How AI, Media Mix Modeling & “Analyst-in-the-Middle” Design Will Revolutionize Media Planning & Buying
Since it hit mainstream consciousness in 2023, there has been no shortage of both predictions and potential applications of generative AI (GenAI) for the marketing industry. Most notable among them was Sam Altman’s prediction that 95% of all marketing jobs will be replaced by AI in March 2024.
Most applications seem to highlight GenAI’s potential impact on the creative aspects of marketing, like chat-assisted customer communication, content personalization and new creative content generation. In fact, Gartner® predicts that, "By 2025, 30% of outbound marketing messages from large organizations will be synthetically generated.”1 These content applications of GenAI are novel. But, on the media side – where the management of media buying has long left Mad Men-style advertising behind for automated toolsets like automated bidding, lookalike modeling and programmatic advertising – the focus on bringing creative up to automation parity misses the best applications of AI in marketing.
In the standard marketing process, critical tasks like budget allocation, strategic media vendor selection and audience creation are still heavily reliant on human intuition and interpretation. Yet, Gartner also predicts, "By 2025, organizations that use AI across the marketing function will shift 75% of their staff’s operations from production to more strategic activities.”2 To get there, it’s crucial that marketing teams integrate GenAI into the media planning process – yet few are doing so. We’ve noted two major trends occurring in unison that signal now is the time to invest in applying AI’s multitude of capabilities to the media side of marketing. The reaction we’ve seen to these trends, along with our experience applying AI in marketing departments, has shown us a new method for implementing a human-AI workflow experience – one that combines the opportunities that marketing mix models (MMMs) and AI offer with the uniquely human oversight that only people can provide.
The Rebirth of Marketing Mix Models
As we wrote previously, MMMs and prescriptive analytics have enjoyed a resurgence as organizations look for ways to understand the effectiveness of different marketing channels in a privacy-friendly way. MMMs deserve the hype, as they help forecast the possible impact of different marketing strategies on sales and customer behavior without mining into private customer data. AI has increased the popularity of these models by accelerating their insight cycle and solving economy-of-scale issues when matching the big, siloed data sets common in media planning. More recently adding to the already vast ecosystem of commercial-off-the-shelf (COTS) tools, powerful open source solutions have appeared that allow users to obtain quick prototypes (e.g., Meta's Robyn and Google’s Meridian), like the always-on MMM we proposed in our previous blog.
Brands Taking Back Control of Their Marketing
In addition to the rebirth of MMMs, marketing departments have also had to grapple with the ripple effects of another event: the pandemic. The pandemic reminded both consumers and brands about the benefit of having a direct relationship with each other. Brands moved quickly to consolidate their marketing efforts in-house, building dedicated teams and technology to interlock siloed business units into a singular marketing technology ecosystem. According to the Association of National Advertisers (ANA), 82% of brands have an in-house agency, up from 58% in 2013. The “desire to own, control and protect customer data” was the top reason for moving in-house. Brands who invested in themselves realized a newfound and nearly complete view of their customer base and the expertise and technology to directly act upon this insight. Travel and retail media brands realized additional revenue streams by monetizing their owned data and online properties to non-competitive enterprises. The agencies who supported them also had to change, finding new ways to provide specialized services to their brands through offering specialized data, exclusive media partnerships or more automated media buying processes.
All this newfound ability is not without its problems. Even those who invested in a full-stack marketing ecosystem to automate their marketing and media processes realize they’ve only solved half of the equation, such as when they find that their beautiful, custom marketing KPI dashboards sit unused by business teams who prefer to trust their manual, Excel-based methods. Paradoxically, this revolution in data availability and AI-supported big data analysis that was meant to improve our insightfulness is creating an expanding gap between the depth of information that machine-derived insights can provide and the level of information that the typical media planner can both trust and act upon with confidence.
Moving to an “Analyst-in-the-Middle” UX Design & Workflow Model
Before you hire more data scientists, AI engineers or ML operations specialists to educate your staff on the finer points of Bayesian analysis, we propose a more sustainable way to close this data literacy gap. All “algo” joking aside, it’s important to make a meaningful effort to educate your business divisions on the basics of data analysis – and make sure that that process equally focuses on educating your data analysis teams on the basics of your business.
Once that education process is underway, businesses can leverage their team’s overwhelming interest in AI applications to help stakeholders re-think how they can equip themselves with the best insights available to make business-impactful decisions. For example, how would you develop your interface if the human is not the primary user but instead another data source the AI needs to complete its calculations? What if you could deploy AI/ML for predictive analysis and to coordinate data into intelligent, intuitive and actionable marketing scenario plans? Could you do that while congruently developing a human interface with AI that only requests human intervention when necessary — like to provide external context or suggest decisions they should make in the real world?
In that future, a daily dose of campaign information is the “prompt” from the “user” to which the AI must respond. The AI then is asked to manage the reasoning: Is this good data? Do I have to react to it? Should I notify my human counterpart about it and ask them to do something about it? It’s root cause analysis (RCA) and next best action — but leveraging your AI engine.
The future of media planning is no longer Software as a Service (SaaS) or even Service as Software (SaS) – it’s Analyst in the Middle (AitM) design. An AitM design framework provides a comprehensive approach to media planning, mixing the accuracy of AI with human intuition and creativity. It appreciates the collaborators' role and their power to apply innovative solutions where AI cannot. While AI handles the hard work of sorting through massive amounts of data, humans play the important role of providing context and reacting while teaching the AI a decision tree to apply to its calculations in a complex market environment.
The Road Ahead: Future Proofing Your Media Planning
When the different disciplines of AI are appropriately applied and combined with an optimally designed interface, a marketing department’s media planning accuracy can be greatly improved. This will enable the creation of flexible strategies that can grow and adapt to market dynamics. With this approach, harnessing both the power of AI and the beauty of human intuition, you can be confident that your organization will be equipped with a competitive edge for your marketing initiatives.
Regardless of where your brand is in realizing the benefits of these opportunities, investing in your marketing team to create a proper blend of AI and human intuition will ensure a successful customer marketing strategy. By harnessing AI-driven analytics, privacy-friendly predictive modeling and strategic adaptability based on real-time market changes, you can not only optimize your marketing spend but also future-proof your media planning.
1 Gartner, Critical Capabilities for Content Marketing Platforms, Nicole Greene, Jeffrey L. Cohen, Rene Cizio, Carlos Guerrero, May 8, 2024.
2 Gartner, Critical Capabilities for Content Marketing Platforms, Nicole Greene, Jeffrey L. Cohen, Rene Cizio, Carlos Guerrero, May 8, 2024.
GARTNER is a registered trademark and service mark of Gartner, Inc. and/or its affiliates in the U.S. and internationally and is used herein with permission. All rights reserved.
