The Data Maturity Model

Why invest in data?

At the end of the day, good data management usually comes down to one thing: making the best decisions possible. Data is a key part of living in the information age. Increasingly, workers and companies both must be knowledge workers and knowledge companies. Roles where no judgement is called for become automated.

However, there’s a wide gulf between having data and being able to make wise, data-driven decisions about the future. This is especially true when you are dealing with multiple, interconnected internal departments, a changing economic and cultural landscape, and the impacts of the changing actions of customers and competitors.

There’s a lot of urgency today around finding a way to become data-driven, and fast. And some companies we talk to want to jump straight from simply storing data to having a fully functioning predictive model that can make customized predictions about individual users. While it’s certainly possible for organizations to reach this level of sophistication, they have to go through multiple steps to grow their data maturity first. In this blog post, we will cover the different levels of organizational maturity when it comes to working with data, and the steps it takes to move from one level to the next.

Level 1: Monitoring

The first level of the data management maturity model is Monitoring. At this level, an organization is working primarily with raw data, with no applications or tools to assist with processing or analyzing it. Raw data is stored in tables in a database.  In order to have this level of maturity, the technology you need is a database.  

The primary activity that you can perform with data at this level of maturity is monitoring. You can see the data being produced, but without a tool to process the data, any kind of calculations you want to perform will have to be done by hand.  So while you can see all the data that is available to you to learn from, seeing and observing is about all you can do. If you have even a thousand rows of data (not that many for a database), performing any kind of calculation to understand the data requires a massive investment of manual effort.

To move from the first level to the second, an organization needs to make an investment in software tools, and the training and skills to operate them, in order to process the data and perform basic analysis. You will need a reporting tool. These can range widely in price and complexity, from Microsoft Excel to the open-source Metabase, to advanced Business Intelligence software like Looker.

Questions you can answer at this level look like this: How much money was spent in this particular transaction? What device did a user use to login for this particular session? How long did this user session last?

Level 2: Organizing

At this level of maturity, an organization is able to transform its raw data into information. The organization can organize its data into calculations and visualizations, such as sums and averages, as well as charts that can track change over time.

Information organized at this level can answer questions like, “Who”, “What”, “Where” and “When?” The kinds of questions you can answer with this information are: “Where are customers making transactions?”; “At what time of day do most people log into the program?”; “Which users purchased this bonus feature?”

When an organization is working at this level, most of the company is not interacting with raw data, though there may be some people in engineering that do. Decision makers are going to be primarily working with calculations and visualizations rather than tables. 

Being able to organize data and visualize it as information is a big step up for organizations. Companies can get a lot of value out of being able to view and understand customer behavior or market movement or weather patterns.

The data maturity of an organization also depends on the maturity of all of its departments. Most organizations spend a fair amount of time in this level of maturity, because it takes time to bring all of their departments up to speed and used to driving their decisions using information.

This maturity level becomes limiting when an organization wants to move up the maturity level and be able to detect patterns and trends. This can be especially tricky, as human beings are pattern-makers and may see and interpret trends where there are none, as we’ll see in the next level of data maturity.

Level 3: Analyzing

At the third level of data maturity, an organization is able to analyze relationships between factors that affect their business in a way that is scientifically sound.

Moving from the second to the third level of data maturity usually requires investing in a contractor, consultant or employee. The same software and technology tools are used between maturity levels two and three. However, while you only need basic math and analytical skills to organize data, in order to perform verifiable statistical analysis, you need a deep background in mathematics and statistics.

At some point you may want to ask questions like “What is the relationship between time spent on my application and purchases made in-app?”

In order to answer this question accurately, you need both a statistical model, which is easy to produce in BI software, and an understanding of how to apply statistical models, which requires a deep understanding of math. The work involves using linear equations and models, as well as creating hypotheses to test.

Data Analytics is a big investment for most organizations, as it requires investing in staff instead of technology. This means longer times for onboarding and getting the team used to working together. Analysis is also labor-intensive for knowledge workers, and takes time because it involves developing and testing hypotheses to answer insightful questions. Because of this process, each analytical question answered (especially for a large organization) can be an investment of months.

The questions that require statistical analysis are ones for which there is a high yield or return if the organization knows the right answer. Some organizations will never reach this stage of maturity simply because (at least with the technology available in 2020) it requires significant scale to be profitable.

Stage 4: Predicting

The final stage of data maturity is prediction. This is the only stage where data science and machine learning comes in.

At this stage of data maturity, you have monitored your data, organized it to gain information, and analyzed it to create statistical models of relationships between factors that affect your business. You have used data to understand your business and drive many, many decisions. At this stage, you are trying to inject all of that understanding into an algorithm that will be able to make many, many decisions based on your understanding, and intake new information based on your rules to improve its performance.

Predicting answers these types of questions: “Given what we know about user ABC and their last four interactions on the platform, and the rules for how we know users like ABC behave, what is the next advertisement we should show user ABC to generate the greatest likelihood of engagement?”

This is a very advanced use of data that requires large investments in both technology and staff. An organization needs to go through all the previous stages or it won’t have the insight and statistical models it needs to teach an algorithm how to behave. Data Science is a new and specialized field, and it is very rare for a data analyst or engineer to also be a data scientist, so you will also need to bring on new staff.

Predictive modeling also requires massive scaling to be a meaningful investment. There needs to be enough incremental benefit that if you teach a machine to make many decisions just a little better, then the sum of those improved decisions will reach a profit that outweighs the large investment to create it.  As more and more organizations become global and deal with data at scale, and as more technology to support machine learning becomes available and cheaper, this will be an easier investment to make.

Komodo’s team has experience advance their capabilities as each level of maturity, and and has helped organizations jump to the next level. The key to working with a data consultancy is finding one who can meet your organization where you are and that can help your company advance from where you are now, not where you aim to be in the next year.

To have a conversation about leveling up your company’s data capabilities, contact us today.

Read more of our thoughts on the Data Maturity Framework here!