Transforming IT Cost Management with FinOps and Unit Economics

In this episode of Business Reporter’s Best of British Business, we explore the power – and pitfalls – of data-driven decision-making with David Harding from MagicOrange.

 

MagicOrange, a leader in IT spend optimisation, believes every CIO should have full transparency in their IT budget, enabling smarter investment and divestment decisions. But can business efficiency truly be reduced to numbers? And how do we ensure data-driven insights are trustworthy?

 

David takes us on a journey through the origins of MagicOrange, and how its role in helping enterprises uncover the real cost of their technology investments has evolved over time. He explains why cost transparency beyond IT is critical, how unit economics can transform financial decision-making, and why AI is becoming the CFO’s smartest assistant.

 

If you’re looking to optimise IT spend and drive profitability, this is the episode you can’t afford to miss.

 

 

---

Podcast transcript

Rachel Hicks: Welcome to Business Reporter’s Best of British Business podcast, which challenges broadly accepted ideas. Today we discuss the optimism surrounding data-based decision-making with David Harding from Magic Orange. Magic Orange is one of the market-leading experts in IT spending optimisation, and they advocate that any C-suite can—and should—have full transparency of the IT budget and then make insightful decisions to invest in or divest from technologies.

 Can business efficiency truly be boiled down into raw numbers? Indeed, is there more to success than pure calculation? David seems to be the perfect candidate to explain whether decision-making—and the analytics behind it—are exact sciences, whether we can trust data on our spending, and what efficiency, transparency, and insightful investment really mean. David, welcome. How did you come up with the name Magic Orange?

David Harding: It’s a long story—a 20-year-old story. I was running a cost transparency project at the time in a large bank. Steve Norman, who was CIO at Merrill Lynch at the time, stood up in a conference of CIOs in the US and addressed the room: “I can’t be the only CIO here who works for a CFO who thinks his IT division is a magic orange.”

He went on to explain that every year his CFO would approach him and say, “Steve, tech just costs too much. We have to squeeze 5% out of the technology cost.” That resonated with me 20 years ago and, I think, still resonates with most CEOs today. It goes to the heart of the challenge—that most organisations, or the business side, don’t really understand the value they get from technology investments, even though that’s where much of their money goes.

Rachel Hicks: What does Magic Orange actually do? Tell us about it.

David Harding: We try to solve that very problem. We translate the cost of technology—joining the dots from an organisation’s general ledger and financial systems, which are designed for external reporting, not internal. We translate that financial data into the cost of a product or service to help the business truly understand the cost of its product lines, services, or applications.

Rachel Hicks: So, is what you do only important to the IT department, or do people outside IT care?

David Harding: It’s certainly not just relevant to IT. While the tech division is often where we start—because it’s the most complicated of all shared services and often carries the largest cost base—we also layer in other costs, such as operations. This allows us to calculate a much more comprehensive cost per transaction. That’s what we do for many of our clients. Most organisations understand the revenue derived from a given product or service line, but very few truly understand the real cost underpinning each of those products and services.

Rachel Hicks: So let’s look at how unit economics applies in the real world. Can you give us tangible examples of where this works?

David Harding: Unit economics is the cost per transaction—cost per trade, per customer, etc. When we layer in costs beyond IT, we can offer much deeper insights. For example, with one of our insurance clients, we included operations costs to help them understand what it actually costs to write a life policy or process a claim. That brings us to a new level of cost and financial analysis—something traditional finance systems just don’t offer.

Rachel Hicks: So how does that unlock better transparency and value for money?

David Harding: In two key ways. First, it helps organisations move beyond the “squeeze the orange” mentality—just cutting tech costs—and instead use the information to inform margin improvement and drive profitability. Second, it encourages broader business engagement, beyond the tech division, as part of a value chain.

We provide cost transparency and allocation within the tech division or CFO’s organisation, but if it’s not seen as part of the value chain driving accountability in the business, then it’s just interesting to accountants—and of limited use. Once the business understands that costs—such as compute usage—are driven by the applications or services they build and sell, they have no choice but to take responsibility for it. This, in turn, informs decision-making and can help improve margins and profitability.

