The Death of the Customer Persona

Longtime readers of this blog will know that I believe passive investing works better than active investing. Active investors spend too much effort researching their investments and picking stocks. But the vast majority of them fail to beat a simple passive investing strategy.

The meta-lesson here is that putting more effort into something doesn’t necessarily translate into better results. I believe something similar is happening in the field of Marketing. There are many aspects of Marketing that used to take a lot of effort, but are becoming less critical in the age of AI.

One such area is the idea of target personas.

The Old Way Of Consumer Research

Remember that time we all thought we needed to know our customers down to the last detail?

When design thinking was all the rage 15 years ago, we used to focus a lot on creating our ideal customer persona. The idea was to create personas with real characteristics, preferences, and hobbies. That way, you could better tailor your product, messaging and media buying to appeal to them.

In my first job at an airline company, we’d spend hours in workshops brainstorming our target personas. We even gave them names like Manager Margaret, a 45-year old business traveller who skips her inflight meals to work. And then there was Demanding Donny, the high-net worth retiree who only travels First Class and takes Krug with his caviar.

Once we had those personas, we’d try to figure out how to reach them on media. We’d buy banner ads on LinkedIn and CNN to reach Manager Margaret. And we’d buy the front page of the Straits Times to reach folks like Demanding Donny, because that’s what he reads while munching on his caviar.

The problem with this approach is that it leaves out a whole universe of potential customers who don’t fit neatly into these segments. Life isn’t a sitcom. People are complex and they don’t fit into our perfectly crafted personas.

There are 15-year olds who only travel First Class. There are CEOs who only buy flights when they’re on sale. And there are Gen Zs who love travelling but hate social media. All of these unique folks didn’t fit into our customer personas, so we left them out, leaving money on the table.

AI Eats Personas For Breakfast

Instead of spending hours creating fictional characters, AI is busy analyzing real data. It’s looking at everything from what you bought last week to the websites you visited yesterday. And it’s using that data to figure out who is most likely to buy your stuff.

Instead of manually segmenting your customer into an arbitrary persona, you’re better off asking AI to find ANY customer who’s likely to buy. In other words, Marketing is moving from a philosophy of “Sell my product to a 45-year old single mother of two who likes to play pickleball on Wednesdays” to “Sell my product to whoever’s interested in buying“

The second approach might give you less control, but it drives better results. Think about it: Would you rather have a robot analyse millions of data points to find your ideal customer, or would you rather spend your time daydreaming about Manager Margaret’s favourite vacation spot?

But this is really hard to accept: Many Marketing teams will claim that “we use AI to power our Marketing“, but very very few will admit that “we’re willing to step back and let AI do its thing“. It requires a tremendous shift in how we think and work.

Replace Your Personas With Data

So, instead of manually segmenting your audiences, what can you do? Maybe try giving more CONTEXT to your AI.

You know how ChatGPT gives you better results when you give it better prompts? For example, if you want ChatGPT to draft you an email, don’t just say “Draft an email saying XYZ.” Instead, you should say, “Act as a B2B software seller to SME clients. Draft an email saying XYZ, and keep the tone friendly but professional“. The second approach gives you better results because it gives the AI more context.

In the same way, uploading your first-party data is a powerful way to help AI perform better for you. This means:

1. Collecting the details of customers who’ve bought from you before, or who have expressed interest in your brand (e.g. by signing up for an email list)
2. Encrypting it so that everything stays privacy-safe
3. Uploading it to your ad tech platforms

In a world where everyone is using AI, this is a key way to differentiate your performance. It allows you to tell AI what’s worked for you in the past, figure out a pattern from it, and apply it in the future.

I once heard someone say that, “In the not-so-distant future, all of media buying will entail uploading a list of your customers, and telling AI to find more of them“. It’s a bold statement, but I think there’s some truth to it.

Remember: AI isn’t here to replace us. It’s here to make our lives easier and our marketing more effective. So, maybe it’s time to experiment with it and see where it takes us.