What Marketers Get Wrong About Personalization

Don't believe the hype. Personalization isn't the Holy Grail of marketing.

Not every marketing message needs to be — or should be — personalized.

There are some compelling and specific use cases for personalization, such as when someone uploads a product to the shopping cart and doesn't follow through with a purchase. The Bayard Institute reports that the average cart abandonment rate is a staggering 70 percent. However, with the right messaging delivered at the right time, we’ve seen a 9 percent conversion rate uplift for a retailer’s abandonment baskets.

But the goal of personalizing every message for every person at every moment is a fool's errand.

How I react to a product recommendation sent as a text message or while perusing a retailer's website is all about my unique situation and particular context. There's no deep learning engine, no artificial intelligence model in the world that can predict a person's frame of mind at a specific moment. This is where personalization gets it wrong. Humans are unpredictable beasts; you've lost them if you hit them at a bad or wrong moment.

Also, personalization will never enable that cultural zeitgeist possible with an advertising campaign that hits all the right notes, like Nike's iconic "Just Do It" campaign or Budweiser's infamous 2000 Super Bowl "Wassup" commercial.

A good example of when to use personalization is when you have a very predictable context. For instance, if I log on to a sporting goods retailer's website, put a hockey stick into my basket, and then leave, then it makes sense to deploy a personalized message that says, "Hey, your hockey stick is still in your basket; how are you going to score any goals if you don't have this latest hockey stick?"

Personalization is effective here because it's guaranteed to be contextually relevant. There's a linguistic connection between the product and the message: "How will you score goals if you don't have this new hockey stick?" And it's delivered at the perfect moment — when someone is at that moment of purchase. That last little nudge to make them purchase something is when personalization can be the most effective.

McKinsey's research discovered that companies that excel at personalization generate 40 percent more revenue than only "average" practitioners. Therefore, personalization alone isn't enough; how it's implemented is equally important. The report cited that 59 percent of customers wanted timely communications tied to critical moments. They want to see that brands are invested in the relationship, not just the transaction.

The key is to create contextually relevant language at scale that performs strongly. That's the challenge. You cannot click a button and pray it works.

Don't personalize unless you must; when you must, go deep. When personalization isn't hitting the mark, it's often because the message is created as an afterthought and isn't contextually relevant.

Personalization is based primarily on past behaviors and assumptions based on the demographic or the segment we fit into. But not everyone will align perfectly with their demographic or segment. Relying solely on personalization commodifies the entire purchasing process; it devolves everybody into the next-best action.

Another problem with personalization is that it doesn't allow your customers the opportunity to be surprised and delighted. When's the last time you went shopping on a major e-commerce website and, although you were shopping for one thing, you saw something else, and you're like, "Oh, wow, that's interesting; I'd never considered buying something like that before." So, you purchase something you wouldn't have expected or wasn't part of your original intention.

Instead, we always get the most logical next-best action while shopping online. Personalization assumes that humans are logical and rational, but we aren't. Humans often make unpredictable purchases. If you try to personalize everything perfectly, you're giving up those serendipitous moments that can lead to discovering new things.

I've yet to see a company do it, but what would happen if a marketer picked 500 products the company would like to sell, then randomly selected three to put in each person's email? So, each person gets a different combination of three random products. I'd love to see what the numbers look like.

People need to be surprised and delighted; let's help them get there.

It's easy to recommend a product. But it takes a lot of work to recommend a product compellingly. And when you can marry the generation and optimization of creative with personalization routines and underline it, that's extremely powerful. However, if you personalize without optimizing the creative simultaneously, you're wasting your time and money.

Parry Malm is the CEO of Phrasee, a platform that uses generative AI to generate billions of the best marketing messages across the digital customer journey.