Whenever we try to create something new or work in an existing environment, we make assumptions about data: you hand in a receipt for your expense report because that is data relevant to this process, when you plan your workload in your ERP system you give it certain data points that are relevant to this process, when you create a new product you make certain assumptions about how customers will use it and what data becomes relevant at which point.

What's interesting is that few organizations seem to make those assumptions explicit. In contrast to Leonardo da Vinci, who created a form of data model for humans - the Vitruvian Man shown above - in order to make his paintings and sculptures more realistic, when we look at IT solutions in the organizational context, the job of defining the data model often falls on the vendor of the solution, not the customer who would best be suited for this.

Combine this with a trend that obsesses over using certain tools doing one thing particularly well instead of holistic solutions and you get a modern workplace that looks like the screenshot below to many employees.

On top of a fragmented tool landscape, you often have a fragmented data landscape which means that organizations have a bunch of siloed information they can hardly make any sense of, something they discover at the latest once management wants to embark on an AI project and the reason for the increasing demand for companies like Palantir that claim to help organizations with ontologies, i.e. descriptions of what data is actually available.

Both developments are a nuisance to organizations and both developments stem from the fact that all too often organizations do not think about their data models. What is it that we are actually doing? What information do we need at what point in time for whom? Where is information coming from and where is it going?

These might sound like boring questions that take time and effort to answer. This might be the reason many organizations do not actively want to think about those questions. But it would be worth the hassle.

If organizations spent some time properly figuring out their data models, it would be a fast-track to not only better processes but also make the onboarding and everyday lives of employees easier. It is also a factor in better IT procurement and better digital sovereignty. Knowing your data models is an integral part of knowing your tech stack, hence enabling you to no longer fit your organization to the available data models of solution providers but the other way round. You also don't need somebody else to explain your data to you, becoming less reliant on sharing your "crown jewels" with third parties.

So, when prioritizing IT projects for the coming year, might I suggest paying particular attention to the question: what are our data models? This would be a sound investment in a sustainable digital transformation of the organization, increasing understanding and alignment across the organization, identifying duplicates and blind spots.

Subscribe to this newsletter and follow me on social media for more publications and thoughts:
LinkedIn
Mastodon