The online advertising industry has to evolve, and the best way to fuel that evolution is by automating the entry-level day-to-day processes. Last time, I began a conversation on this topic with Jeff Einstein, author of "Put God First: A Pocket Guide to Quality of Life in the Great Age of Excess." Einstein posits that quality, the intangibles of advertising, is more important than efficiencies and scale. Today, the conclusion of our conversation.
Eric Picard: Much of the lower-level work will be automated and go away. New, next-generation roles will be more analytical and involve more technical ability than we typically see today (replacing adding machines with analysts). This will start happening within five years; the evolution of the industry will take longer, maybe 10. The average salesperson, for instance, will not bring junior media buyers out to expensive lunches. Sales roles will be more evangelical of the intangibles.
Note that while the average traditional media buying group deals with maybe 40 publishers a year, and the average digital media buying group deals with 100 to 200 a year, things will shift very quickly. Right now ad networks are dealing with thousands, even tens of thousands of publishers a year, very efficiently. But even the newest ad networks are using old-school technology. Nothing like what Google or Microsoft are working on. What happens when next-generation technology is applied against the problem?
Efficiency and scale are exactly what media buying are about, spending the dollars of the advertiser efficiently in a profitable, scalable way. Agencies are going to radically change their approach over the next few years, and efficiency and scalability are going to improve significantly. But efficiency doesn't always mean cheaper. It means reaching advertiser goals with more effective ROI (define). And I'm not talking about CPA (define) here, either. I'm talking about reaching advertiser goals on scales we haven't considered before. The issue is to solve that problem by giving agencies new tools, new models of doing business, empower them to evolve. And interestingly, some of the most senior people in the agency business are thinking this way.
Jeff Einstein: Media planning and buying has always been about efficiency and scale. And convenience. But let's be clear: media planning and buying is not advertising, and advertising performance doesn't decline because of inefficiency in the media channels. Advertising performance declines the very moment we decide that ROI is somehow the primary objective of advertising. Advertising performance declines the very moment we decide that who we talk to and where we talk to them are somehow more important than what we want to say and how we say it. Advertising performance declines the very moment we put the media cart before the creative horse.
EP: Of course creative is important. But in the world I envision, media agencies become more strategic, working more on the goals of advertising for the advertiser, mapping goals to business rules, and letting the business rules apply to automated optimization technologies that will very efficiently figure out how to spread dollars to impressions -- and continuously optimizing the spend throughout the campaign. Humans can't review the results fast enough. We're at the point where this can be done programmatically.
In the world I see coming, the manual work of media buying goes away. Much of that work today is literally rekeying data across multiple, nonintegrated workflow tools. Much of it is human review of data that is so complex that it frequently is a money-losing proposition. The work I'm talking about is mind-numbing. It's literally driving people out of our industry; the attrition is incredibly high among junior media and ad operations people at agencies (and publishers). And they're not leaving one agency to go to another. They're dropping out of the industry altogether. Nobody wants to do the work I'm talking about removing from the plate.
JE: My problem with your vision of an automated future predicated on efficiency is in how it all but eliminates the human dynamics, the intangibles that drive the real profit potential. Any vision that's predicated on what we can measure and analyze is by definition a narrow vision. Performance will not only continue to decline, but the quality of life that you profess to protect will decline with it.
It therefore seems to me that the very last thing advertising needs is a way to scale and more efficiently manage display. The very last thing we need is another reason to automate or otherwise eliminate another opportunity to negotiate -- because value is only assigned through negotiation. The fact that most online commercial inventory is never negotiated testifies to its real value: next to nothing and still losing value every day to the relentless tyranny and cynicism of digital scale and efficiency.
EP: If a buyer determines the appropriate cost/value success factors ahead of time, the process we use today to negotiate is replicated but automated. I'm talking about this happening for every impression, evaluated in real time and the instant negotiation happening between the buyer's optimization system and the publisher's optimization system.
In this world, much of the cheap remnant inventory will evaporate, at least that inventory that exists today. We'll always have remnant. But what I believe will happen in the next generation of our industry is that the super-premium inventory will become more expensive due to increased scarcity. The second tier of inventory will become more expensive due to targeting and automated optimization. And about half the remnant inventory will become more expensive, as we can apply very deep targeting and optimization to it. The remaining remnant will become even cheaper, but frankly most of it will go unsold as the audience that makes it up will be those who don't interact with advertising -- and who don't buy things from existing advertisers.
JE: Wrong, Eric. Both super-premium inventory and remnant inventory will continue to decline in value and performance on a per-unit basis relative to the total universe of inventory, especially as more and more of it goes unnegotiated and is rendered friction-free. The mere fact that the sheer amount of inventory will doubtless exceed even the demands of DR is a wakeup call and testimony to reduced value and performance in the media channels. Friction is a good thing, irrespective of scale.
EP: I believe there are ways to replicate the experience of friction in the sales channel in this massively scalable automated vision, but now we're treading on areas I don't want to discuss publicly.
I believe that we will see the world evolve to a point where spend and placement of media will be optimized automatically. And creative will be both interactive and customized based on the audience. Here's where I think the conversation nets out for me:
JE: The days of automated media buys and customized content delivery are already well upon us, although admittedly nascent. I don't see the need to redefine negotiation. People who insist on more value will negotiate. Those who don't won't. The automated systems you describe are all but inevitable but will not only strip a lot more value from each exchange, they'll leave a lot more money on the table. Meanwhile, increased scale and efficiency will continue to come at the expense of per-unit performance.
May 22, 2013
1:00pm ET / 10:00am PT
June 5, 2013
1:00pm ET / 10:00am PT