Take risks in marketing to craft stories.

Could you please briefly talk about how you came into your current role?

Quint: I worked as a legal assistant in Washington, DC, after college. I liked the combination of language and logic that made up the law, and was considering law school. Unfortunately, I witnessed some of the downsides of the legal profession in my year-and-a-half in that role: my peers committing to go to law school for financial stability, not because of any passion to study it, biased and egotistical decision-making, and a Federal judge falling asleep repeatedly during a trial. So, I answered an ad in the Washington Post for the Assistant Counselor (Nuclear) position at the Australian Embassy, focused on science and nuclear policy. I learned a lot from being involved in such a controversial topic.

After some time, my now-wife and I moved to New York City, where I got a part-time job at Columbia University and could do a master's degree in strategic communication. I have stayed at Columbia ever since. Today, I am the Director of the Center of Global Brand Leadership at Columbia Business School. I am focused on thought leadership, bridging what happens in academia and daily marketing and brand development practices across industries.

Could you define brand development for us?

Quint: People tend to think of a brand as a combination of marketing, communications, and image. Brand development, however, is about attending to everything that affects the relationship an organization – or a person, place, or thing – has with its stakeholders. Technology, supply chain, and operations, for example, impact how people perceive your brand alongside marketing and communications.

How did the Brand Leadership field transform from when you started 15 years ago to now?

Quint: I moved into marketing and brand development at the industry's inflection point, when the Internet and ubiquitous connectivity became accessible to everyone. Reputation was no longer just about good media coverage. Consumers, businesses, and the media could all reach people with information and opinions about your products and services.

Plus, everything done in the digital space created data. Suddenly, decades of marketing strategy bound, in part, by the Wanamaker quote of not knowing which half of one’s marketing efforts was wasted, was no longer seen as a given. The causality between an ad and a decision to buy a product could be tracked. Segments could be more precisely targeted. Customer insight could be gathered from comments and ratings on lots of different platforms.

Plus, this wide visibility and two-way communication raised the understanding among CEOs and CFOs of the impact of brand perceptions. Enough impact that pressure was put on the marketing and communications field to develop better metrics and to run more test and control efforts to validate the effectiveness of a marketing effort.

This shift brought changes in the profiles of marketing leaders and a bit of tension that is now discussed as performance marketing vs brand marketing – where and when to lean on trackable, call-to-action marketing vs creative storytelling to drive top of mind thinking and loyalty. Today, go-to-market leaders understand underpinning business models, the financial basics of how their firm and their industry operate, and how what they're doing ties into the big picture of their company’s competitiveness.

I remember talking recently to the CMO of a multinational brewery. He confessed he wasn't clear on which parts of his marketing budget influenced customers' buying decisions. It was like throwing spaghetti at the wall and seeing what stuck.

Quint: And this brings up a favorite academic phase given what I just said above, and that is “it depends.” The beverage industry, and particularly alcohol, is on the complex side of things! Direct sales rarely exist. Local regulations influence advertising practices and sales channels. How do we determine the impact of advertisements against multiple factors, e.g., weather or socio-political trends or a brand’s own promotions efforts that are influencing consumer behavior?

Still, the future lies in continuing to be better at modeling proxies and running pilots to test marketing efforts in situations where you can control these other factors to some extent. Run different types of efforts in similar regional markets (similar populations, similar weather, etc.) and compare the effects.

The key is to try to gather as much information about the scenarios going on in society and in the immediate surroundings of your marketing efforts to be able to analyze what factors might be involved in the shifts in your product sales.

Listening to you makes me think about the importance of context while building AI models. But what about intuition in decision-making? In the wine and fragrances business, we discuss the role of a great 'nose.' Is there something similar in marketing?

Quint: AI has already and will continue to help analyze context and other factors in modeling the effect of marketing, communications, and brand. The improvements in AI over 15 years, for example, have created even greater capabilities to use social media to gauge public sentiment about a brand’s products and services.

Historically marketing was heavily influenced by intuition, or effective pattern recognition around what trends were affecting consumers and what storytelling style would resonate for a brand in that context. That hasn’t gone away, but now it’s possible to combine human intuition with AI analysis. I like to use Richard Dawkins and his evolutionary arguments that began with the "Selfish Gene” as a metaphor sometimes. It goes something like this: If an organism cannot see at all and suddenly its next generation can see light versus dark, it is incredibly far from the sight provided by an eye, but the impact of that improvement is still enormous in comparison to darkness. Little changes can create significant advantages and can then build up over time.

