woman-holding-magnifying-glass-paper-stacks

Data Talks Episode 19: The Critical Importance of Data Context

Lack of Context Can Spell Disaster for Data Analysts

Host: George L'Heureux, Principal Consultant, Data Strategy
Guest: Matt Schroeder, Data Advisor

Whether it’s a seemingly straightforward indicator field like an out-of-business indicator, or a field containing the number of employees, data elements have context that can affect the ways we interpret and use them. But those details aren’t always obvious – and data users don’t always ask the questions they should.

Dun & Bradstreet Data Advisor Matt Schroeder explains that failing to consider the context of the data you’re using can have disastrous impacts on your downstream analysis. He also shares tips for making sure you get the information you need, and how to share the context you know with other team members to ensure that your entire organization benefits.

 

Read full transcript

George L’Heureux:
Hello, everyone. This is Data Talks, presented by Dun & Bradstreet. I'm your host, George L'Heureux. I'm a Principal Consultant for Data Strategy in the Advisory Services team here at Dun & Bradstreet. In Advisory Services, our team is dedicated to helping our clients maximize the value of their relationship with Dun & Bradstreet through expert advice and consultation. On Data Talks I chat every episode with one of the expert advisors at Dun & Bradstreet about a topic that can help our consumers of data and services to get more value. Today's guest expert is Matt Schroeder, a Data Advisor at Dun & Bradstreet. Matt, how long have you been with Dun & Bradstreet?

Matt Schroeder:
Just over 22 years.

George L’Heureux:
And you've recently moved into this role on the Advisory Services team. Can you tell me a little bit about what you're doing in this new role for you?

Matt Schroeder:
Really I'm focused in helping our strategic customers really on optimizing the match. Because we know that when they come to D&B with their customers, suppliers, prospects, it's important to get an accurate match so that they get the data that they need, and get accurate data.

George L’Heureux:
So Matt, we wanted to chat a little bit today about how understanding the full context of data that you're working with really helps you to be successful at accomplishing what your business goals are. Before we get a whole lot deeper than that, though, let's talk about what we mean when we're talking about this data context.

Matt Schroeder:
With data context it's really, what's the definition, what's the shape? What does the data elements that you're receiving actually mean?

For example, if you're looking at sales figures or any type of revenue number, is that summarized by millions, billions, or is it the exact number? If you're looking at other data elements such as employee figures, any type of thing to size a business, is it modeled or are they actual numbers?

George L’Heureux:
That type of numeric precision, definitely one example of context, but it's far from the only one, right?

Matt Schroeder:
Exactly. I mean, sometimes it could be simple as a flag, whether you're looking at ... D&B has a public-private indicator or an out-of-business flag. And knowing what that definition is, is key to utilizing that data. I mean, just by listening to the flag you might think, "Oh, I know what it means. I'm 100% right."

But let's say the true definition of that public-private flag is really, is that particular company, that site, traded on a stock exchange? If your definition is different, that's putting risk in any decisioning that you're making. Or if you look at the out-of-business indicator, is it the whole company's out of business or are they just not operating at that particular address? Two completely different definitions that sometimes people don't think about.

George L’Heureux:
Great points. And when I think about it, I'm also thinking about things like how complete is the coverage, globally, nationally, just regionally. Or are there special values that really are the equivalent of null or something else? That's the type of stuff that I also think about too.

Matt Schroeder:
Oh, exactly. It's really important to understand the scope of your project when you're looking at those data elements. Is it a global project? Is it just a certain region of the world? Is it just the US? Is it on customers, suppliers, or including prospects as well? Because depending on what that scope is, where the geography, are those elements populated? I mean an element that's in the US might not be in Europe for example, or in Canada, or it may be just in part of the world. And what are you going to do with those null values, as you mentioned.

George L’Heureux:
We've kind of talked through a couple of really good examples, numeric precision, those indicator definitions, coverage, completeness. But what if that doesn't happen? What if you don't have that context, what sort of problems can you run into?

Matt Schroeder:
It could throw off your entire analysis or decision making. If you're looking at revenue or sales amount for businesses, for example, trying to define your own sales people's sales territory, so trying to balance it out. If that information's off in any way, you could have more customers, larger opportunities heavily weighted in one area and not the other. And then you're overworking employees.

If you want to have things well-balanced or whatever your goal is, it's important to know the context of that data because it could just throw off your project or what you're trying to accomplish.

George L’Heureux:
You've stated good reasons why we need to have a good grip on data context, but how do we go about getting there? How do we go about getting to a point where not only do we have data context, but we've got enough, and we know that it's enough in order to make those types of good decisions?

Matt Schroeder:
It's really talking to the teams that are providing the data. Whether it's internal data or external third party data, talking to the team that's gathering or providing that. They're going to be the experts, in this case data advisors on that case. Usually in those situations there's some sort of data dictionary, some sort of documentation that can be provided that walks through, hey, what does this particular data element mean? What's the scope of it?

And utilize those. Rely on experts, rely on your ... If it's D&B data, rely on your data advisory team to help you with this. Rely on the delivery people that are providing the information and that data dictionary. There's really no dumb questions. It really comes down to asking so that you understand so that you can properly use that data for the use case and accomplish your goal.

