of massive databases containing huge amounts of data about individual users based on cookies and pixels, including what sites they visit, what pages they view, what actions they take and what search terms they use, either based on the IP addresses of their computers or on registration data that identifies them personally.
- The rise of ad-serving software that can dynamically determine which ad to serve to a given user on a given web page and serve it in milliseconds.
Connect these developments, and you get massive and instantaneous targeting and retargeting power, unlike anything in the history of advertising. Call it machine buying, call it real-time bidding or call it targeting — it’s huge.
No wonder Borrell Associates forecasts that targeted display advertising will go from 4 percent of all online advertising in 2008 to 54 percent by 2016, while run-of-site display plummets from 51 percent to 8 percent. And no wonder Forrester predicts that, by 2015, half of all online display advertising will be bought on web and mobile ad exchanges.
Thanks to my Morris colleague Michael Romaner, our EVP of Digital, for these figures. A big thanks to Michael, too, for the narrated overview of machine buying you’ll find here — an excellent introduction he has presented to several Morris audiences.
In practice, how does this stuff work?
Suppose an auto dealer in Augusta, GA, wants to reach people in the Augusta area who are actively shopping for Toyotas — or for SUVs, hybrids, Hondas, or cars in a certain price range. In the “old days,” the solution might have been to list the dealer’s inventory on auto sites and place banner ads on sites containing auto content, which might be read by the kind of people the dealer wanted to reach. Note that these tactics involve placing advertising on specific sites because of the type of content there.
In this new world, however, the dealer — or the agency or media sales organization assisting him — can place an order through an ad platform specifying what CPM he’s willing to pay, and checking off a long list of detailed user characteristics that can include geography, age range, family characteristics, probable income and more. And they can add retargeting to the buy, placing ads in front of people who have just searched for the desired type of car through a search engine or on a dealer’s website.
Note, however, that these choices are all about user characteristics and CPMs, not about what web sites to use. It barely matters what site or web page the ads appear on — huge national sites like YouTube, Yahoo, Facebook and AOL or the local newspaper or TV site — as long as the person looking at the page fits the desired target characteristics.
The advertiser truly is buying an audience, not a website.
For advertisers, this is a huge breakthrough. It’s the end of the old saw, “I know that half of my advertising dollars are wasted; I just don’t know which half.”
On the sales side
For local media companies, it should be immediately obvious: We need to start selling these powerful solutions in our markets as soon as we can. So far, these targeting methods are being used mainly by sophisticated national advertisers, but local businesses need them, too. And the power of these solutions is far greater than the ROS display advertising that has been our bread and butter in digital for so long.
To get going, we need to hook up with the right targeting technology partners, and we need to train and incent our sales teams to sell these solutions effectively. A number of newspaper companies are moving to do that. And if we don’t, somebody else will.
In that sense, hyper-targeting is a big opportunity. Our local sales forces can be our best competitive advantage, if we can mobilize them to sell this stuff well.
That’s an immediate imperative. But as we act on it, we need to be thinking about the larger implications.
On the audience side
In a world where advertisers can reach their desired audiences with pinpoint accuracy and instant immediacy, anywhere those people may be roaming the web, how will our current business model be affected?
Local media companies have been selling their web sites to advertisers pretty much the way they’ve always sold print advertising — as a local audience of premium value. They’ve been able to charge premium CPMs because of the perception that there’s more value in a local news audience than in other sites. Never mind that most newspapers’ web audiences are tiny compared to the local audiences of Facebook, Yahoo, YouTube and a dozen other nations players.
But as real-time buying of highly targeted audiences becomes common, why will advertisers pay $10 CPMs to reach newspaper site visitors when they can pay less to reach precisely targeted active buyers on other sites?
It seems highly probably that data-driven machine buying will drive down the CPMs newspaper companies can charge, because a motivated Toyota shopper is worth no more on their news site than on Youtube or Yahoo.
And if there’s going to be strong downward pressure on our CPMs, this calls into question our very model of audience creation. It costs a lot to create audiences by gathering, reporting and editing news. Even at $10 CPMs, this model of audience creation has been severely challenged. Drop the CPMs to $3 to $5, and it gets extremely tough.
Two conclusions seem obvious.
First, we will more and more find ourselves in direct competition with the huge local audiences of the national players. More than ever, we will need to multiply our own local audiences dramatically, to get as big a share as possible of machine-based, data-driven buys being made in our markets. Our sales staffs must learn to sell into everybody’s web sites through the DSP’s and exchanges, but we can earn both the seller share and the publisher share of the revenue on any part of the audience reached through our own sites. So our numbers need to get big enough to command reasonable shares of the machine buys. At Morris, we want to multiply our audiences by at least a factor of five.
Second, as we multiply our audiences, we have to do it at far less cost than traditional news coverage. So we need to be much more aggressive and creative in developing user-generated local content that creates high interest and high engagement among the people in our markets. At Morris, we’re challenging ourselves to figure out what aspects of living in a local community can generate enough passion and participation to drive much higher levels of engagement. News alone won’t cut it.
Boiling it down
We’ve got three big jobs ahead of us: Become potent sellers of hypertargeted local audiences wherever they may roam; multiply our own audiences so they can capture bigger shares of targeting’s exploding revenue, and figure out how to drive huge local audiences at much lower cost so we can make money at commodity rates.
This coming wave of disruption offers legacy media companies no special exemptions, but it does offer opportunities to those willing to reshape their sales and audience models. Let’s get after it.