Something similar happens when browsing through the noise in a particular context, human intuition about human behavior and the emotions that drive decision-making can be very good. But AI can now bring an additional edge to analyzing scenarios and open up new discoveries to improve both marketing creativity and effectiveness.

But the 'nose'-factor will still apply, I believe. Experienced marketers and brand builders have all gone past the 10,000-hour rule and achieved a level of proficiency. And now most of them will add further capabilities to their decision-making with the use of AI. But, taking the NBA as an example, people like Michael Jordan and LeBron James still stand above the average professional player. What gives them the edge? The diversity of their experiences and some added natural gifts around how fast their neurons fire and how quick their muscles can react that let them perform above their experienced peers. We might call this edge a 'nose' or 'intuition' and such things take place in business professions as well.

Let's discuss marketing's focus on trust. Edelman publishes the Trust Barometer annually, and we see diminishing trust across industries and geographies.

Quint: There is an overall downward trend in public trust in institutions, which is the core area the Trust Barometer analyzes and that has certainly put pressure on brands. There is also another story within Edelman’s data. When it comes to leadership, for example, there is a mismatch between how people generally trust CEOs and what they think about their own leaders. The first score is 22 percent, while the second is 55! Direct interactions and deeper awareness of something or someone can break through the overall feeling of distrust.

And trust does clearly influence brand strength. My colleague Prof. Michel Tuan Pham, Research Director at the Center on Global Brand Leadership, did some research a decade ago using the well-respected Brand Asset Valuator (BAV) data, a long-running survey of consumer sentiment on thousands of global brands run within the WPP group.

Michel did a classification analysis of all the different characteristics people noted as something they associate with the individual brands (rugged, trusted, fun, innovative, etc.) and found that brands viewed as trusted had, by far, the highest lift in purchase intent and buying decisions. Other core brand sentiments like innovativeness, classiness, fun, energizing or caring had impacts, but at much lower levels. It is interesting to note that only about 50% of the brands were associated with any specific characteristics by the survey respondents.

While thinking of trust, it is also something specific to a core audience. Here in the US, most Republican-leaning people trust Fox News and distrust traditional media while Democrat-leaning folks deeply distrust Fox News and then may also be critical of the financial pressures that have changed traditional news outlets as well. You can build a strong brand even when some people hate you, as long as enough people trust you.

The internet and the growth of social media platforms have, as everyone knows, created propaganda at scale. While partisan divides have always existed, with engagement as the core KPI of online content, algorithmic pushes towards content that drives strong emotions has shifted socio-political discussions and trust.

Would AI further move things to the dark side?

Quint: It’s possible and everyone, including me, is worried about it, but we are still in the early days of its potential transformative effects. It is all about predictive content—what is the next word or pixel? What piece of code is coming next? And all of that is governed by the data it has sourced and the weighting and rules that the model developers have created.

What concerns me, and others, is how most of these genAI systems developed without any care to any IP protections as it sucked up content – nobody is sure of all the data that is in these systems, and the lawsuits are in play around whether it was true theft of protected content. I think of it as an ‘original sin’ of these systems, and when you build on top of such things, there will always be challenges to society and the economic effects they have.

In addition, so much data within these systems and other AI tools has underlying human biases, and thus decisions made by such systems will continue to perpetuate that bias. What I hope for is the use of AI to analyze human bias and bring it to light more and more in ways that can help people recognize and manage it. I don’t really trust that these hopes will come true, though.

In 2024, Google's Gemini emerged with the black founding fathers while trying to de-bias the foundational database. LLMs don't understand what you are asking. They entirely lack causation and reasoning.

Quint: Yes. This is a great example of how new these tools are, how important underlying developer rules impact them, and how hard it will be to combat bias in their outputs. The laudable goal was to take prompts asking to create images of people and not have them always be light skinned, which is what the model had been doing. The problem was that the system still wasn’t sophisticated enough in the way it fills in the blank(s) after a prompt to accommodate historical and contextual elements of the prompt. I’ve tried several image-generating systems and find them to have all sorts of flaws like this. Prompt them to create images of women of various shapes and sizes -- or with something out of the norm like a beard – and they cannot do it.

