Assessing a customer’s lifetime value (LTV) usually requires observing that person’s purchase behavior over time.
But there are more efficient ways to calculate LTV earlier in the customer journey, said Emad Hasan, CEO and co-founder of Retina, a startup that uses machine learning and data analytics to help brands predict LTV and lower customer acquisition costs.
On Thursday, Retina launched an insights tool that builds personas for new customers before they make their first purchase so that brands can optimize their marketing without having to wait for those customers to complete multiple transactions.
The tool, which Retina originally started to develop in partnership with one of its clients, DTC hair care and coloring brand Madison Reed, was in beta for the past six months with a small group of brands, including Brickell Men’s Products and legal cannabis delivery app Eaze.
The offering evolved during the beta process. In the past, Retina used an algorithm based on RFM modeling, which is a method for analyzing customer value that looks at when someone last purchased, how often they buy and how much they’ve spent overall.
But that approach can take weeks, months or longer before a customer generates enough purchase data to reach statistical significance.
To rectify that problem, Retina’s updated insight tool also tracks behavior data leading up to a first purchase, including products that were previously viewed.
The tool can also make predictions immediately following a first purchase using bits of information before a customer establishes a pattern of behavior.
For example, a billing address that differs from the shipping address indicates the purchase is a gift. The LTV trajectory for someone buying a gift, which is usually a one-off, is very different than it is for someone buying for themselves.
“That’s just one variable,” Hasan said. “But if you combine multiple variables like that together, they can inform what a customer’s future journey might look like.”
For Eaze, figuring out a customer’s potential journey is extra challenging, because the cannabis industry is changing so rapidly, said Steven Hester, the app’s director of marketing.
There’s no typical cannabis consumer, he said. Baby boomers, for example, are a fast-growing segment of the market, and their needs are very different from millennials which are different again from Gen Xers.
“The market is changing year by year,” Hester said. “In 2017, medical marijuana was the law of the land, and that’s so different from today – we don’t really have historical data to rely on.”
Using Retina, which is integrated directly into Eaze’s data lake, Eaze has started to refine its targeting for specific customer personas and to confirm a few old hypotheses, including understanding the value of more stereotypical cannabis consumers who buy joints or “flower” versus someone who buys a topical product, such as CBD cream or hemp balm.
“We can see how they respond to deals on the platform or certain types of advertising content, like content to do with sleep or CBD,” Hester said. “That helps us understand how we should market to people who come in from different channels based on early data signals.”
Eaze learned, for instance, that customer LTV is noticeably higher (no pun intended!) among users that take advantage of deals and offers, and that with only a few exceptions, deal-forward marketing drives incremental engagement across the entire customer base.
“Being able to classify those users as early as the first deal when it’s less clear if they’ll respond to site-wide deals helps us figure out how to engage from that first customer touchpoint,” Hester said. “We’re making sure that the narratives we create are actually confirmed by data.”
