Using AI Intelligently, or How to Avoid the “Zombie Marketer” Apocalypse

April 15, 2019

All marketers are storytellers. You’ve heard this a million times because it’s true. As a marketer, you identify key audiences, get to know them, then craft a narrative that helps each of those audiences picture themselves using your product or service and being happier as a result.

It’s no different when the thing you’re marketing is actually a story. It may seem like the marketers on the front lines of Hollywood and the publishing industry have it easy — after all, how many potential ways are there to sell people on a new superhero movie? But the truth is they’re dealing with a lot of the same market pressures and limitations as the rest of the marketing industry, and their job is only getting tougher.

The Plot Thickens

Entertainment is one of the most oversaturated markets in existence right now, with mass amounts of brand-new content competing for audiences’ time and money every day. And with digital marketing ushering in a new era of targeting capabilities, marketers are able to take a much more surgical approach to reaching the people who will ultimately be interested in their content.

The new challenge is getting those audience profiles as accurate and detailed as possible. Anyone who subscribes to Netflix has probably found themselves laughing at some of the categories the streaming service identifies as “interesting” to them. Maybe you were surprised to find out you liked “strong female leads in stories of revenge” or films about “wine and beverage appreciation,” but however the service categorized your interests, they went way beyond just recommending something because you were a “black female” or a “male under 35.”

Most of us would agree we’d rather be targeted based on our actual personalities and preferences than broad stereotypes. And this is how entertainment marketers want to approach their jobs across the board.

But it’s challenging. Because unlike Netflix, marketers do not have built-in algorithms. They don’t see every movie or TV show that comes out each day, and they certainly don’t read every book. They’re also likely to have biased views on the content they do consume, making it hard to “categorize” things properly. (Is “The Handmaid’s Tale” appealing to people with strong feminist ideals or fans of dystopian dramas? The answer is probably both, but if a marketer only identifies with one of those groups, they may ignore or downplay the other category.)

So if a marketer hasn’t seen “Beginners” (or seen it through the eyes of a diverse audience), they may not realize that it shares thematic and cinematic similarities with the project they’re currently promoting. And therefore, they may totally miss out on some valuable audience profiling data.

How AI Can Help

You may see where this is going already. To help entertainment marketers be more effective, we can equip them with the data and algorithms they need to make informed, surgical marketing decisions.

My company StoryFit has a database of millions of film scripts, teleplays, and novels, with new ones being added every day. Our AI breaks down those scripts to make them easier to categorize across detailed metrics like the personality profiles of the characters, thematic elements, the type of story arc (rags to riches, etc.), and even things like whether they pass the Bechdel Test.

Marketers can find comps to their current project, see exactly where the similarities exist, and use that information to target audiences who responded well to the comparable content. They can also look back at marketing campaigns for those properties and see, for example, how they were structured tonally and what aspects of the story they focused on to make more informed decisions about their own campaigns. Comps can even be helpful when determining — and defending — marketing budgets.

Can StoryFit’s database help in the development process as well? Of course, but it’s a tool that has to be wielded wisely. Producers who find themselves reverse engineering scripts from scratch based on what has worked in the past are likely to find the results disappointing. The human brain is well-tuned to identify unoriginality, and playing it too safe will come back to bite studios and publishers over time.

Ideally, if producers or writers are using the tool, they’ll come to it with a list of questions or benchmarks. Are my characters registering as complex, and are they sufficiently different from each other? Do female characters have at least 50% of the dialog? Do they have characteristics that are usually reserved for male characters (are they trailblazers)? Which famous character is my leading male most similar to, and am I happy with that comparison, or should I rethink a few things to make him fresher? In other words, it’s a tool that can help creatives see their stories objectively and decide whether they like what they’re seeing.

As AI starts to become more of a fixture in entertainment marketing and marketing as a whole, it’s important that we use it to augment and improve our decisions, not make them for us. AI is one valuable voice in the room. It should not be the sole decision maker or sole opinion we listen to, any more than we’d listen to one team member at the exclusion of all others.

Machine learning should help marketers across industries become superheroes, not zombies. So as you start to craft your marketing stories, look to AI where possible for data that makes you smarter. Then let your humanity take the wheel, with clearer knowledge of the best routes and the lay of the landscape ahead.