How the Data Team at Trane Advanced from Reporting to Machine Learning
Written by Kara Bobowski
Director, Decision Science & Advanced Analytics
When Karla Hillier joined Trane Technologies as the director of decision science and advanced analytics, she was on a mission to partner with sales to provide a key data tool for sales leaders. “We wanted to create an interactive tool for sales leaders that would help them make better decisions about how to help our dealers grow their business,” she says.
What she learned over the last two years was that taking the agile approach worked for developing and launching a dynamic reporting tool and for leading the team on a path to adopting decision science and machine learning.
Originally, Hillier went looking for a way to organize the team around common goals and accomplishments – and landed on using an agile methodology for their delivery of data to internal customers.
The Business Challenge
The sales model at Trane is through company owned and independent distributors, who sell air conditioners and furnaces to dealers, who then ultimately sell to end consumers. “The dealers are independent businesses,” Hillier explains, “but if we help them grow, then that helps our business growth. So, giving our sales leaders more detail about the dealers’ business performance, and about what they were purchasing from us, was going to help everybody create a win-win situation.”
The challenge was that sales leaders were spending a lot of time going to different data sources to find this information and then creating their own pivot tables to bring data together. Hillier says, “It was taking them a long time to get the information to run their business, and they weren't spending their time where they needed to — with their customers.”
An Agile Approach to Data
Hillier came up with a somewhat novel solution for her data team by adopting some aspects of agile methodology and applying it to their data production. As Hillier says, “I gave some thought to a better way for us to work together and landed on using a Kanban approach to help get us all on the same page.”
Kanban was a new concept for her team, but she had support from a coach at Trane. She and her team also recognized that because they weren’t doing traditional software development, they might have to make some adjustments.
“In fact, our team really struggled with assigning points to tasks,” Hillier explains. “That was one of those classic agile behaviors that we ended up not doing. It was really hard because so much of our work is not repeatable. So, we ended up doing hours instead.”
Hillier’s team stayed flexible in terms of adopting pieces of the agile process that fit with their work and leaving behind the parts that didn’t. Hillier notes that the process continues to evolve as projects develop and as there are different needs from developers and stakeholders.
Agile isn’t one-size-fits-all in the data space.”
One of the most positive outcomes for her team? “Knowing what we wanted to deliver at the end of those sprints,” she says. “That helped us to have interim deliverables because one of the dangers of analytics is that you end up working for a long time on something that you've never shown to the customer because it's not perfect.”
In fact, one of her key learnings from the process was to get something in front of their internal customers regularly to help drive iterative behavior and continuous improvement.
As Hillier notes, “It's really hard to draw a direct line between a dashboard and a business outcome.” Instead, they are tracking the adoption of their new sales tool, as well as how often the sales team returns to it.
In a given month, they have 50% of their sales leaders log in and use the tool.
They see some positive correlation between usage of the dashboard and business results, but Hillier does not want to solely credit the tool. Even more exciting, their success with reporting and adopting agile enabled them to advance into predictive analytics and machine learning.
“By starting with historical data, we helped everyone get comfortable,” she explains. “The team built up trust with sales by providing something of value with the dashboard tool. When we moved into predictive analytics – something more complex – there was already trust there.”
What’s really exciting is that the success with reporting opened up the door to machine learning.”
Hillier’s team can now predict dealer attrition. “We found out what the behaviors or purchase patterns are that are predictive of when a dealer is going to stay or go. And now we have a dealer attrition model, predicting which dealers might leave the brand,” she says.
The team provides that information to sales so they can intervene, find out what is happening, and ideally prevent that dealer from leaving.
In addition to having a flexible approach to how they implemented agile methodology and a great partnership with internal stakeholders, these were some of Hillier’s key takeaways about their journey.
- Provide a Hands-On Rollout – And Never Stop Training
Hillier explains that there were a number of layers within the data on the sales dashboard they created – from a high level all the way down to specific dealers. “They had to put in a little bit of time and effort to get it to answer their questions,” she says. So, Hillier and her team made the rollout to the sales team very hands-on – first with sales leaders, then the entire team.
And, remember that you’re not really done post-launch. “You need to remind people that the tool is there and to take advantage of it, as well as introduce new features,” she notes. “It’s an ongoing commitment from both sides to continue to get value out of it.”
- Early Wins & Interim Deliverables
“We identified a couple of early adopters and worked with them on a one-on-one basis to help us answer some questions,” Hiller shares. “That real customer feedback helped us to improve the end product.”
In addition, having short-term or interim deliverables helped to keep her team in front of their internal customers. She continues, “Picking things like dealer attrition that people can wrap their heads around was a good way to start. It's not that hard. People understand what it is.”
- Focus & Discipline
“It takes leadership discipline to create and protect a team like this,” Hillier points out. If her team had been constantly asked to do other projects, they wouldn’t have been able to complete their critical initiative. “Key leaders really protected us,” she says. “We had permission to say no to everything else.”
Hillier sums up what she and her team learned this way: “I would say having a big project or big objective out there like we did to deliver the sales leader dashboard was a really unifying and energizing goal to have. We proved out that if you put multiple people behind one, big objective, you can deliver something pretty phenomenal.”
Special thanks to Karla Hillier and Trane Technologies.
by CDOs, for CDOs
Join the conversation with peers in your local CDO community.
BLOG | MAY 18, 2022
BLOG | MAY 18, 2022
The CDO’s Perspective: Data-Driven Culture
Creating a data-driven culture is the number two priority for CDOs in 2022. What are data and analytics leaders telling us about their progress on data culture? Here are some CDO perspectives on how their organizations are adopting a more data-driven culture.
TOWN HALL INSIGHTS | MAY 11, 2022
TOWN HALL INSIGHTS | MAY 11, 2022
Harnessing Unstructured Data with AI
The amount of data created and consumed across the globe is growing at an unfathomable rate. But, enterprises can utilize AI to channel this unstructured data toward innovation and sharper data-driven decisions. Read insights from this New York CDO discussion on harnessing unstructured data.
BLOG | MAY 5, 2022
BLOG | MAY 5, 2022
3 Areas of Focus for C-Suite Leaders in 2022
After a couple of highly unusual years filled with business uncertainty on the back of a global pandemic, what are C-suite leaders focused on this year? Read the 3 areas executives cited in our survey as their most important priorities for 2022.