How to develop a data-driven culture
Blog: Colin Crofts - Business Process Improvement
Recently, a merchandise manager at thredUP, a San Francisco-based online retailer of second-hand clothing, wanted to better understand how items moved through the company’s supply areas. He would personally inspect how long it would take items to go through intake, get inspected, and be packaged and shipped out. But there was a better way to do the job.
l Ghorai, the company’s vice president of finance, planning and analytics, told the manager that he could track all of the steps through data. ThredUP records everything about its products from the time it agrees to take something from a supplier to the time items sell. This tracking enables the retailer to put every aspect of its operations under a magnifying glass, and frontline workers are starting to embrace this kind of data-driven culture.
“If we can show the value of data to people in their jobs, it’s going to take off,” Ghorai said.
Being a data-driven company
has never been more important as data and the ability to understand it becomes a difference-maker. In a survey
reported last month by The Economist
, 59% of respondents who described their organization as “data-driven” said their company is more profitable than competitors, compared to 40% who said they do not work in a data-driven environment.
It may sound simple to make data a bigger part of the decision-making process, but it’s no small feat for thredUP. Being a retailer of second-hand items, it deals mainly in one-off items with unique tracking numbers, which makes recording and analyzing purchases and sales a major undertaking.
To meet this challenge, thredUP utilizes a pair of MySQL databases to store all its data. Also, the company uses a self-service BI tool
from Santa Cruz-based Looker Data Sciences on top of the data store. By giving access to the tool to workers in all areas of the business, the company has become a more data-driven organization.
Ghorai said data analysis is used in the following areas of the business:
- The distribution center uses the BI tool to manage the intake and shipping of clothes
- The merchandise team uses it to understand what clothes are popular and how to price them
- The marketing team uses it understand customers and craft targeted marketing campaigns
- The executive team uses it to monitor all this activity
But data wasn’t always so ubiquitous at thredUP. Ghorai said the company implemented the BI tool in 2012. Prior to that, any data query
was developed as a one-off project. This approach put strain on the development team and made it difficult or impossible for most workers to ask questions of the data.
When the company implemented the self-service tool, it made sure to let everyone access it, Ghorai said. A lot of companies start building a data-driven culture either from the top down or from the bottom up. But thredUP took the approach of trying to engage everyone right from the start.
This approach doesn’t mean it was a smooth process to get buy-in from every worker. Ghorai said some employees have less interest than others inusing data-driven tools
, and only a small portion of the reports he and his team create see much adoption. But when something does work, it can improve a process, just as in the case of the merchandise manager. The key is making the tools available for those who are interested and making sure the data is relevant to people’s jobs.
“As you get more data, more people start to ask questions and you can go deeper,” Ghorai said. “With all of these tools, if people see a use they’ll go for it.”
He agrees somewhat with the conventional wisdom that one key to getting frontline workers to buy into BI tools is for senior management to use them. But this isn’t enough on its own. Making tools available to everyone and supporting interested workers is also important.
Ultimately, building a data-driven culture can help an organization take advantage of opportunities in the modern marketplace, Ghorai said.
“What really drives a lot of our analytics is understanding our business,” he said. “We have to be able to quickly identify where our profits are. The organization has to be data-driven.”