SCOTTSDALE, Arizona. – When you mention data science, you get one of two reactions: excitement or glazed eyes.
Data science is a broad term that many people don’t quite understand. But according to Matt Teddy, VP of Amazon Private Brands, data science is the key to unlocking business value.
Speaking at the Amazon Business Reshape 2022 conference at the Hyatt Regency Resort and Spa on Thursday, Teddy explained that data science — and understanding how to use it properly — was key to helping Amazon (NASDAQ: AMZN) is developing a private brands division that includes Amazon Basics, Essentials, Elements, Wonder Bound and Mama Bear. In total, there are more than 40 private Amazon brands worldwide.
“Amazon uses science intensively in everything we do to solve customer problems,” Teddy told the audience of about 400 people. – We approach problems scientifically [approach].”
When Teddy joined Amazon, he said the goal was to “inject technology” into areas of the company that didn’t have it. Speaking to Amazon Business customers in the session, Teddy pointed out the keys to using data science, noting that many are using it incorrectly.
“Machine learning covers a very broad spectrum, but like leaders, it reveals patterns,” he said. “Data from past models can inform future changes.”
After saying that machine learning by itself is not that useful, Teddy illustrated several examples of how Amazon uses data science to improve products and pricing.
“It’s very rare to look back at the data [and find it useful for making decisions going forward]. How do you understand customer response to pricing?’ he asked. “It’s also not that helpful for how the customer will react to future prices.”
Strict adherence to machine learning in this case can lead to erroneous decisions. Science will tell you that as prices rise, so do revenues. But this is not useful information in itself. Teddy said Amazon adjusts the prices of many items weekly and does so randomly so that it can gauge how the changes affect sales.
“[It’s] to not just do business as usual, but to have actual data on how customers react to price changes,” he said.
Giving advice to conference attendees, Teddy said to use machine learning to get the data you need to make decisions, but keep the human factor in mind. His second point is about creating good data.
“Having good metrics and goals against those metrics is how you develop quality leadership,” Teddy said. “As executives, you need to know that you have to be skeptical of these metrics if they are really moving us toward our goal of improved customer engagement.”
Having the right metrics is important, Teddy said, citing profitability as an example. Clothing and high value items have the highest return rates. But if a customer wants to lower those rates and only uses data related to return rates, the solution is obvious: stop selling clothes and expensive items.
Obviously, however, this is not the correct answer. Teddy said that Amazon has addressed this issue and created a new metric – profitability exceeding expectations. This measured the return rate of an item in relation to the expected return rate.
“What interests me are the indicators of what the customer does next and whether it leads to a long-term relationship,” he said.
Teddy advised the audience to choose where to deploy machine learning, suggesting that the first area to attack should be the cost structure.
“How much do the materials cost? It’s a productivity issue with a lot of hard metrics,” he said.
Finally, as data is generated, Teddy said it’s imperative to send it to production and create a “loop” of the entire process to check for understanding. Failure to do so creates a one-way flow of data and limits the ability to use this information to create change.
You may also like: