Motive: Why programmatic will kill digital agencies
by Nevo Hadas (@) I am going to start at the end, so those who already know how this turns out don’t have to be bored, hoping that I will have an inspiring twist…
In between 8–10 years’ time, the greed which drove digital media companies to implement programmatic buying ends up killing much of the digital media industry. Massive companies will disappear and the power will reside in enterprise marketing tech, proprietary data clouds and walled garden media such as Google, Facebook, Snapchat etc…
In 1996, we had no idea what would happen with the internet; we made money by writing HTML code in Notepad and making blinking text. 2016 seems to be driving us the same way… There is an ever-increasing push by agencies to move everything to programmatic, reminiscent of the zeal of an eight-year old holding a spoon in one hand and a tub of Nutella in the other.
Inevitably, this will end very, very badly…
“Where there is mystery, there is margin”
A couple of decades ago, agencies lost control of the briefing process: the magic of turning the customers’ business problem into what the agency should do. Agencies demanded better briefs because they made their margin on media commission and, if they could spend less time figuring out what their clients needed, they could be more profitable. Marketing departments agreed — and presto, a whole lot of mystery disappeared. Once clients understood the mystery, it became a process; once it became a process, it became manageable; and, before you knew it, agencies became outsourcing firms selling hours as a business model and being treated like an interchangeable life form.
I cannot remember who coined the term “where there is mystery, there is margin” but it has been proven as true as Murphy’s Law. Luckily, advertising has the attention span of a Mad Men episode and the same thing seems to have started happening again.
It is a better business model… but for whom and for how long?
Agencies love programmatic for a whole lot of reasons, simplistically put as:
- It’s cost-efficient ie doesn’t need that much effort to run (no I/Os, no media sales people, no proposals to read through, no complex media-planning using arcane planning methods such as GRPs, fewer lies to sift through, no bargaining, etc) so you can teach a 20-year-old with a bit of experience how to do it fairly easily
- It’s measurable ie it does reports for you and can integrate into your clients’ analytics, making the buying far more-targeted with a better attribution model of what worked and what didn’t
- It’s often real-time ie you can optimise campaigns far more efficiently
- There are hidden margins ie clients haven’t figured out where all the bodies are hidden so far and, while it has gotten far cleaner, there are still a lot of unscrupulous agencies out there milking margins from their clients
- Clients don’t understand it and think it’s very, very complicated
However, most importantly, it’s not a people-based business model but a highly efficient product-based one. It is predictable and very profitable.
When you look at the value chain for a normal publisher, it looks fairly bleak, with more money going to the middlemen than the actual publisher (this excludes walled gardens such as Google, Facebook, Snapchat, WeChat, etc). This means that, over time, there will be a significant change in publisher capabilities; they will improve their data sets (making it easier to buy specific audiences) and focus more upon what makes money (worthy of a whole other article).
Currently, media owners woo the buyers, which creates market efficiency ie the buyers have lots of clients and they know lots of media companies. It is valuable to the clients, as they wouldn’t have the time nor the expertise nor the contacts to meet so many media owners. It’s valuable to media owners as it’s a cheaper way to sell.
As buyers become more programmatic, smart publishers will go direct to large clients because there won’t be anyone at the agency able to integrate custom deals into the ‘system’. They will undertake their own content development and become, effectively, an agency. It’s already starting to happen with Vice and many others…
Agencies are losing a key skill, which was valuable, as they plug in ad networks and focus upon how to make things easier for themselves.
A client with enough technology to make programmatic work well doesn’t really need an agency in the middle
At a high level, programmatic promises to allow clients to target the best potential customers at the lowest possible cost at the right time. To do this, customers need to implement a marketing tech stack with DMPs, and lead nurturing systems to build up data on whom these customers could be. Rather than simply targeting women between the ages of 24 and 60 who live in the US, a marketer with sufficient tech on their side may pinpoint who buys, when they buy and a better attribution model of the purchasing process.
In fact, a client with marketing tech may build and expand segments in a way that is superior to how an agency at a buying desk can. With the shift in technology spend rapidly moving from the CIO to the CMO, and huge enterprise players coming into the ‘marketing cloud’ business, it’s clear that there will be massive investments in this space.
The more technology you have, the tighter the integration between your media spend and your effectiveness. More-advanced marketing tech now allows me to publish directly from my DMP/campaign orchestration straight into a DSP (or buying platform) to undertake purchases, which then dynamically links straight back into analytics or DMP and gives effectiveness reports based upon the segments targeted and results reached. It even tells you where you have underused spend.
The beginning of the end
While agencies and networks were always the guardians of advertiser spend, this will shift and, over a series of years, supplier agreements will come to an end and publishers will connect directly to exchanges owned by marketing tech companies, cutting out middlemen and increasing their margins.
With a bit of machine learning, media-buying optimisation stops needing people
The way this rollercoaster ends is in an R&D and tech-pricing war, which agencies will have a tough time winning, purely because of capitalisation ie Google has more and cheaper money than WPP (larger market cap and better multiples). A high-level estimate is around US$13bn in marketing tech acquisitions since 2014; it’s probably been a lot more investment with VCs, etc.
Agencies, which have worked so hard to create the programmatic market, will be left out in the cold as marketing tech will communicate directly with exchanges and use a learning algorithm to optimise the media spend. With all the data available, these models will become highly efficient and far more effective than a human in the middle.
Ultimately, the future value chain won’t have a digital media buyer in it at all…
Digital agencies will go back to making production-based revenues, just like in 1996.
Republished and edited with permission.
“Motive” is a by-invitation-only column on MarkLives.com. Contributors are picked by the editors but generally don’t form part of our regular columnist lineup, unless the topic is off-column.