Walking into a Dan Murphy’s is not entirely unlike browsing Netflix on a Friday night - after endlessly perusing a huge array of options, quite often you’ll just settle for a rerun of something that you know you like.
While there’s nothing wrong with sticking to what you know in the vast and nebulous realm of wine tasting terms, it leaves little wriggle room for Dan’s to upsell its customers.
Dan's, now part of Woolworths Group's spin-out to be Endeavour, knows that most Australians aren't drinking more, and they are getting fussier. So nudging customers up a price bracket is imperative.
Fashionable moderation front of mind, Dan Murphy's recently launched a new Discovery Engine to score its 3.8 million customers against the 5000 wines it has on offer every week, recommending the top product each person hasn’t previously bought.
Andy Sutton, head of data and personalisation at Dan Murphy’s parent organisation Endeavour Drinks, said that the plonk purveyor reached out to two of the world’s streaming media giants to find out how they successfully nudge new things to users.
“Netflix doesn’t recommend you the same stuff over and over again. If you’ve watched the film, you don’t get it recommended, you get something like it,” Sutton said.
“You essentially see [recommendations] that are similar but not the same as products, films you’ve seen before, whether those are the same actors, same genre, same colours… there are multiple ways you can classifying this as being the same.”
Speaking at the Oracle Modern Cloud Day event in Sydney, Sutton volunteered the example of watermelon cider (why anyone buys it, the discerning Briton doesn’t know), as a relatively simple example.
Dan Murphy’s has tagged the pink packaging, sweet flavour, and buyer demographics so that if you buy the watermelon cider, you would receive recommendations for strawberry and lime cider or something with sweet blackberry flavours as something to try in future purchases.
However, Sutton said that Dan’s is still on the journey to developing a capability as robust as Netflix and Spotify’s, especially in regard to wines.
He brought up his mother-in-law’s wine purchasing history as an example, which included a mixture of pinot grigios and sauvignon blancs across a number of different price points - typically between $10 and $20.
“If you wanted to, you could just recommend within that quite niche set of products. So, ok, offer me a $15 bottle of sauvignon blanc, it’s fairly simple to do,” he said.
“What we’ve tried to do with recent items we’ve built in is look at what’s similar to a pinot grigio from a grape perspective, from a region perspective, from a taste perspective and brings those attributes into our model.
“Then you end up with wines no one’s ever heard of like albarino, petit chablis, gruner veltliner - wines that relate back to pinot grigio and sauvignon blanc but actually they bring a much wider range of grape varieties to customers.”
A quick search shows each of those varietals sit at the top end of his mother-in-law's typical spending range, or just above.
It also has the added benefit of “wallet stickiness,” Sutton said, as customers buy more of their alcohol from Dan’s.
While the Discovery Engine doesn’t operate in real-time (befitting of liquor it's still batched) Sutton believes it could one day be ramped up from weekly email recommendations to in-app offers that get sent to customers through the Dan Murphy’s app when they enter a shop.
Having a more nuanced understanding of customer’s tastes and spending habits could have wide-ranging benefits across the larger, vertically integrated Endeavour Drinks group which includes Cellarmasters, BWS, delivery service Jimmy Brings, and the ALH Group of 323 licensed venues and 537 retail liquor outlets.
Completing the vineyard to doorstop experience is Pinnacle Drinks, which operates a number of brands that produce alcohol exclusively for sale through Endeavour’s retail outlets.
By tapping data from Dan’s as to what trends are building - or not, in the case of bubbly - Endeavour could scale its alcohol production accordingly to meet growing demand ahead of other producers.
However, those capabilities will take time to mature in the same way it takes Netflix a while to adapt to new markets.
“We had a call last week with some guys from Netflix, and they had the same problem when they started, which was that it was very good with repeat viewers but when you go to a new country or you’ve got a brand new customer on board, it’s up to six weeks to get a view of what you should be promoting to them.
“They’ve now got that down to three days.”