fbpx
HomeNEWSTop retailers share how they’re powered by data

Top retailers share how they’re powered by data

At The Register’s ‘Powered by Data’ event on Wednesday morning, we partnered with Paydar to share insights from Briscoe Group’s general manager of ecommerce Dave Hughes; Trade Me head of data and insights Dr Kathryn Hempstalk; and Paydar co-founder and TRA partner Anthony Ede. Here’s some snippets from their presentations.

At The Register’s ‘Powered by Data’ event on Wednesday morning, we partnered with Paydar to share insights from Briscoe Group’s general manager of ecommerce Dave Hughes; Trade Me head of data and insights Dr Kathryn Hempstalk; and Paydar co-founder and TRA partner Anthony Ede. Here’s some snippets from their presentations.

Dave Hughes, general manager of ecommerce for Briscoe Group.

“You can get really good analysts that don’t understand business and really good businesspeople that don’t understand data,” says Hughes.

He says that when presenting data internally, it’s important to keep the presentation simple: “If you get too clever and include too many numbers, you lose your audience.”

Rather than continuing to throw numbers at their audience, Hughes says, presenters should keep bringing it back to what those numbers mean for their company, and what course of action is being communicated. He suggests the presenter selects a single number that’s key to their business, such as topline sales, and then keeps reconciling their presentation back to that number.

Hughes is passionate about data governance. He says it’s important for companies collecting data to make sure the data they’re using is recent enough to be accurate and relevant, and also to be regularly reassessing their privacy policy so as to keep up with evolving legal requirements in that area.

He recommends companies commit to an “experiment and learn” approach: “Data goes really well with experimentation.”

When Briscoe Group first experimented with offering click and collect at a single store in Botany, Hughes says, the pilot resulted in “a page of problems”. The group was able to resolve these problems before rolling out a nationwide programme which was “far superior” to the original Botany offering.

Hughes says retail data tools are now accessible enough that retailers should treat their selection like buying a car – “It comes down to what you like, not what they do,” – but noted that just because a sales rep tells you that your competitor is using a piece of technology doesn’t mean you need it.

For example, he says, Briscoe Group has so far chosen not to invest in AI: “Just because everyone else is using AI doesn’t mean we can monetise it.”

Dr Kathryn Hempstalk, head of data and insights for Trade Me.

Hempstalk says Trade Me has a much greater scale of data collection than many retail businesses. It sees around six billion searches every year, which, combined with other data sources, gives it around 30 terabytes of compressed search data to use.

The extent of Trade Me’s market penetration means the company can make inferences about what’s going on in New Zealand based on what’s happening with their site. For example, traffic drops during rugby games and episodes of Shortland Street.

“Anything that’s a major event in New Zealand, we can see it reflected in our traffic patterns on the site,” Hempstalk says.

However, like many businesses, Trade Me hasn’t always used its data creatively. Hempstalk says that historically, analytics was part of finance, explaining that changes in how data was handled began about five years ago: “We really only used data to inform about our business, not to change the product.”

The biggest contributor to Trade Me’s working more with data and analytics has been putting it all in one place on a data platform rather than warehousing it in different locations, Hempstalk says. She shares Trade Me’s main points of interest from data: understanding the business; understanding customers; and understanding product.

Trade Me is now using its data to create new products, such as its Property Insights tool, and the algorithm that auto-recommends product listing categories. Some have needed tweaking – Hempstalk says an early version of the “Did you mean…” algorithm which aims to circumvent misspellings was stumped by New Zealanders’ idiosyncratic typing.

“There’s a lot of words that Kiwis use that you wouldn’t find in a normal dictionary, so we actually built our own dictionary.”

Anthony Ede, partner at TRA, co-founder of Paydar.

Ede says New Zealand retailers should be paying attention to data and AI as these tools offer them an edge in a market where they’re small fish in a big global pond. 

“It’s at a level where New Zealand as a whole isn’t going to be able to compete, let alone individual companies, so we have to do something different.”

Google is spending nearly twice as much as New Zealand’s national research and development budget on AI alone, he says. 

“The silver lining to these big tech companies investing so much in AI is that they want to establish it as a platform… they want everyone else to use it too.”

The technology is currently in active use in retail activations like Amazon Go, the Mashgin intelligent self-checkout, and Imagr’s Smartcart technology in New Zealand, Ede says.

It’s also got applications in logistics, where Amazon, DoorDash and Domino’s are experimenting with drone or robot delivery. In the payments space, there’s ecommerce platform Granify and pricing and promotions product Eversight.

“AI is not something that’s going to happen in 20 years and our kids are going to worry about it, it’s happening now,” Ede says. “New Zealand needs to adapt to this in a different way, we can’t just go head to head.”

Getting started with AI isn’t hard, says Ede. He recommends that New Zealand retailers leverage their physical stores, explore local partnerships and use external data sources to create innovative products: “Start small and start now.”

“You don’t have to do everything in this space just because some other vendor is doing it, but you should be doing some experimenting.”

Rate This Article: