3 non-negotiables for data project success

Mar 09, 2022
AnalyticsBusiness IntelligenceData Science

Identifying the right data projects takes discipline. At workwear company Carhartt, these are the must-haves for setting projects up for success.

thread weave needle
Credit: Thinkstock

Going into the pandemic, workwear company Carhartt was “probably as far away from being a virtual collaboration organization as you could be,” says Chief Digital & Information Officer John Hill. But like so many organizations, Carhartt has had to rethink how work gets done in today’s hybrid workplace.

Part of that, says Hill, is “push[ing] down decision making to the lowest level in the organization”—a move that will increase reliance on data.

“So that is going to put a lot of pressure on the data organizations to make sure that data is available to be served up to that decision maker.  And yet, at the same time, it all ties together, so we know that we’re not having inconsistencies across the organization.”

Hill and his colleague Jolie Vitale, director of BI & analytics at Carhartt, sat down with CIO senior writer Thor Olavsrud at the recent Future of Work Summit to discuss how Carhartt is changing to a workplace culture of access to data & analytics.

What follows are edited excerpts of that conversation. Watch the full video embedded below for more insights.

On building a data science function:

Jolie Vitale:  We have two data scientists today, and a couple of the things that we learned.  So, when we first grew the team, a lot of what we were doing is we were going to the teams within the organization, we were socializing what data science could do and trying to demystify it and make it a lot more approachable.  So, we developed a really nice meaty backlog.  And then we looked at all the projects in the backlog and we said … we need to develop some way to know what the right types of opportunities are for us. 

Aside from aligning with our company goals and the value that we’re trying to drive, there are really three non-negotiables.

The first one is that we need to know upfront what that model is going to need to be able to achieve, if it’s going to be adopted or drive the value that we’re hoping from it.  So, the clearer we are on that upfront, the more that we’re set up for success. 

The second thing is that we need a business partner within the business team who’s fully invested and the implementation is also resourced.    

And then the third piece is that we do data discovery on a small scale at the beginning of the project, and that discovery needs to basically prove out that we think that things will be achievable, that we have the data that we need, that the processes are something that can be modeled.

Mostly, what we’ve learned is that it can’t be done in a silo.  And the more that that future end state is crystal clear, the more successful we’ll be in achieving it.

On adopting a product team approach: 

John Hill: As Jolie mentioned, it’s really important to have that product owner, somebody who’s making the calls on something so they can quickly get to minimum viable product and build from there. 

The interesting thing that we see is that internal clients oftentimes will go to a Jolie type of person and say, “I need help with analytics.”  They’ll go to an application person and say, “Well, I need help on application enhancement.”  And what we want to do is build—we call them squads—that essentially are end-to-end to support a particular business function or a business unit inside Carhartt.  So, they would have everything that’s needed from a business analyst, a data analyst, data engineers, BI engineers, all the way through the application folks and even security and kind of the DevOps side of things.  So that we’re able to think about the work for that organization as one entity, and we’re able to line up resources to do it. 

On managing transformative change:

Jolie Vitale:  Transformation is, by definition, something that requires big change.  And so the more that you can really gain adoption early and often, the more successful you’ll be.  So those feedback cycles and the iterative approach allow us to be able to deliver things in very small manageable chunks, so that’s not one big change at once.  And the collaboration gives us that feedback when you do that, so that we know that we’re moving in the right direction. 

And especially when we are choosing who that collaboration partner is going to be and who we’re going to involve in that process, what we’re really looking for is somebody to partner with who has high trust with people who are going to be impacted by the change [and also has] their own transformation mindset, that they’re not just looking at how we’ve always done it but what things could be.  And once you can find those people, and you can bring them on the team, that really helps to make sure that you’re going to be successful once the transformation is complete.