The problem with AI-assisted marketing
Three similar personal invites landed in my inbox within 12 hours.
Is it possible to reach peak personalization? Because I think I found it in my email inbox.
It’s not that the personalized emails I received from a handful of full-service restaurants were too personal… it’s that they were far too similar.
Here’s what happened: My birthday is coming up. I’m not particularly secretive about it and have grown accustomed, almost indifferent, to the string of marketing emails, birthday discounts, and well-wishes that arrive each June. Last Friday, exactly one month before my birthday, I received three emails from managers at restaurants in different cities. Each email addressed me by name, acknowledged my upcoming birthday and invited me in to celebrate. Nice, right?
It was nice! But a little strange since I’d dined at each of these restaurants just one time. None are located in my home city. One, I visited in New York about a year ago. Another, in Los Angeles, I visited at least three years ago. (It may have been four.) The third, a delightful spot in Nashville, I visited nine years ago. Each hoped I’d come back, they said.
I wish I could! But I’ll be more than a thousand miles away from all of these places on my birthday. (And most other days, too.)
As best I can tell, these emails arrived via a third-party loyalty platform that taps into restaurant customer data to optimize and personalize engagement. It’s powerful hospitality when done right, but a little unnerving when done en masse.
I must’ve ended up in a specific marketing segment inside the system, something like: Frequent diner with upcoming birthday who’s only been to this restaurant once. I assume the loyalty platform knows enough about me to understand that I am fiercely loyal to a handful of my favorite spots, and I do tend to linger, first over wine, then tea. I think I’m a good tipper and a fairly low-maintenance customer. (Sorry to that one restaurant where my children once broke two glasses by spilling water on the tablecloth twice, I definitely don’t expect a personalized return invite, oof.)
Instead of making me feel special, these emails made me feel like a commodity. Taken together, the invitations, arriving seemingly out of the blue and within hours of each other, diluted their message. Good tech and great data can make for excellent personalized service. But it’s hard to surprise and delight a diner when lots of restaurants are using the same information to do the same things.
Everyone wants to scale a personalized touch.
AI-fueled personalized marketing and engagement efforts might improve this. They might also make it worse.
According to National Restaurant Association survey data released earlier this year, 56 percent of full-service restaurant operators believe marketing technology has had a “significant effect” on their restaurant over the last couple of years. Things like personalization and increased segmentation and even automated messaging help engage diners and reduce the lift for operators who are already stretched thin. That tech is only getting smarter; in the same survey, operators indicated they used artificial intelligence tools for restaurant marketing more than any other function.
This was easy enough to see coming. Advances in generative AI have commoditized written communication, and smart marketing tools segment a dataset into targetable buckets of consumers. I once asked a restaurant tech company that offered suggested, AI-assisted messages to restaurants if they were concerned too much repetition might dilute the message.
“We have a template library of best-performing campaigns that operators can use as inspiration,” they said. “All of our customers […] are very good at bringing their distinct brand voice and imagery to their marketing efforts.”
Will it be enough? Maybe not.
I did write back to one of the restaurants. I appreciated the invitation, I said, but was surprised to receive it given how far away I live. I promised to come back and visit when I’m back in town, I said.
I never heard back.