George L’Heureux:
I mean, we regularly deal with customers who have hundreds, if not thousands, of data elements that they're working with. And in order to really get anywhere, you have to be able to prioritize. Are there types of fields that maybe in general are more susceptible to these types of problems, with lack of context?

Matt Schroeder:
That is a great question. I really think every data element has some level of susceptibility. It really comes down to what’s your use case and what you're trying to accomplish. When you're defining that and the data elements you're gathering, you'll be able to parse out.

And really, that's where our expertise comes in or whoever's providing the data. That's where their expertise comes in to help you define what's more important. Some will be elements that are just coming along as reference. Others are going into decisioning and modeling, which will be key on what you're trying to accomplish.

George L’Heureux:
It's interesting, the more and more of these conversations that we have here on Data Talks, the more and more we realize that things really just continue to come down to what's your use case. And it sounds like that's what you're saying here again.

Matt Schroeder:
Oh, definitely.

George L’Heureux:
We've been talking a lot about how to use context information to really make more effective use of data, but let's approach it from the other side now. How do we make sure that if we have data, that we can convey that type of context to downstream users who have to try and make use of it?

Matt Schroeder:
It really comes down to documentation. I mean, if you are having high level reporting, having some sort of document or context in there. Or if you're presenting it, explaining that as you're going through there.

Documentation's also important, along with those data dictionaries, to help represent those values, depending on how often you do these types of projects or reporting. You might not be the one that's doing the project the next time or the next year or the next quarter. So having that documentation, whether it's you or another party at your business, it just, it's going to make your job easier. You're not going to have to start from scratch. You're going to have that fundamental of documentation to be able to run your report and make your decisions.

George L’Heureux:
Yeah, you help others to avoid having to reinvent the wheel from scratch every time. And we know that Dun & Bradstreet already does a lot of this. We have a full data dictionary that goes into that sort of detail like you were talking about.

George L’Heureux:
In your experience, are there any stories that jump out where you've seen the power of data context really take a front seat? Really show how important it is and why it made a difference in a project?

Matt Schroeder:
Yeah, actually just recently, you'd mentioned I just recently moved into this Data Advisor role. And I was brought in to help a particular customer after the sale. They were using a lot of our data, diversity data and other information for reporting and compliance on their suppliers. And they were getting some feedback that they didn't trust the data. There were some questions that had came up.

We rolled up our sleeves, went in, what were the issues? What were the questions? And it came down, I know I mentioned it earlier in our discussion, that public-private flag. You think it's simple. And as we dove into it I found out that, we were matching their suppliers, we were appending the data to it, but the particular customer didn't understand, first of all, it's site specific. And when they were doing the reporting, they would provide information at the parent or the ultimate level. And so they were getting some questions that came back to say, "Hey, you've got this flagged as a private company and it's public."

And what we discovered is, when I went through the definition of the data element them and, "Hey, this is site specific. And is this particular company traded on a stock exchange?" And what we discovered is a lot of their suppliers were owned by larger public companies. And with the definition of this flag, it wasn't. This particular company was not traded on the stock exchange. It was a wholly owned subsidiary by a large public company. So that created some confusion with them and the reporting.

We also discovered, with their suppliers, they were global. In that case, this particular element is only in the US at this point. Just because when you go outside the US, different countries have different definitions of what public and private are. On all their global suppliers it was just coming back blank. Once we rolled up our sleeves and basically help them define what this field was, they were able to adjust their reports for the US. And then globally, we were able to work with them and provide additional data elements to really help them understand which global customers were truly public and which ones were private.

George L’Heureux:
I think, in that answer I found something really interesting, which is that you'd said, "You'd think it'd be simple." And it's not that it's not simple, it's just that your simple understanding may not be the same as my simple understanding, which may not be the same as the real simple understanding.

Matt Schroeder:
Exactly. And that's why it's important to understand the definition and ask the questions. And it may not come down to an element by element level, but whatever project, whatever use case you've got in mind, utilize your teams. If it's internal data, utilize them, tell them what you're trying to accomplish.

Work with your sales and consulting teams at D&B. Explain to them, "Hey, this is what we're trying to accomplish," so that you can get the data elements. And that way they know your scope. They're the experts on their data elements, and they'll be able to appropriately say, "Hey, this is what you need and this is the definition."

George L’Heureux:
Matt, as we prepare to wrap up here, what's the bottom line takeaway that you would want someone who's watching this or listening to this to walk away with?

Matt Schroeder:
It's really just, know your data. Based on your use case, based on what goal you're trying to accomplish, make sure you have a good understanding of that data. There's no dumb question. I've been here 22 years. I learn something every day. Every day that I work with the data, I'm learning as I go. And so that there's never a dumb question. Make sure to ask it and then document it. By making assumptions you're just adding risk to your project.

George L’Heureux:
Well, thank you, Matt. I appreciate you taking the time today to sit down and share some of your perspective on this. I think it's been a really valuable conversation.

Matt Schroeder:
Oh, thank you.

George L’Heureux:
Our guest expert today has been Matt Schroeder, a Data Advisor at Dun & Bradstreet, and has been Data Talks. I hope you've enjoyed today's conversation. And if you have, please let a friend or a colleague know. And if you'd like more information about what we discussed on today's episode, you can visit www.dnb.com, or you can talk to your company's Dun & Bradstreet specialist today. I'm George L'Heureux, thanks for joining us. Until next time.