Personalization is the key ingredient to helping build trust with a customer. With trust comes increased growth and engagement. But personalization can also go too far. Companies can risk personalizing so much that users might then have a narrow understanding of what these personalized AI tools can actually offer. And beyond that, really great personalization can be downright creepy. What’s the best way to address this?
“I think the trick is to ask permission, you have to be transparent about it. It can’t be a surprise,” said Chris Williams, chief product officer at iHeartMedia, at Transform 2019 in San Francisco today.
Thinking about personalization from the experience level of the customer, including what’s comfortable for the customer, is key for growth, added Chris Hansen, senior director for digital at TGI Fridays.
Hansen focuses on three steps for personalization, he said. 1) identify a problem; 2) tie it to your business goals; and 3) find how the technology will help solve that problem.
By observing user habits — like which meals they tend to order on a particular weekday evening — and then suggesting that same meal the following same weekday evening, Fridays has been able to come off as cool, not creepy, according to Hansen. Most customers are aware that apps collect a number of data points, and need data in order to make recommendations. Offering a solution to the problem of trying to figure out what to eat also dovetails into Friday’s business goals.
With its personalization tools, TGI Fridays has increased engagement on social by more than 500%, and online revenue growth has grown by more than 100%, Hansen said.
Similarly, Doordash also relies on identifying user habits to key in context and make the right recommendations. Doordash found that email reminders were a simple way to stay in the minds of their customers, said Jessica Lachs, vice president of analytics at Doordash. It helped the company not only improve its click-to-open rate, but also its conversion.
Similarly, customer experience was at the forefront of everything it was pursuing. “All of the testing that we’re doing is to improve for customers,” she said.
From the email click-throughs, Doordash was then more comfortable to better predict the kinds of restaurants customers might be interested in, based on previous places they had dined, Lachs said.
What companies have to look out for are contextual clues, and remember that being straightforward about what its tools can do is key to making the customer feel comfortable, not creeped out.
iHeartMedia, for example, will alert a user to its audio recommendations for multiple activities as opposed to just one — since users might have different playlists for their morning routines than they do workouts, for instance, and so personalizing for a specific user is actually personalizing for multiple contexts.
“And we found that the more and more we got it right, the byproduct of that was that users who are subscribers to one of our on-demand services had a higher retention rate, because we built up radio trust with them,” Williams said.
“Hopefully it’s not too creepy for you.”