Part 2: Leveraging AI to unearth and distill potent audience definition
Also, Cannes and creativity and new music
📂 File this one under “Proud Parent.” 🎶 🎸 If you’re into Bossa nova-infused guitar pop, check out “Cherry,” the latest single from my son, Ellington, who is 16. He wrote the song, sang the vocals and played all the instruments. Older son Maks (age 17) engineered and produced.
AI + Creativity News: The All Cannes Digest
And we did kind of laugh at Publicis back in 2017. Mostly because the comms around its Marcel platform were overhyped. I’m still not certain what it actually enables. Good of them to ask, “Is it ok to talk about AI at Cannes now?” Because apparently that’s all anyone is discussing. The Cannes Lions Advisory State of Creativity Study 2023 notes, “65% of respondents cited AI as the most important tech trend in 2023;” perhaps because, “Last year, you told us that mundane tasks are eating into your opportunities to get creative.”
Digiday has a solid primer on the culture of the festival—and a great podcast, with the first episode featuring Sir Martin Sorrell, founder and chairman of S4 Capital, and Tara Agen, head of marketing for HP. Ms. Agen offered what I took as the money quote regarding AI, creativity and mundane tasks:
“Humans really will be a huge part of this [AI] process, we’re just trying to find and define the workflow that takes the [mundane?] work out—so that we can actually have them do the work we want to pay them for, which is their creative mindset.”
Can we start with time sheets?
Shortlisted winners in the advertising competition get posted throughout this week. Always worth the time to see what gets held up as exemplary work.
Plain old Creativity
1️⃣ Last week I joined Neol, a community-powered creative platform. IDEO co-chair Tim Brown describes the intent here. I’ve had the chance to engage in similar venture over the years, but this one seems different in all the right ways. More to come, for sure.
2️⃣ Neil Perkin, who writes a wonderful blog, hosted an insightful interview with Dom Boyd, UK Insights and Marketing Effectiveness lead at Kantar MD for Google Firestarters. The episode’s focus—the marketing confidence gap—is really the age old hesitation around investment in creative potency. On the one hand, “we’ve never had more knowledge around effectiveness,” and yet campaigns struggle to gain investment enthusiasm. However, as a creative person, it’s wonderful to hear an effectiveness expert state:
“Creativity really is the biggest lever you can pull on driving not just your sales but your profitability and commercial impact.”
Now if we could just get everyone to agree with Dom and move on.
Part 2: Leveraging generative AI to unearth and distill potent audience definition
It’s time to continue sharpening our thesis via AI-generative tools. Last week we talked through the benefits of AI-generative prompts to help refine business issues (Part 1 is here). Now let’s focus on the audience.
AUDIENCE DEFINITION
Someone has the power to impact the [business] issue by changing behavior.
Illuminate the personality, role or tribe who can best impact the issue, and specific behavior we’re trying to change.
Who’s it for?
Whose behavior stands between us and the change we seek to make?
If a creative brief could only answer one question, I’d always pick “Who’s it for?”
That answer—provided it’s well-researched, keenly written, and provocative—is the stuff which drives dramatic change. Time spent honing the response to “who’s it for?” is the best allocation of resources a marketing team can make. It’s the surest route to persuasive creative.
But you’re overwhelmed with competing projects. And there’s soccer practice and summer camp to chauffeur. And you forgot to eat lunch. Thank goodness then, for LLMs and AI-generative technologies. This particular phase of marketing is literally custom made for tools like ChatGPT and Bing.
One of my favorite movies is David Mamet’s Heist. The dialogue is angular, cutting, pungent. And there’s a moment relevant to ChatGPT and prompt-writing, I swear. In it, Gene Hackman’s character, Joe Moore, a con artist and a boat builder, explains how he accomplished an engineering feat:
BOAT BUYER “You’re a pretty smart fella.”
JOE MOORE “Ah, not that smart.”
