Alexis
Hi everyone, welcome to Talk Tech with Data Dave. I’m Alexis, and I’m here to talk with my good friend Data Dave about all things tech, all things data. 

Data Dave
Good morning, Alexis. This is Data Dave or Dave Wilkinson, CTO of D3Clarity, and I’m pleased to be here to see whether I can answer some of your questions. 

Alexis
This question’s a good one, so I’m excited. So, our question today, Dave, is critical data elements, what makes them so critical, and how do I know if it’s not critical? 

Data Dave
OK, so that’s a very good question, Alexis. When we talk about, certainly commercial data and data governance, right?. So, when we talk about data governance and data that describes an organization, that describes what is happening in a business, the term “critical data elements” really comes back to what we talked about a little bit last time. And drawing an enterprise data model and understanding the data that describes that enterprise.  

So, a data element clearly is an element of data. What makes it critical? That’s a good question. 

Alexis
It seems like data elements can be kind of everything but… 

Data Dave
The data elements can be everything. There can be an awful lot of things as we talked about before, data is just evidence that something occurred, and the data element is a particular little piece of that evidence. So, a critical data element, the way I define critical data element is as a data element that is required in order to transact with an entity.  

Now there’s a lot baked into that. Let’s unpack that a little bit. An entity, as we discussed before, is a thing, a person, an organization, a product, a thing, a critical thing that can be identified. So, a critical data element for that thing could be one of the identifying pieces of information. Right?  

The key, the UPC number, your Social Security number, that’s a critical data element for identifying you as Alexis, right? Your name, your address – they’re critical data or could be critical data elements to identify Alexis, right? So, a critical data element. When you go a little bit further, it becomes the thing that allows you to transact with that. So, if you become a customer or you have an interaction with you, then what did you buy and who are you become the critical data elements.  

If you think about a retail situation where you go into the store, and you purchase a loaf of bread. From a business perspective, the store that sold you that loaf of bread, what are the critical data elements, the ones that really identify that that occurred and really tell you something about it?  

Well, realistically, the only thing that really matters in a true retail situation is that one loaf of bread was purchased for a price. You don’t actually have to know anything about the person who did the purchasing. You don’t have to know who was operating the till. You don’t have to know. You might want to know what day of the week it is or what time it was, right? That might be useful when it comes to inventory, but the critical data element is one loaf of bread.  

So, what was it exactly? What was it? So, was it white bread? Was it wonder bread? Was it whole-grain bread? What kind of bread was it? That’s probably interesting from a, “How do I restock this item,” perspective, right? The fact that one was sold and the fact that it was sold for this price, those are probably the three critical pieces of data in that transaction. Does that make sense? 

Alexis
So, when it’s critical because it’s, it’s critical, it’s the thing that really matters. 

Data Dave
It is the smallest set of data that is necessary to describe that transaction took place. 

Alexis
That’s beautiful. The smallest set of data that you need to describe that something took place. 

Data Dave
And that’s exactly right. You use the word “need” and that’s absolutely crucial in this. I tried not to use the word “critical”. Yeah, that is absolutely crucial in this definition of critical data element. It’s the minimum amount of data necessary to describe that transaction. Or to describe it so.  

So, that leads you into the thought, “There has to be a purpose behind the transaction.” For it to be critical data, you have to identify the purpose that you’re doing it.  

We just had a purchase, the creation of a new customer, or a new entity. The creation of a new employee. All of these actions are going to have their own critical data elements. Right? Because there is a purpose inherent behind it. It is critical for purpose, right? This is very important when we get into data governance. And start to look at if we are going to govern the data of an organization or put some more discipline around the data, what data do we put that discipline around?  

There can be an awful lot of data as we’ve discussed previously. So, there can be a tremendous amount of data. Where do we focus? What do we put our efforts into?  

Well, if you start to say, “What are the critical transactions, the critical things most important to this business.” A sale? Right. A purchase? A new employee? Whatever it might be, then it becomes. “How do those come into place? How do we know the high quality? What does high quality mean? Who owns them? How do they get defined?” 

If you think of bringing on a new employee? You’ve hired this person, but you just starting this person, what do you need? Do you need the person’s name? You probably need the tax information, so you need the form. You need the date that they’re starting, and you need who is their hiring manager. Right? And that should have been gathered through the hiring process. As you hired this individual, you should have confidence that by the time their first date and you’re doing the onboarding training that these critical elements are there, and they are trustable. Now, you might validate them and whatever, but you’ve made decisions based off of them.  

So, this is where you start to say, “Okay, this critical data element for this purpose originates here, and it originates with this owner because…” and so you start to build this, this tracing of those elements. Does that make sense? Does that help? 

