In several descriptions of the Information Trust Exchange, the word “signaling” is used. This post provides some context, and also introduces a new acronym — “CHEDDAR.”

Academic and marketers earn degrees and spend careers trying to understand what makes an advertisement effective — either directly selling a good or service, or building a brand or ideas for longer-term purposes.  And the content of the ad — print, audio, video, outdoor, digital — isn’t the only factor.  Other factors include:

  1. The type of product or service
  2. The nature of the audience members and their receptivity
  3. The timing or location of the message

All three come into play in comparing the form and delivery of digital advertising.  The third one is where the concept of “signaling” comes into play.  It refers to user-perceived cost of the advertisement as a hard-to-fake message about the advertiser’s confidence in its product’s ability to sell.  An ad on a major network or in The New York Times has “high signal” by that definition.  And ad on a unknown blog may not, at least by its context.

Data-driven, network-driven advertising placed in real-time on millions of web services that subscribe to third-party ad networks has the potential to end up in places the brand advertiser might not have expected, or sought.  Or it can “follow” an individual user from website to website. But such “followme” ads are turning out to perhaps be costing publishers rather than bringing dollars, according to Digital Content Next CEO Jason Kint, who analyzed new data from Microsoft.

Agencies have the option of placing advertisements specifically on publisher’s web sites or services — such as Google, or Facebook or major news or publisher branded sites.  This latter approach allows the advertiser to be specific about the “signal” sent by the context around their advertising message.

It’s a fair question to ask which is more effective:  Advertisements that follow a specific targeted user across all manner of websites, or advertisements that live in a particular environmental context on the site or service of a single or collaborating group of publishers?

Ogilvy & Mather UK vice-chairman Rory Sutherland argues that behavioral economics shows people make decisions based on the idea of signaling.  “We don’t just ask a question about what is it that  is being sold,” he said in a March 17, 2016 video interview. “We ask arguably the easier question which is: ‘What is the reputation of the person that is selling it’ ”   (FOR MORE FROM SUTHERLAND, SEE BELOW)

The Information Trust Exchange is pursuing an audience-segment advertising system that is intended to find a balance between the two approaches. In the ITA framework:

  • Advertisements can’t “follow” a particular user. That’s because advertisers will never have a way to address an individual user. Instead, they will have the opportunity to direct their ads to be shown to cohorts of users with similar demographics, interests or other attributes. This addresses a growing reason for users to install ad blockers — concern about privacy and “creepiness.”
  • Advertisements can be directed to specific publications and pages. Each publisher will prepare an Audience Profile Book (APBs)  for their members/users.  The books will aggregate user profiles in segments too large to identify any individual but small enough to be able to target those with similar interests. And the APBs will connect specific publications or sections of publications to particular user interests in a uniform and scalable way.
  • Advertisements can be targeted across broad geographies or interests. Because the ITE will allow APBs to be aggregated into larger and larger books, and segmented by unique interests or attributes, advertisers can select cohorts of potential customers to reach efficiently at scale.


A small but growing group of technologists and publishers believe that the kind of unrequested, “ads-that-follow-me” targeting of individual users is making it hard for reputable publishers to sell advertising directly within their services.  Even if you don’t care about the future of publishing, they perceive this as a looming public problem because of the way targeting is done now — through the use of dozens of tracking “cookie” placed on user’s devices.

The services that place these cookies use database to mix and match your activity to come up with consolidated — some say inaccurate — portrait of your appeal to various type of advertisers.  Their concern is that over time, this is eroding consumer trust and confidence in publishers and the web generally, leading to such behavior as ad blocking or just abandoning web services with annoying ads.

Two solutions are seen as looming:

  • Regulation by the U.S. Federal Trade Commission (or European authorities) to restrict outlaw tracking of user behavior without explicit permission.
  • Voluntary, market-driven changes in the way advertising technology works to end tracking of individuals who do not wish to be tracked.

The Information Trust Exchange is working with Project VRM at the Harvard-Berkman Institute, and others, to spread awareness of the CHEDDAR initiative, a basic set of technical choices that make web ads work in a signal-carrying way, and restore market power to news sites.

“It’s hard for publishers to enforce standards when an original content site is in direct competition with bottom-feeder and fraud sites that claim to reach the same audience,” writes Don Marti at Digital Content Next’s website in a blog entitled: “Service journalism and the web advertising problem.” 

“So how to turn web advertising from a race to the bottom into a sustainable revenue source, like print or TV ads?” asks Marti.  He continues: “How can the web work better for high-reputation brands that depend on costly signalingC.H.E.D.D.A.R is a basic set of technical choices that make web ads work in a signal-carrying way, and restore market power to news sites.



“WE don’t just a question about what it is that is being sold. We ask arguably the easier question which is what is the reputation of the person that is selling it? . . . economists would tend to think that brand preference is entirely irrational or it is driven by things like status symbol. I completely disagree. I think a preference for established brands is ecologically rational. I think it shows at an extraordinarily intelligent level in a sense saying ‘I can’t solve the main problem, it is simply too complicated or time consuming . . .  a preference for advertised goods is not irrational once you understand that costly signaling is a very, very common feature in nature. That we interpret the signficance of a message not only by the information it ostensibly contains but the cost in terms of effort and expense that’s gone into the creation of that message, and also by the cost consequent on the sender if that message is wrong.”

Sutherland goes on to argue that the advertising-technology business, which attempts to measure everything in terms of number of views and accuracy of individual-user targetting is misguided.

“This fetishation of measurement is dangerous . . . If you want to appear rational, you can take a complex problem, strip out most of the aspects of irrationality, so you are left with a frame of reference, so that it now becomes a simple optimization problem. You now solve that problem, reimburse it on the whole system as if you’ve necessarily improved the system. And it is a very, very dangerous approach. So you can always achieve a kind of rational consistency by simply narrowing your frame of reference.”