MillerKnoll is a massive and influential producer of office furniture, equipment, and home goods. Today, they have 16 sub-brands, including Herman Miller, Design Within Reach, HAY, and Holly Hunt.
With so many brands, MillerKnoll struggled with multiple ERP systems, duplicate data, and no clear way to see customers across brands. DEPT®’s data team worked alongside MillerKnoll to address their technical debt, integrate data, and build custom dashboards for their entire organization.
An aging analytics infrastructure
MillerKnoll had an old data warehouse managed by a single data advocate.
Their system had functionality problems and technical debt that was slowing down the organization. It had technical dependencies that were difficult to support, including file export formats that needed custom scripts to decode.
The biggest challenge with the old system was that databases and systems were never integrated. Instead, every few years, new systems would launch, adding to the total collection, spreading data out, and creating incomplete pictures across several “sources of truth.”
This aging and diverse analytics infrastructure hindered MillerKnoll from having a unified view of customers and products.
The Northstar platform
MillerKnoll’s diversity needed to be organized into a coherent picture. To achieve this, we built three core data products.
Cross-channel sales reporting dashboard
With so many sales channels, products, and data streams, it was difficult for them to figure out how many (and which kind of) office chairs they sold. This custom dashboard gave them a picture of retail and contract sales across their suite of products.
Email and cross-channel marketing performance reporting
MillerKnoll’s email marketing program was only a few years old, but they also started dabbling in social media marketing. What ads were converting? And where were they wasting money?
We created two separate reporting dashboards (one email, one social) and then one mixed dashboard so they could view all email and social media marketing activity and results.
When millions of people started working from home during COVID-19, workers bought office chairs in droves.
Out of the millions of people who purchased an office chair, who would likely make additional purchases down the road? How could their team find the best people to target?
Our data science team created a machine learning model that could predict what customers would be more or less likely to make additional purchases in the future. This analysis identified roughly 45,000 email addresses and a whopping 90 million potential revenue for MillerKnoll.
Consolidated data warehouses & better insight
With modern data systems, streamlined data collection, and integrated dashboards across multiple brands, MillerKnoll was able to retire some of its old data warehouses and start fresh.
With the ability to see customers’ activity across brands and how those affect conversions, MillerKnoll’s marketing and sales team can focus on the initiatives and channels that return positive ROI while reducing ineffective marketing/sales efforts.
CTO at DEPT®/AI