Alexis
Yeah, let me ask you this phrase in the form of a podcast. So, we have a podcast now because we’re super fancy like that. And the critical data elements, if I’m understanding you correctly, the critical data elements of this podcast would just be the name of the podcast. The date that we publish and maybe the date that we record, but probably not even that. Would you say that that’s right? Or did I miss something? 

Data Dave
I would say that’s right. Now you might want to have a subject. You might say the subject is critical, but I mean, is it critical or is it just interesting?. The subject is actually just interesting. It’s not going to be very good without it. You might say that your name, my name and the subject are. They’re not necessarily really highly critical. Quite simply, the fact that it exists in the data publication and where it is published, the link to it is probably a critical data, because nobody can watch it. What’s the point in publishing it if nobody can watch it? 

Alexis
Right. 

Data Dave
So, where it is published is probably another critical data element. 

Alexis
That’s the, what you said earlier – like the smallest amount of information that you need. That was totally life changing for me because that really helped me understand, you know why something is critical. 

Data Dave
Right. Yeah, it’s the minimum set of data in order to do something with purpose, right? So, it’s the minimum and nothing less. The absolute bare minimum, nothing less. 

Alexis
So, you just said a little while back, something along the lines of like when we get into data governance, that’s when identifying these critical data elements is so important. We’ve talked a little bit about data governance before. We’ve kind of hit on it during a lot of our podcasts, including the last one we recorded. But at what point in that kind of data governance journey, if you will, do you try to identify those critical data elements? Sooner or later? 

Data Dave
I’m going to give you some bad answers. All the time. And early – as early as possible, right?  

Alexis
Early and often, yes. 

Data Dave
But yeah, exactly. But let me clarify that a little bit and be a little bit more meaningful in my answer. The purpose of your critical data elements is very, very important, right? And for identifying this… So, as we talked about last time when we talked about the enterprise data model and that building that map. What we often look at is the priority of the business as well. So, as you can imagine, when you’re looking at an entire business, there’s a lot going on. There’s a lot of data, there’s a lot of information flowing around an organization, right? When you start on a data governance journey. You try and start with some amount of priority, right? “Where is my data in the worst shape?” is one question, but equally as important or more important is, “What are the most critical transactions that my business performs? And how do I increase the confidence in those critical transactions?” 

If you start there, then what you want to do is not boil the ocean on these data elements, right? You don’t want to look at all of them. Let’s look at the critical data elements for the most critical transaction. If you say, “We think we can make our sales more efficient as a construct,” then we say, “Okay, what are the critical data elements for sales in order to complete this? Does the salesperson have the right data to effectively perform that sale? Does the accountant have the right information to send the invoice?”  

So, invoicing becomes that transaction. So, does accounts receivable have the correct minimum set of data with high confidence that allows them to send those invoices at the end of the month or whatever, whenever they do it right? And who owns that particular piece of data. Right. For example, I own my timesheet.  

Alexis
I own invoicing on it, which is why I bug you every month.  

Data Dave
You own invoicing, right? Right. Exactly so. So, that is actually a critical data element for you to be able to send that invoice, but you don’t own that. So, it is actually a pain in your neck when I don’t enter my timesheet because that is data that is owned by me that you need.  

You require it is a critical data element to send that invoice right that my time that I put in that month is a critical data relevant for that piece, and you need it in order to do your job. If I don’t supply it in a trustworthy, timely, efficient manner, then you’re the one that feels the pain. Right, so that is a critical data element for that purpose. 

Alexis
That’s the truth. 

Data Dave
Right. Now, which project I’m on, and other things those are also critical data elements surrounding that, but that invoice, that is a crucial part of that structure.  

So, you start to see how this builds up, and where you look to start to say, “Okay, how do I identify them? When do I identify? And what do I do with them?” Right, so it’s, “What are the critical transactions, right. What do I put? Let’s put those in priority order for how to fix them.” Understanding they can only be one priority number one. Right, so the top of the heap and then,  “What are the critical data elements? Who owns them, and how can you have higher confidence in them?” And you have to think about it for that purpose from that perspective, “What is the purpose? What do they describe? Why do they describe them? Who owns them? How do we make them more complete?” Does that make sense? 

Alexis
Yeah. Let me sum it back to you and tell me if I heard this right. If a client would come to you and say the question that I just asked, “How do I identify my critical data elements?” Before we did anything we would say, well as always, “What answer are you trying or what question are you trying to answer?” and then, more than that, “What is the most critical part of everything that’s going on?” 

Data Dave
That’s right. So it would be, “What transaction? Yes. Can we identify critical data elements? What part of your business do you decide is most critical?” Right? So, we’d follow up with that question, and we usually end up with a priority stack of we want to improve: the customer onboarding process or the sales processing or the invoicing process or the account reporting, the business process, or something like that.  

