Data Science Foundations Chapter 2: Who Are Your Stakeholders and Why They Matter

You built a model. It works. The numbers look great. But nobody uses it.

Because you forgot the people around it. Chapter 2 of “Data Science Foundations” by Stephen Mariadas and Ian Huke is about exactly that. Stakeholders. The humans who care about your project, fund it, get affected by it, or can shut it down.

After 20+ years in IT, I can tell you: ignoring stakeholders kills more projects than bad code ever will.

Not All Stakeholders Are Equal

Here’s the thing. When you start a data science project, many people have a stake in it. Some contribute. Some benefit. Some get affected without ever asking. But they do not all matter in the same way.

The authors use a great example. Imagine your company wants to predict which employee phones will break in the next six months. The phone manager cares a lot. They have power over the project. But a regular employee who just uses the phone? Almost zero influence.

And what about former employees whose phone usage data sits in the analysis? They left the company. But their data is being used. They are stakeholders too.

Your stakeholder list is always wider than you think.

The Stakeholder Wheel

Mariadas and Huke introduce the “stakeholder wheel.” Put your project in the center, then map categories around it: partners, suppliers, regulators, employees, managers, owners, competitors, customers, data subjects, and society.

The book uses the classic smoking and cancer research case. A tobacco company commissioning that analysis has stakeholders far beyond the research team. Tobacco farmers, production workers, directors, shareholders, health boards, smokers, potential smokers, health providers. All have a stake.

And it spreads further. Insurance companies, taxpayers funding healthcare, politicians. The ripple effect can be huge.

The authors mention a fun example. The team that built dynamic pricing for concert tickets probably did not expect the UK prime minister to comment on their work. But that happened after the Oasis 2025 tour pricing blew up publicly. Your stakeholders can surprise you.

How to Actually Find Your Stakeholders

Knowing that stakeholders exist is one thing. Finding them is another. The book suggests two methods: talk to people and check documents.

Start with the org chart. Find departments that might be affected. Then talk to colleagues. Do your homework first. Nobody likes answering questions from someone who did zero preparation.

As you talk to more people, the network expands. Each person points to another. And remember: someone who does not care about your project early on might become very interested later. Stakeholders shift over time.

For external stakeholders, the authors recommend PESTLE analysis. It covers six areas: Political, Economic, Socio-cultural, Technological, Legal, and Environmental. Running your project through these lenses helps spot stakeholders you might otherwise miss.

The Power-Interest Grid

So you have a list of stakeholders. Now what? You cannot treat them all the same. That is where the power-interest grid comes in. It is a simple 2x2 matrix.

On one axis you have power (how much influence they have over your project). On the other, interest (how much they care about it). This gives you four strategies:

Keep close (high power, high interest): These are your sponsors, the people funding and driving the project. Stay in regular contact. Share updates. Listen to their concerns.

Keep onside (high power, low interest): Think regulators. They do not care about your project day-to-day. But cross a line and they will come down hard. Make sure they have no reason to worry.

Keep informed (low power, high interest): These people care about the outcome but cannot change it much. Give them enough information so they feel included.

Keep track (low power, low interest): Monitor them. Their position might change as the project moves forward. Employees who did not care at the start might care a lot when they learn their favorite phone model is getting replaced.

Build a Communication Plan

The power-interest grid feeds directly into a communication plan. Who gets what information, how often, through which channel.

For the phone project, the sponsor gets regular one-to-one meetings. The compliance team gets periodic updates. Employees get informed at appropriate stages.

It does not have to be complicated. Even a regular catch-up with your manager counts. The point is to be intentional about it.

One More Thing

Different stakeholders have different levels of technical understanding. The CFO does not need to see your model parameters. The data engineer does not need a business case summary. Adapt your language. Meet people where they are.

This sounds obvious. But I have seen senior data scientists present confusion matrices to boardrooms and wonder why nobody approved their project.

Key Takeaways

• Identify stakeholders early and revisit the list throughout the project
• Use the stakeholder wheel to make sure you are not missing anyone
• Use PESTLE analysis for external stakeholders
• Map stakeholders on a power-interest grid to decide your engagement strategy
• Build a communication plan, even a simple one
• Adjust your language to your audience

Stakeholder management is not glamorous work. But it is the difference between a project that ships and one that sits in a Jupyter notebook forever.

Previous: Chapter 1: Introduction Next: Chapter 3: Project Delivery