BOAT BUYER “You’re not that smart, how’d you figure it out?”
JOE MOORE “I imagined a someone smarter than myself. Then I tried to think, ‘What would he do?’”
Here’s the thing.
We can leverage AI-generative tools to imagine and play a role or personality in near endless contexts, circumstances, and conditions who is “smarter than” ourselves. And then ask it, “What would [X] do?”
The trick is your imagination.
The audience is the problem.
More to the point, their behavior is the problem we seek to solve, address, affect, inform, nudge. Our job at this juncture is to write about behavior.
If we work among professionals, we likely have a decent starting point. In other words, if we were able to distill the business issue, then “who’s it for?” ought to be somewhat clear, and is typically demographic: Age, gender, job or role, living circumstances, etc. A starting point, but not inspiring. We’re missing culture, attitude, belief systems, history, hopes, fears—all the salient nuance of a problem worth solving.
This is where AI-generative prompting can help paint a psychological picture, infer inner motivations, and suggest inherent drama. And help us write an audience definition which elicits robust agreement among stakeholders (i.e. “you’ve described them perfectly”) and attracts greater effort from our creative partners (i.e. “these people sound fascinating”).
Mix, match and contrast as you see fit:
PROMPT 1 to expand the understanding of a particular audience: “You’re an audience research expert. I want to understand what motivates my target audience: [paste demographic info here]. Use insights from firms including Nielsen, Kantar, GfK, Gartner, and Ipsos to illuminate the target audience’s behavior as it relates to [business issue].”
PROMPT 2: “We’re conducting audience research. Let’s examine how firms including Nielsen, Kantar, GfK, Gartner, and Ipsos have defined the behaviors which motivate our target audience: [paste demographic info here].”
PROMPT 3: “You’re a market research expert. Our target audience is [paste demographic info here]. Please discuss the target audience’s top five films, songs and TV shows of the past decade and what motivated those choices. Use markup.”
PROMPT 4: “I want you to act as a market researcher who specializes in social media usage. Our target audience is [paste demographic info here]. Please discuss the target audience’s top ten social influencers, and hashtags related to [business issue]. Use markup.”
PROMPT 5: “We’re conducting market research and our target audience is [paste demographic info here]. How does the news media describe the ambitions, fears, eccentricities, and habits of this audience over the past decade? Please cite relevant quotes and sources. Use markup.”
PROMPT 6: “I want you to act as a journalist writing a biographical story. Our subject is [paste demographic info here]. Use newspaper and magazine articles from the past decade to define the motivations, habits and preferences of the subject as it relates to [business issue]. Use markup.”
PROMPT 7: “I want you to act as a theater playwright writing dialogue about our subject: [paste demographic info here]. Use non-fiction sources to discuss how you would research and describe the daily routines, frustrations, ambitions and emotions of our subject.”
I guess this one was inevitable… and you can view ChatGPT’s results here which are surprisingly useful!
PROMPT 8: “You’re David Mamet, writing a new play about [paste demographic info here] and are involved in the [business] industry. Discuss how you would 1) research, 2) evaluate and 3) ultimately write about their motivations related to this industry.”
Julian Cole in his Strategy Finishing School suggests another approach you could explore.
Prompt 8: “We’re conducting market research about [paste demographic info here] audience. What are 10 commonly held beliefs our audience has about the [business] category? Also provide the polar opposite of those beliefs. Use markup.”
Of course you’ll want to double-check any claims, facts, quotes or sources the AI generates.
The point of using AI-generative tools here is speed, range and diversity. We’re trying to characterize a problem rooted in the behavior of an audience. You can cover a lot of ground using the prompts above, and adjusting variables as you go to explore and express that character.
And if you’re still wrestling with this concept, ask the Seth Godin AI bot.
PROMPT: “I want to know how to describe a specific audience in order to inspire creativity within my team.”
We’re going to focus on actionable insights next week.