So, we put those in priority order as to what purpose we are solving, then we start to look at, “Okay, what is now the minimum set of data that you need in order to perform this action for you? What is the minimum set of data that you require and nothing more? What is the absolute minimum that you require in order to send that invoice at the end of the month?” Right. And now, “How much confidence and how much checking do you do on each one of those data elements?” Cause you probably… I’m just picking on that as an example, right? You probably think about, “Okay, which client am I sending it to? Do I have the address of the client? Have I sent it to this client before? Is it a new client?” You put different scrutiny if it’s a new client versus when you’ve sent to before. “Who is the resource that you’re including the hours for in our business, right? Or the product that has been sold? Do you know which product has been sold? Do you know how many of them? Right. And do you know who the client is?” And that is probably the critical data element. 

Alexis
Yeah, I was just thinking that I was like, if I went through my list for invoicing of what I would need, like probably the only thing I would add to that is like the rate at which we we bill our clients, yeah. 

Data Dave
Right? So that’s the price. Right, so, it’s the price, but that is derived from the number of hours in our business. It’s derived from the number of hours that have worked and who the resources are within the context of the project. So that’s a derived set.  

So how do you have confidence in those, and how do you track them is the question.  Who owns them? Where do they come from? Right? So a new client was created.  

How do you know the billing address of that client? That’s a critical data element, as you trace that backward and start to add governance behind that, we make sure that you know there has actually been a postal validation or a validation on, if we do electronic invoicing, a validation on that e-mail address. Have you sent an e-mail to that accounts payable organization previously to make sure that it can at least get through? How much confidence do you have in that? What data governance was created when that customer was signed? Who created that? Do you trust it from Mary or Joe more, right? Et cetera.  

And it’s what processes and what procedures do we put in place to have high confidence in those critical data troubles. Once you’ve done the critical data elements, of course, from a data government perspective, you can start to say, “Well, now that we’ve done the critical ones, and we’ve got confidence to the critical ones, then let’s expand our definition of critical.”   

You might say, well, “What work was done in this hour?” You don’t actually need to know what work was done. For that number of hours or what the product was used for, or what color the product was in that initial realm, but it is useful information as you go further down the road. Further down the journey, right, and you decide whether as a data governance organization, you say, “Okay, well, I’ve done that process and got that one. I am out now. I’m going to move on to another process. Right? 

Alexis
So, so critical data elements first. Everything else after. 

Data Dave
Everything else after. Yeah, exactly. And you’ll cascade through them. Because you’ll start to say, “Okay, my critical data every month for invoicing is knowing which customer I’m going to send the invoice to. Okay, so that’s an input. The customer is an input to the invoicing process, but it’s an output to the customer onboarding process.” 

So we might say, well, “If we don’t have confidence invoicing time on who our customers are, then the next process we should look at is customer onboarding.” 

Alexis
To make sure that we’re getting it right. 

Data Dave
To make sure you’re getting the right data. And the right focus is on it from those owners so that we’re building this efficiency and there’s data flow, and there’s data chain through the whole organization.  

So, you’re not having to chase down X, Y, and Z people from a customer onboarding point of view because you’re not confident in the data that they supplied you. And you walk through the organization in this way to identify the critical data elements, and again, it’s the minimum set of data necessary to perform the task. 

Alexis
It’s really helpful for me. Listeners, if you don’t do invoicing, I’m sorry that Dave used a very Alexis-specific example here, but I understand it a lot more now I must say. If you can’t tell, it’s the beginning of the month and I need to invoice.  

Well, thanks Data Dave for talking me through that today. I really, really appreciate it. That was super helpful and really gave me a better understanding of why maybe one thing is critical to a situation or one piece of data element is critical when another piece might not be. 

Data Dave
And that is a really interesting point that you just raised, is that, one person’s critical data element is not another person’s critical data. So an organization has tons of these critical data elements, but when you look at it with purpose and start to say, “What am I doing?” then it gets very narrow very quickly, and you start to focus.  A lot of people will say, “I’m going to do critical data elements. There can’t be more than 100 of them.” Awesome. Well, okay, that’s a little bit of a meaningless conversation, right? Because it’s kind of “Okay, you’re declaring there can’t be more than 100. If you got a very critical business, than there’s a lot of stuff. There can be more than 100 when you look at every process that you perform as a business.  

Now, you might say that for a sale or for customer onboarding or for employee onboarding or whatever it might be, there cannot be more than a handful, but you can’t necessarily put arbitrary limits on it. And you have to pay attention to the purpose because one person’s critical data element is not another person’s critical data element. 

Alexis
That’s great. Thank you, everyone, for listening to Talk Tech with Data Dave. We really appreciate it, and I hope you have a great day. 

Data Dave
Thank you. Thank you, Alexis and everybody, have a great day. Thank you.