SAN SEBASTIAN: US mobile network Sprint used artificial intelligence to personalise messages to customers – and not only boosted retention rates but created a ripple effect that slowed churn rates among its customers’ friends.
The project, carried out in partnership with AI marketing platform Amplero, won the AI category of the recent I-COM Data Creativity Awards, where Glenn Pingul, VP scientific marketing strategies at Amplero, explained that the project was based on the premise that “no customer is an island”.
“The strategy was ‘do not only look at the customers that you usually target with marketing, but understand the influence on the customer connected to the customer we targeted’,” he explained. (For more, read WARC’s report: How Sprint’s ‘ripple effect’ slows churn rates.)
The targeted customers were those in the early stages of their relationship with Sprint – those who’d signed up within the past 60 days – as they were most likely to churn.
An SMS campaign using over 120 combinations of messages was run to take account of different users’ phone usage and payment patterns; artificial intelligence monitored responses so the team could test, learn and adjust as the campaign went on.
“What it is doing is it’s progressively learning what is the right combination of attribute of the customer, with the right kinds of messages,” said Pingul.
As one might have expected, the closeness of the contact between people was a far stronger indicator of how much ripple effect the campaign would have, rather than the sheer number of contacts any one customer had.
And Sprint was able to put some numbers on the effect. “We saw a 35% increase in engagement of the targets,” said Pingul. “But what we also saw was a 10% rise in engagement among their friends.
“That basically translated to a 29% retention lift and we translated that to an additional 85 cents revenue from their contacts.”
The project, carried out in partnership with AI marketing platform Amplero, won the AI category of the recent I-COM Data Creativity Awards, where Glenn Pingul, VP scientific marketing strategies at Amplero, explained that the project was based on the premise that “no customer is an island”.
“The strategy was ‘do not only look at the customers that you usually target with marketing, but understand the influence on the customer connected to the customer we targeted’,” he explained. (For more, read WARC’s report: How Sprint’s ‘ripple effect’ slows churn rates.)
The targeted customers were those in the early stages of their relationship with Sprint – those who’d signed up within the past 60 days – as they were most likely to churn.
An SMS campaign using over 120 combinations of messages was run to take account of different users’ phone usage and payment patterns; artificial intelligence monitored responses so the team could test, learn and adjust as the campaign went on.
“What it is doing is it’s progressively learning what is the right combination of attribute of the customer, with the right kinds of messages,” said Pingul.
As one might have expected, the closeness of the contact between people was a far stronger indicator of how much ripple effect the campaign would have, rather than the sheer number of contacts any one customer had.
And Sprint was able to put some numbers on the effect. “We saw a 35% increase in engagement of the targets,” said Pingul. “But what we also saw was a 10% rise in engagement among their friends.
“That basically translated to a 29% retention lift and we translated that to an additional 85 cents revenue from their contacts.”
Sourced from WARC
