The Emergence of Audience Selling

Data-Driven Thinking“Data-Driven Thinking” is a column written by members of the media community and containing fresh ideas on the digital revolution in media.

Today’s column is written by Jonathan Mendez, founder of Yieldbot.

“So far audience targeted media optimization has been a huge party that only the demand side has shown up for. The supply side guys need to make data work for them.”

    – Terence Kawaja, GCA Savvian Advisors. IAB Networks & Exchanges Keynote, May 3, 2010

The idea of buying an audience is native to advertisers so it’s understandable that display technology has being trying hard to craft narratives around this vernacular from the earliest days of behavioral targeting. The thinking has always been more advertiser awareness of targeting brings more ad dollars into display. However, targeting is not selling an audience in the true sense. It is tracking people across the web and matching cookies.

This is why there remains a huge disconnect with advertisers. 74% of advertisers cited “targeting” as the primary reason they make site-specific buys. On the media side, on a dollar basis, direct buys will continue to dwarf audience buying through networks and exchanges. $7 billion of the $10 billion total display dollars in 2011 will be site buys (JEGI Estimates 2/10).

The other 30% of the media will have to manage against a myriad of cookies, third-party data, inventory, pricing, segmentation, creative and optimization. That is a fair amount of services and no DSP wants to become known (or valuated) as a services business. There are also a number of statistical inefficiencies and technology challenges managing across the entire web. As industry observer and investor Jerry Neumann recently commented “A DSP is going to need to be buying tens of millions of dollars of media to know anything at all about what works and what doesn’t.”

This presents publishers with a unique opportunity. In fact in Terrence Kawaja’s IAB Networks and Exchanges Keynote called it out as one of the largest opportunities in the space. Operating within a single first party domain or network publishers already have an audience. In fact, the huge opportunity is they have many, many different audiences they can sell, brand or target, to many, many different advertisers in many, many ways.

Even better, the performance and engagement of those audiences will be far superior to demand side audiences due to brand, timing & context. Simply, as all media becomes performance, publishers can thrive since they have better data, more ways to slice it and easier optimization.

How do I know? Look at Search. It operates from the same dynamic that publisher audience selling will.

In Search, marketers first create segments based on the actions of the audience (in this case keyword queries). Only then do they match ads and landing pages relevant to the segment created. This is a much more effective practice than Display where the marketer first creates an ad then they tries to find an audience across the web that might respond. The starting point is what is critical and therein lies an insurmountable edge for publishers segments and audience selling. Starting with the ad is a strategic legacy of old media. Performance here is akin to finding needles in haystacks. Starting with people’s interests or intent is the unique advantage of a user-controlled medium. It is akin to shooting monkeys in a barrel.

Publishers have two other critical performance factors in their favor mentioned above  (and also present in Search) — timing and context. With first party data event driven ad targeting is reality. In much the same way Search referrer actions can drive dynamic landing pages visitor events can drive ad matching. Best of all, these ads can be placed in context to the interests or intentions harvested from the visitor. As the advertising sage Howard Gossage famously said, “people don’t notice ads, they notice what interests them and sometimes it’s an ad.”

More intelligence has always made campaigns more effective. The issue has always been at what cost. With a higher CPM cost basis to begin with site-specific audience buying is well positioned to succeed, possibly even better than exchange buying where the costs of buying and optimizing data are increasingly becoming higher than the media itself.

Ultimately this may lead to the integration of demand side platforms with supply side platforms in an “end around” of Google and the myriad of middlemen in the Display ecosystem. This would provide higher returns, revenue and performance for advertisers and publishers, and more relevance for everyone on the web. In 2010 Publishers are only beginning to wake up to these ideas but I don’t expect they will sleep on their data much longer.

Follow Yieldbot (@yieldbot) and (@adexchanger) on Twitter.

Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!