These challenges are what have driven the new “reasoning” additions to model development at several genAI companies, and such efforts have improved many of the outputs to conform better to requests and reduce some of the hallucination effects. But they are still far from perfect all the time. What this means for businesses is that as they use AI models, they need to be very careful about the data within them and they may need to spend a good bit of time to fine tune the model and add data to effectively use the systems in ways that will be true to their brand.

We are beginning an international academic collaboration, with our Center and colleagues at St. Gallen University at its core, to research and examine how businesses are developing responsible AI and AI governance rules. And we are not alone on such efforts in academia. Some best practices are developing around these issues, but some gaps are still out there, and it will probably not be a one size fits all scenario for organizations. And there will inevitably be unavoidable compromises. For example, data transparency is often a key element of responsible AI rules within organizations. But if they are using the main public LLM models, they have no transparency around the data within those systems.

Regulations as well will continue to come down the pipeline and how they are enforced as AI capabilities continue to grow is still in question.

What is the role of storytelling in the AI age?

Quint: In 2024, Coca-Cola did an AI-generated ad leading into the holiday season. It assembled iconic images of winter and Christmas themes mixed into Coca-Cola’s own red trucks, trusty dog, and polar bears, which they used in previous campaigns. In a sense, AI implemented a montage of historic Coca-Cola ad scenes. The marketing community hated the video, claiming it was both a poor ad that did not tell any story and was hurting the creativity community in the interest of lower costs. Efficiency over quality. But when testing companies showed the ad to the general public, they scored it favorably, likely because it had familiar Coca-Cola themes and fit well with holiday nostalgia.

Twenty-five years ago, before we all moved into the digital age, determining good and bad stories was very different. The 1979 "Have a Coke” ad with ‘Mean’ Joe Green comes to mind as a great example of the nostalgic feeling that ads used to mean something. So much is packed into those thirty seconds! Age and race differences, fame and feeling small, anxiety and tensions are all there. And there is a magical moment when drinking a Coke smooths any differences. As I noted earlier, many creatives feel that the world of data and tracking drove immediate ad performance to be heavily over-prioritized. A very difficult to collect but great empirical analysis would be to look at video ad campaigns over the past 50 years to determine trends in their storytelling level. I personally think people remember the good ones and forget that even before the internet, plenty of ads pushed function and promotions and weren’t telling great stories.

The big question is whether AI will be a research tool that brand leaders use to tell better stories, or whether it will be used in ways that dumb things down to the lowest common denominator to make sure everything has a general mass appeal. I believe, that in the content age we are in, the key is really about whether management leadership is willing to take risks in marketing to craft stories that stand out from “the crowd” or not. Truthfully, that has always been a key factor in whether great ads are made.

Authentic stories aren’t indeed limited to to those with deep pockets.

Another source of magic might lie in the ability to ask great questions.

What pieces of advice would you give to the aspiring CMOs?

Quint: The first one goes back to my earlier comment about understanding the business model, business goals, and financial metrics. Can you talk to a CFO and use the language of your industry? Over the past few years all the CMOs I have spoken with talk about making the CFO your best friend.

The second would be around how well you understand your internal stakeholders. Do you have access to product development or supply chain management teams or their data? Do you have the capabilities to share insights and work in partnership with them? If you can, you have a much greater chance of “walking the talk” when you craft stories and make brand promises.

The third one is about understanding data. You don't need to be a data scientist, but you must know how to ask the right questions while working with data scientists and people building machine learning models. And coupled with that making sure to craft experimentation into your marketing campaigns. Test ideas, get back data, and then learn from it and refine your decisions before scaling up any effort.

The fourth is making sure to be a lifetime learner, and hunting for insights and practices outside your category. For example there is a growing push around foresight thinking -- what changes are going on in the ecosystem of your industry, among your consumers, or in the world at large, that may affect your company’s business model, and then how can you think strategically about the likelihood of these actions and provide an appropriate budget towards activities that can help prepare for them.

In his 2007 book “Big Think Strategy,” for example, our Faculty Director at the Center, Prof. Bernd Schmitt, has an exercise called ‘killing your sacred cows,’ which challenges an organization to think about its “standard business norms” for its category and then break them entirely to think innovatively about how they could do something different. These last points are reminders that disruption is a regular part of the world right now. You need to understand your routines and rules, and then consider how they might be broken, or how you can break them to remain competitive.