Jeremy Adamson is a leader in AI and analytics strategy across industries, and author of Minding the Machines. Jeremy has worked with major organizations to establish leadership positions in data science and to unlock real business value using advanced analytics. 

This article features a Q&A from the Data Chats podcast episode featuring Jeremy Adamson on how organizations can leverage data teams and how data teams can work more effectively.  

What prompted you to write about great data and analytics teams in Minding the Machines? 

I had the good fortune to be part of the data and analytics journey for a lot of different companies and what I noticed with the more companies I saw, regardless of if they were big or small, tech-savvy or not, they were driven by their gut. 

At the same time, whenever they made an investment in data and analytics, it tended to backfire. They weren’t getting the value out of it that they were expecting. And what I saw is if you look past all the symptoms and you get to the root cause, it was the lack of connection between the data practice and the business. 

If you spoke with the data practitioners, they would usually say we don’t have the tools to be successful. Data quality isn’t great. They would have a lot of technical reasons, and if you spoke to only them, you would think there’s a technical issue with a technical solution. 

But if you spoke to their stakeholders, they would say that the data people don’t speak our language and they don’t understand the business. 

There was a complete disconnect. Bridging that disconnect was what I was hoping to accomplish with the book.

Do you have any examples you could share of ways that business strategies and technologies didn’t work well together?  

Business leaders may hear their peers talking about data scientists being the sexiest job of the 21st century and feel a need to hire to be competitive in the future. They’ll go out and they’ll hire the most talented group of individuals they can get.

They’ll invest a ton of money, and then maybe start a cloud migration. They’ll get all of this pre-work done, and then at the end of the year, they look at their P&L statement and realize they lost a lot of money on this initiative, and they’ll wonder “what happened?” 

As a result, they may steer the data teams towards things they understand more like reporting.

It’s not fair for the data scientists that are put in that position because they’re thrown out there with a mandate to just go and do data science. They don’t have the relationships in place, nor do they have the foundations in place. They don’t have a seat at the table. There are all these elements that aren’t in place for them to be successful. 

At the same time, business decision-makers lack the technical or analytical foundation to be able to articulate what they need in a way that can be understood and solved by the data team.

So again, right from the start there, there’s no connection between the functions.

What advice do you have for young graduates who are entering data practitioner roles?  

When you’re early on in your journey, you need to go slow and you need to start with low hanging fruit. And for the most part, those aren’t technically complex problems.

A lot of those problems can be solved with Excel. And there’s no shame in that. 

You need to start somewhere, and you need to build credibility. You need to build excitement in the organization, analytical literacy, and you need to steer the culture of the organization towards data maturity and making data-driven decisions. 

I also imagine a new grad from data science – that’s been poached from Google, let’s say – has all these amazing skills and may be worried they might need to do something beyond the spreadsheet or connected to AI or they’re not going to impress people. 

And to be honest, early on in my career, that’s how I felt I could add value was by doing the most complicated thing, showing my work, and really blowing them out of the water! 

I learned that most people care more about the impact than the details. 

What are some effective ways to put together data teams?  

A lot of that depends on the organization and on the culture and level of data maturity. But I’ll say first off, it’s definitely not always a technical problem. 

You need data to be successful. Your data is never going to be perfect. You can’t wait for it to be perfect. 

I would say you need some foundations in place to be successful, but from there I would focus mostly on establishing an understanding of the importance of data.

I think new leaders need to be out there. They need to be intentionally building relationships, and if leaders aren’t spending half their day in sales mode, I’d say that they’re not getting it right. They’re focusing way too much on the nitty gritty. 

In terms of structure, I would say that the Center of Excellence (COE) model tends to work best. Having a COE allows you to have those technical foundations that are a shared resource in the organization, as well as technical specialists throughout the company that can draw on and partner with them to get projects across the line. 

Of course, you still need that raw horsepower. You need those people that can do the hands-on keyboard stuff. But unless you have that functional knowledge or domain knowledge to be able to contextualize the solutions and to understand how to operationalize things, there’s going to be a breakdown on one side or the other. 

What are some of the techniques that you might encourage people to do in order to be more effective communicators when they’re working on a project? 

Stakeholders want to understand what’s happening. They want a very clear connection between the model that you’re doing and the business outcome. They want to know that you’ve considered all of the operational risks with what you’re doing, that you understand the business, that you understand their motivations and their pain points.

All of that stuff is much more important to them than if you used TensorFlow or a neural net. Nobody cares about the technical details at the end of the day. That’s kind of a harsh truth when you get in front of stakeholders. You might want to talk about the cool ensemble technique or XGBoost you’re using and throw buzzwords out there that people like us like to nerd out about, but that just decreases trust. Stakeholders often don’t understand these types of explanations. 

You need to always focus on business value and find the simplest possible way to get the point across. Once you’ve established credibility, you can start to ramp up sophistication. But if you start that way right out of the gate, you put people on a defensive footing. Instead, focus on empathizing with your stakeholders.  

How do you stand out? What can a data practitioner do to make them stand out as excellent? What might they want to avoid?  

That’s a good question. It depends a lot on the types of projects and where they fit in the group. 

One of the things that really impresses me is when somebody gives me something that I can just pull the trigger on right away. I find quite often when people come in, they will present all their work to show they are smart and capable, but only a very terse final answer at the bottom.

They present their work when the outcome was supposed to be a presentation or speaking notes that the executive sponsor can have to defend the project. So then the executive has to turn that work into a presentation. 

If a person is able to understand my needs as a leader of an analytics function, if they’re able to understand what I’m going to do with this in the business context, and they arm me with a deliverable that is fully thought out, fully scoped, and I don’t need to give it any more thought – I find those type of people are just really great to work with. 

I would say one more thing that makes someone stand out is creativity and confidence in their approaches. There are so many great outlets for analytics in companies, but stakeholders quite often can’t identify those opportunities. It’s the people that are at the front lines, the ones that are working with their peers within the functions that can see those pain points, that come up with those ideas that can make those innovative recommendations.

Translate Data Insights into Business Strategy 

As a business leader, it’s important to understand what problems you can solve with data and how to leverage your findings to make better decisions. Data Science for Business Leaders shows you how to partner with data professionals to uncover business value, make informed decisions and solve problems. 

The world of data is moving fast. Understand how business leaders and data practitioners contribute at each stage of data projects to drive results that have a real impact on your organization. 

Learn More