  1. Great to see you again last week Jonathan, I hope you enjoyed SF.

    I have to take issue with your statement: “The other 30% of the media will have to myriad of cookies, third-party data, inventory, pricing, segmentation, creative and optimization. That is a fair amount of services and no DSP wants to become known (or valuated) as a services business.”

    Managing for those problems at scale and with a technology product is exactly what a DSP does. By it is only by productizing, automating and streamlining all those interconnecting pieces that we can do our job. Anyone who solves for those problems with services is doomed to failure.

  2. Great post Jonathan. Publishers need better tools to tap and leverage data that cannot be found in ad networks. Tapping knowable data such as minutes/month by section and topic for each visitor can provide insight and better target ads. Getting deeper into the data improves value for both advertisers and publishers.

  3. Jonathan–

    I like how you’ve turned the DSP logic in its head: advertisers know what audience they want to target and use data, etc. to find it, while publishers can know what audience they are reaching and should try to find products this audience will like. DSPs: know the product, find the audience; Yieldbot: know the audience, find the product.

    As you say, somewhere in this dialectic probably lies some great synthesis. But until then each approach will provide the larger margin to the party implementing it: to the advertiser with the DSP, to the publisher with your approach.

    Zach– I agree that the goal of the DSP is to automate, but the reality today is that they mainly do what they do through the accumulated learnings of their people, the very definition of a service. Even with the few DSPs that are actually platforms, the value of the service component of the buy (whether in the agency or the advertiser) vastly outweighs the value of the technology. This is changing but, imho, the media buy will never be a completely automated process.

    • Regarding the comment: I like how you’ve turned the DSP logic on its head: advertisers know what audience they want to target and use data, etc. to find it, while publishers can know what audience they are reaching and should try to find products this audience will like. DSPs: know the product, find the audience; Yieldbot: know the audience, find the product.

      This is interesting to me. What DSPs allow is to de-average the audience. “Media” people know who their audience is based on averages, but what DSP’s allow us to do is de-average the audience and specifically go after audiences with specific messages. Most advertisers know who their audiences “should be” but it when results come in, it’s typically a very different base.

      • I agree with you. When I said advertisers know their audience, I meant advertisers using agency buying platforms and DSPs, natch.

  4. Jerry, as a member of the DSP group of companies I respectfully disagree, please don’t define the product category by how some of the players that you are familiar with operate. I can’t speak to what our competitors do, but the whole point of a dsp is to use technology to automate the process and bring efficiency. That is what we do here at Triggit and it is the only way the makes sense for the long term. As far as I am concerned the service component is a small portion of the value vs the technology.

    Hopefully we can connect in person sometime and have this discussion at length. I really appreciate your thoughts and have been enjoying reading your blog.

    • Zach– Would love to chat. DM me. Loved your last post also (and not just because it mentions me :))… hope you keep posting.

      I was having this same discussion with an investor in the space who comes from Wall Street. His view was that there would eventually be an algorithm that decides. I disagree. And I thought it ironic that someone from Wall Street–where the commodities are much more cut-and-dry, the underlying value more readily ‘knowable’, and the information and data much more complete–would think that. Hedge-fund managers and Wall Street traders get paid huge amounts and, despite how cynically we might want to look at those salaries, these amounts must be tied to value created somehow.

      I don’t think the analogy is complete, but it deserves some thought: if the companies that buy and sell securities on Wall Street have most of their value created by people, why would we think that would be less true of companies that buy and sell media?

      This is not to say that the technology isn’t valuable. Bloomberg is a billionaire, after all. But in divvying up the ‘value’ created by Bloomberg vs. Goldman Sachs in any given dollar invested, I think you’d have to say it weighs in Goldman’s favor.

      • Jerry,
        I think your analogy would hold true if all of the participants in the display ecosystem had access to highly automated and interconnected trading system like you see on Wall Street. The problem is that the currently structure of our space is built on a manual trading infrastructure reliant on passing tags and IOs back and forth over fax and email. It is the same as if Goldman still traded using runners and hand written order slips. Where we are building technology to automate a highly inefficient process the technology is the value.