#Shoptalk19: Celebrate failure and your data scientists, says Stitch Fix

Stitch Fix is a retail start-up success story. Launched eight years ago, it has since generated over $1 billion in revenue, secured 2.9 million subscription clients, and in 2017 went public on the NASDAQ.

The business encourages customers to exchange data in order to receive a personal styling service with a curated wardrobe posted out to them.

Speaking at the Shoptalk 2019 conference in Las Vegas this week, president and COO of Stitch Fix, Mike Smith, explained how much of the company’s success is down to its team of 100 data scientists and the test and learn culture the company encourages.

He described how innovation isn’t something you can bolt on to a company, but it has to be core to the company’s strategic values.

“Investment in data science is really important to differentiate and understand your customers,” he told Shoptalk delegates, but warned that engineers shouldn’t be “buried” in a group function, but have direct access to the CEO.

“If you think data engineering is a real differentiator to your business, don’t let it sit behind a group – it has a huge impact for the company and touches all functions from merchandising, planning to operations, so if you want to have a broad impact for the company, don’t bury it in a group.”

Smith said a few years back when the company only had 50 engineers, he spoke to a retailer thirty times the size of Stitch Fix which only had five data scientists.

But Smith, who previously worked at Walmart for nine years, admitted it is difficult to create an environment of innovation within a company, especially if you are working on technologies which could be seen as cannibalising the existing business. He also noted that larger companies which create an innovation hub as a subset to the business aren’t actually innovating at the core.

Celebrating failure

Smith explained how his company celebrates failure, or more specifically celebrates the insight it gains from failure.

“It’s super important to have a culture of test and learn,” he said. “And make sure you are constantly talking about those insights so the next test you do is just smarter.”

He said his 100 data scientists – some of whom may have a PHD in astrophysics sitting next to a stylist or merchandiser – have the freedom to work on projects with this test and learn approach. Discussing testing online conversion tools, he said his team should be failing around 75% of the time.

“75% is actually really good, because when you take big bets you learn which will have the highest impact – you can’t find the hockey stick [growth] unless you give people a lot of room to test and learn.”

“The advantage of being over a billion dollar revenue company is that you can learn super quickly what works. It’s the advantage of scale we underestimated until we got to the scale we’re at.”