Data Value. OK. Hot topic. I’d be remiss if I didn’t say something about how data is the new oil or the new bacon or the new hot fudge lava cake with hand-churned vanilla bean ice cream drizzled with chocolate caramel sauce. Got it.
So, let’s see. I have this pile of corporate data on the table in front of me. What is its value?
Obviously, it has some value, otherwise the company wouldn’t go to all the bother and expense of managing it and making it available. But I can’t pour out an ounce of data and look up today’s spot price. And I can’t log into Amazon or eBay and check out recent data sales like you can baseball cards, leather jackets, and video game cartridges. I’m ready to put pencil to paper, but how do I even begin assessing value?
Well, a great place to start looking for the answer is Infonomics by Doug Laney. And I didn’t find just one answer, I found six, each with a different perspective on modeling Data Value.
The first three are Foundational Models that quantify information management discipline.
1. Intrinsic Value measures the data’s correctness, completeness, and exclusivity.
Correctness and completeness are really standing in for all of the Data Quality dimensions. I’ll have more to say about Data Quality shortly. Exclusivity captures the likelihood that another company has that same data. The less likely that a competitor would have the data, the higher the exclusivity. After all, some data is commonly available, while other data is uniquely captured, collected, calculated, derived, or generated by your company. Those are generally recognized as your crown data jewels.
2. Business Value measures the data’s relevance.
The more useful the data is for making business decisions, the higher the Business Value. This is easy if your company has a business process model that associates data, even if at the domain level, and business process. The more points of intersection, the higher the Business Value. Laney includes certain Data Quality dimensions as components of the Business Value calculation, but I would argue that instead those components should be replaced with the Inherent Value.
3. Performance Value measures the data’s impact on key business drivers.
First off, Performance Value is not a characteristic solely of the data itself, but rather of the data within the context of a specific performance indicator. Calculating Performance Value requires comparing key performance indicators with and without use of the data. Run the experiment. This puts a finer point on the Business Value, recognizing that just because two datasets touch the same business process, their impacts may differ.
The remaining three are Financial Models that evaluate the economic benefits of the data.
4. Cost Value measures the impact to the company of losing the data.
Cost Value has three components. The first is to imagine that the data becomes unavailable, perhaps due to catastrophe, error, or sabotage, and has to be reconstructed or recreated. This is the most feasibly calculated measure, and provides a good, conservative, minimum baseline. You know what data is backed up. You know how many data feeds and tables. And you can figure out how long it would take to collect everything again (or at least collect as much as you can). The second component is the impact to the business if the data becomes unavailable. You can leverage past Business Value Inventories to help here. This component typically dwarfs the first. The third is the damage to the company if the data were to be stolen. This component typically dwarfs the second. Just ask Target.
5. Market Value is the revenue the company could receive from selling the data.
I look at Data Value through the lens of Physics. Everything we’ve talked about so far is like potential energy. The data has potential value. Now we’re putting the data to work. We’re converting the potential energy to kinetic energy. Companies are increasingly looking for ways to monetize their data, including collecting and collating customer propensity data, address data, sentiment data, financial data, demographic data, and more.
6. Economic Value is the data’s contribution to the bottom line.
This is similar to Performance Value in that it compares revenue generated with and without the use of the data.
To get a complete view of the data, you have to consider all of the dimensions, and in the book, each is accompanied by equations, implementation details, and discussion. It’s awesome. I finally have what I want!
And then I realized that the absolute number isn’t the important part. Well, yes, if you are going to REALLY start incorporating data into the balance sheet, but most of us aren’t focused on that.
The important part is maximizing the value of your data. Let’s look at the six value measures in reverse order and see what we can do.
Economic and Market Value can be increased through creativity and initiative. Increase the value of your data by coming up with better uses for it. You’ve made a big investment in the data, and it will only impact your bottom line if you do something with it. Obviously, some data isn’t that valuable and there’s not going to be a market for much of your data, but ask yourself if you’ve considered all of the possibilities. Of course, some data may be too valuable to sell. That’s a decision for the business.
Interestingly, Cost Value is the one measure that companies should be looking to minimize. Sure, you want the business impact and theft value to other companies to be high, but those depend on ensuring that the data is always available and properly secured. Keep business continuity and security plans up to date so that the data is protected from theft and can be quickly restored in the event of a disaster.
When it comes to the Foundational Models, there’s not much you’re going to be able to increase Business and Performance Value. These are a function of the attributes, domains, and KPIs, and how they relate to your business. You can collect or create more useful data, but the impact is fixed or something like that.
That brings us to Intrinsic Value. And you can have a tremendous impact on Intrinsic Value. Put simply and very directly:
Data Quality determines the ceiling on all of your potential financial benefit.
To maximize any Data Value, first maximize its Intrinsic Value. Increase the value of your data by making it more usable. It only makes sense. You’re not going to be able to sell or to make good business decisions using bad data. It won’t matter how great your marketing people are if the data is faulty. Low Data Quality has low value. Data Quality will always limit Data Value.
Similarly for Data Understanding. It’s like you bought a luxury car with a whole bunch of features. But since you didn’t take the time or effort to understand those features, all you do is drive it from home to work and back. Maybe it’s a status symbol just having that car. But you’re not getting the value out of it. Similarly with data. You might have tons of data, but if you don’t understand it you’re not going to get the value out of it.
Identify the data with high Business Value and high Performance Value, and make that the focus of your Data Quality efforts.
The data with the greatest impact and relevance to the business needs to have the highest quality.
Use Data Value as a prioritization tool to get as much corporate benefit out of your data as possible. Returning to physics, increase the potential energy of the data so that more kinetic energy can be released.
1 Comment
Talam Thiam · October 30, 2024 at 9:07 am
I totally agree especially where you mentioned not understanding your date or not putting enough focus on the quality of the enterprise data is Just like purchasing a luxury car without taking the time to understand each feature.
I can personally testimony on your very interest in terms of data quality as in June 2002 at FedEx Services while being your summer internship student , my first project was data quality related within the Information Super Hub between FedEx domestic and international shipments.