My Patreon
https://www.patreon.com/Batman10514
6. Finally, Leverage Your Machine Learning with Purpose
It's easy for executives to see an ecommerce search results page driven by thousands of rules, then ask how many analysts they need to do some search engine optimization. The alternative is a dynamic recommendation engine that takes what's already there and makes it more efficient — automatically.
"It's the whole idea of optimizing the exact right product recommendation and search for one individual," says Hoyne, "but also making sure we show stuff that makes us money for that one individual transaction. One of the reasons people don't make money in ecommerce is they're not showing the right stuff."
Even in the era of data privacy, where you might know very little about customers from the outset, customers do send signals . "Take signals from across the customer journey – even anonymous customers send — where they came from, the products they're looking at, what customer reviews they are looking at and resonate with, and start building that history."
In exchange for a personalized experience, many customers will provide more data.
"It's like going to your favorite restaurant," says Hoyne. "They know what you like and even though they might make more money, the exchange is mutual, as opposed to being handed a menu each time you show up."
Hoyne says that ML and automation shouldn't be viewed as a cost-cutting initiative, but as an efficiency to improve productivity. Too many companies get caught in the cycle of more rules, manual optimizations, and new-hires to sustain it all. Not every component needs to be done manually.
It can be automated under the right conditions — and to the customer's delight.
My Patreon
https://www.patreon.com/Batman10514