Rachel Hicks: How hard are those conversations, getting people outside of IT to understand the relevance of strategic, holistic business planning with Magic Orange?

David Harding: If I flip the question, it’s harder when you’re allocating cost based on arbitrary metrics. If I allocate overhead based on headcount—which rarely reflects actual cost drivers—you won’t take responsibility for the spend. But if I allocate cloud compute costs based on your division’s actual usage, the conversation becomes easier. You recognise it as your own spend, and you take ownership.

Rachel Hicks: Why do we trust—or why should we trust—data?

David Harding: Data trust is about confidence in using it to inform decisions. That trust must be earned and validated. It can’t be taken on faith. For instance, incorrect date formats—like using American vs. European formats—can introduce significant errors. So, data must be validated.

Rachel Hicks: If we’re using data for decision-making, can fallible humans be trusted to determine what’s relevant?

David Harding: Great question. Human oversight is critical. Input validation must come first before trusting the data and the reports built on top of it.

Rachel Hicks: What if a client’s data isn’t “shipshape”? What do you say to them?

David Harding: I often say: the best time to start going to the gym was yesterday—if not, then today. Never tomorrow. Get started. Magic Orange shines a spotlight on your data. Many clients worry their data isn’t good enough, but improvement is iterative. You just need a general ledger or finance system to begin. Data quality improves over weeks and months—not years.

Rachel Hicks: AI is a hot topic. What’s your take on its role in Magic Orange?

David Harding: We see AI as your smartest, fastest financial analyst. When you’re asked to explain a cost change or variance, you should get the answer in seconds—not hours or days.

Rachel Hicks: But are we asking the right questions? How do we know?

David Harding: Excellent point. As I said earlier, answers must be validated. Over time, your confidence in the data and the insights will grow.

Rachel Hicks: You said we’d trust our own reports more than AI. Does AI add an extra validation stage?

David Harding: I’d challenge that. A manually built dataset might be more error-prone than an automated system. Automation and connectivity reduce the risk of human error.

Rachel Hicks: So what happens if different people ask the same question using different wording or languages?

David Harding: That’s where natural language interfaces matter. Our tool—Genie, powered by Databricks—lets you engage with data in natural language, not code. You no longer need to know Excel formulas. If I go on holiday, my deputy can ask the same question, and Genie will still deliver the right answer.

AI should not depend on who is asking the question. Whether you say “please” or use a different tone or language, the system should still understand you. That’s where we’re headed.

Rachel Hicks: How will the use of AI impact tech costs?

David Harding: That’s the multimillion-dollar question. Every boardroom is asking how AI can help—and what it costs. AI requires massive data and compute power. Should it run in the cloud? On-prem? A hybrid of both? Managing these costs is a major challenge.

Rachel Hicks: And what about the environmental cost of AI?

David Harding: That’s a real paradox. AI can help us solve climate challenges, but it also consumes vast energy and water resources. Finding that balance is a serious challenge.

Rachel Hicks: So how does Magic Orange use AI?

We use Genie for natural language queries. For example: “What was my client spend yesterday versus Monday last week?” You can dig deeper by resource type or compare year-on-year. We also use a voice assistant—Erica—who can send WhatsApp updates or call you with daily cost summaries. By the time you walk into the office, you’ll know where to focus. We don’t see competitors using AI to this extent.

Rachel Hicks: What would you say to CIOs or tech leaders who want this capability?

David Harding: Come find us—on LinkedIn, on our website, or connect with me directly. We’d love to show you what we can do.

Rachel Hicks: Finally, what do the next 5–10 years look like for the sector?

David Harding: The next 3–5 years are full of opportunity. The challenge of unit economics remains—what is the real cost per ticket, per trade, per claim? Shifting the conversation away from IT as a pure cost centre to something strategic is game-changing. Use solutions like ours to understand your cost base by product, service, or business line.

Rachel Hicks: David Harding, thank you.

David Harding: Thank you very much.

---