My POV vs. the WSJ: What does the Omnicom and IPG merger foreshadow for the ad industry and its employees?

Since this article was published in the WSJ on Friday, I’ve had no fewer than a dozen conversations with friends both in and outside the industry, where they’ve asked me what my take is on it and if I agree with the points it makes. The premise of the article is essentially that the $30B mega-merger between Omnicom and IPG would theoretically create $750M in efficiencies, which could be re-invested in data, tech and, in particular, AI – and help them fend off the tech giants while ushering in a transition from “Mad Men” to “Math Men and Women”. See below for my take on some of the specific points the article hits on.

“The hot hires instead crunch numbers, run experiments and use algorithms to analyze data and predict what kind of message will resonate, where it should appear and even when someone is most likely to click ‘buy.’”

My take:

Yes, the agencies can hire more “number crunchers,” but where will they put them and how will they get paid for their time? The agencies currently get paid for creative services, production, media planning and buying, and very little for advanced technical services like analytics. Even what the agencies call “analytics” is often low-level reporting. The agencies will have to enter or create new service lines that help clients, for example, create their own AI models to optimize their creative and target their ads. This is not a construct that really exists at any scale within these agencies the way it does within consulting firms like Accenture or pure play data science service providers like Fractal and Mu Sigma.

The challenge for holding companies like Omnicom and IPG is that there is very little historical evidence that they can either organically create or acquire and integrate pure play technology-based services companies at scale that don’t fall neatly into traditional creative or media focused service lines. As Terry Kawaja rightfully pointed out in his recent article in adexchanger, Publicis might be the exception with Sapient creating some hope that this can indeed be done by a holding company.

“The Interpublic deal, which includes data broker Acxiom, will give the company access to more data. Acxiom has ‘a much better understanding of the consumer, given the number of consumer attributes that they bring to the table.’”

My take:

To be brutally honest, I don’t think any of the holding companies (and, candidly, I think this is even true of Publicis with Epsilon) have really done much to evolve their business or create any significant synergies with the acquisition of data companies like Acxiom. Plus, I’m not sure Omnicom has really been disadvantaged on a relative basis by not having one thus far. For the most part, the deep data and technical expertise in companies like Acxiom, Merkle and Epsilon has remained pretty tightly contained in their original walls and has not dramatically increased the data IQ of the rest of their agency brothers and sisters. I know many of my friends at these companies are going to be annoyed that I said this … but it's true.

Thus, I’m deeply skeptical that Acxiom by itself will have any transformational impact on the combined entity since it wasn’t to IPG. That is, unless the company more radically reimagines itself as a provider of advanced data and AI services and solutions, in which case having an asset like Acxiom would in fact be incredibly valuable. Omnicom has a seed for this kind of work in its Annalect unit, but we are talking about something at a totally different level of scale and impact here. And if it’s about doing the same old things –like planning and buying media – slightly smarter because now you have a data company, history has proven this will fall flat.

“But in the long run, the efficiency of AI is likely to reduce the number of people and hours required to deliver for clients—a big impact for agencies that have long made money by billing hourly for employee work.”

My take:

While this will take some time to unfold, I think this is inevitable across almost every industry. For better or worse, the need for the armies of people employed by agencies to proliferate creative versions and manage ad ops, media, and reporting will be reduced by AI. This feels really obvious, and I’m actually surprised at how surprising this is for a lot of people I’ve encountered. Even some of the relatively small number of agency employees writing code will be replaced by AI. However, the highly skilled technical and strategic resources that help brands with a set of invaluable data-centric tasks will be in higher and higher demand as adoption of AI-based technologies increases. We are not talking about media analytics (a.k.a. reporting and insights), but rather services like:

• Data Engineering

• Data Science (predictive analytics)

• Customer Insights and Strategy

• Customer Journey Orchestration

• Media Planning (more modern like an extension of Orchestration)

• Programmatic Activation (not necessarily the button pushing, but the thinking and chess playing much like on Wall Street)

• Creative Optimization (same as above)

• Measurement and Campaign Optimization (same as above)

• A lot of other advanced things I can’t think of on the fly here

This is a melding of more traditional agency skill sets with hardcore technical and data science capabilities that work within cloud data environments like AWS and GCP, data sharing frameworks and clean rooms like Snowflake and Habu, and the various data management toolkits like Databricks where the AI models can be accessed and trained. These native data and data science skill sets married with the world of professionals proficient in adtech and martech to create a uniquely “modern agency” set of competencies and culture.

So, to summarize, I absolutely agree that the march from “Mad Men” to “Math Men and Women” is upon us, but I don’t think a mega merger between two holding companies by itself really has much to do with it. I think companies like Publicis are winning because they already got the message quite some time ago, and my bet is that they are probably thinking a few steps ahead. I also think that some of the acquisitions of data companies by holding companies show that simply taking out your wallet without a lot of strategic and executional focus doesn’t get you very far. What I am more interested in is whether there will be new entrants or new and surprising combinations of non-traditional players that might enter to compete for the attention and dollars of the big brands. My guess is that because AI is so disruptive and requires such a different skill set (and mindset) that the answer will inevitably be – yes.