Let’s start with a question: Why do we store data? Our obligations to govern our organisations forces us. We need to be able to answer to various stakeholders on what, why, how long questions. You need to know what is going on in your organisation to govern it and data is used for that for decades.

Currently also a lot of European legislation like GDPR, Data Act, Intellectual property issues force us to be transparent on what data we have, what we do with it and how it is used. There is more legislation on the way soon like the AI Act.

So, there is a need for data governance. Data models have been a part of data governance from the start.

A data model is a conceptual representation of data, including the rules and conventions for organizing and storing it. It is used to provide a common language and structure for working with data, and to facilitate communication and understanding among people who need to use or analyze the data. Data models are especially important in fields such as computer science and data analysis, where they provide a way to organize and make sense of complex or large-scale data sets.

And there is the main reason, current legislation like GDPR is based on context. So, if you do not know the context of your data, you cannot comply to this legislation.

Today most people are aware that there is value in data, and that combining data can add value. What most people don’t know is that the context of the data can change if you combine data. This risk is bigger if you don’t know in what context the data was collected.

A bonus is the role data and definitions of data can play in interactions between departments and/ or between companies or customers. Explaining the data and its limitations beforehand can prefent serious misunderstanding.

An example

When someone subscribes to your webshops newsletter, is it a client? Or does it become a client when the first purchase is done? And how long after the first purchase is the client demoted to a prospect again? One year, two, twenty? Rules and regulations state that you can only keep the data of an inactive client for a certain amount of time (it differs per region). Do you have this in your processes?

If you want to apply for an investment, investors want to know how many clients you have, not your prospects.

Different rules apply in different situations.

All my information is in the code

If you only know what data you have in your code, you do not have insights in your data assets and how they are used. This used to be a lesser problem, when data was used in specifics processes by people who worked at a company all their lives.

Now, in the age of combining data and job hopping, having a record on what data is there, in what context and how it is used is essential for companies to run.

Preferably in a mostly automated way.

That is where your data models become essential. Automation of data supply chains based on your data models will lift your maintenance and control of data in to the ‘possible’ realm. It still takes time and skill, but it can be done. If you maintain good data management, all current and coming legislation can be met.

If you don’t … be prepared to not only pay your data debt by being behind your competitors, but also a to pay steep fines for not being in control of one your most valuable asset: your data.

In follow up publications the Data Model Collective will elaborate more on our experience on how to get there. For example on techniques and tools like meta data management (MDM) that can be used to comply to the Environmental Social Governance (ESG) policies.