Why Demand Forecasting in FMCG Is Never Just About Numbers
A CIO’s reflections from the inside
If there is one topic that has consistently sparked animated discussion in every boardroom, supply chain review, and CIO peer forum I’ve been part of, it is demand forecasting. Not because people don’t understand its importance—but because almost everyone has, at some point, felt the pain of getting it wrong.
Early in my CIO journey, I remember sitting in what seemed like a routine review meeting. The numbers on the screen looked precise, confident, and reassuring. Accuracy percentages were healthy, charts were clean, and there was a quiet sense of comfort in the room.
A few weeks later, that comfort disappeared. In one of the regions, some states were sitting on piles of slow-moving inventory, while another set of states was scrambling to meet demand. The factors of seasonality, weather, impact of state excise tax on pricing, cost and demand such factors require delicate manoeuvring and compiling of huge data sets.
The forecast hadn’t failed because the math was wrong. It failed because it didn’t reflect how the business actually behaved amid a small blip and all factors affecting the impact of demand forecasting were not considered in the forecast.
Over the years—through many conversations with planners, commercial heads, supply chain leaders, and fellow CIOs across Industry—I’ve come to believe something quite strongly: demand forecasting is not really a forecasting problem. It is a decision-making and cultural mindset problem.
This is not a theoretical piece. It is shaped by lived experience—my own and those of peers I’ve learned from along the way. I want to talk candidly about why forecasting is so hard in some industries and in some it is easy and where it usually breaks down, and how CIOs can approach demand planning in a way that genuinely improves business outcomes.
The Inherent Complexity of FMCG Demand Forecasting
FMCG demand lives at a complicated intersection of consumer behaviour, channel dynamics, and operational constraints. Unlike capital goods or industrial supply chains, FMCG deals with:
Short and shrinking product lifecycles
Rapid SKU proliferation
Heavy and frequent promotional activity and somewhat veiled promotions in Alcobev industry
Strong seasonality and regional variation
Consumers whose preferences shift faster than systems can adapt
In theory, all of this is well understood. In practice, it is rarely captured coherently.
One of the most common mistakes I’ve seen is the assumption that better historical data will automatically solve forecasting challenges. Historical data is necessary—but it is not sufficient. FMCG demand is rarely stationary. What sold well last year may behave very differently this year, even in the absence of obvious disruption.
Long before COVID, I had already seen how fragile demand forecasts could become when faced with sudden discontinuities—state-level regulatory changes, abrupt route-to-market shifts, pricing interventions, or channel disruptions. Carefully constructed models became irrelevant almost overnight.
Those experiences shaped my thinking: in FMCG, disruption is not an exception. It is the default acceptance.
The pandemic, observed later through discussions with CIO peers across the industry, simply amplified this truth. What it exposed was not just a limitation of algorithms, but a deeper over-reliance on historical stability. For me, it reinforced a belief —demand forecasting must be designed to adapt rapidly changing behaviour when history stops being a reliable guide and a cultural mindset change - demand forecasting is done only once a month,
Where Forecasting Most Often Breaks Down
1. Forecasts Built in Silos
In many organizations, demand forecasts are still created in isolation. Sales has one view. Marketing has another. Supply chain quietly maintains a third version “just to be safe.” Finance looks at all of this and asks the most uncomfortable question: Which number do we trust?
When forecasts are owned by a single function, they inevitably reflect that function’s biases—sales optimism, supply conservatism, or finance’s inventory concerns all find their way into the numbers.
Over time, I’ve learned how CIO’s role is an enabler to the entire demand forecasting process, is to enable a shared version of truth— but one version of data - thru a single platform where everyone from Sales Representative to Sales Director, Marketing to Supply Chain and Finance to Logistics contributes in parallel and one platform that brings all perspectives together without allowing any single one to dominate unchecked.
2. Treating Forecasting as a Monthly Ritual
Another recurring pattern is treating forecasting as a calendar event rather than a living capability. Forecasts are created once a month, reviewed once, approved—and then largely forgotten.
This cadence in today’s rapidly changing demand scenario is often too slow. Demand forecasting signals to change weekly, sometimes daily. Promotions get extended, distributors alter ordering patterns, weather influences consumption, and competitors act unpredictably.
A static forecast in a dynamic environment creates false confidence. One of the most important shifts CIOs can enable is moving organizations toward continuous demand sensing, where forecasts evolve, as new signals emerge instead of being frozen in time.
3. Overconfidence in Tools, Underinvestment in Thinking
Modern demand planning tools are powerful. I’ve seen that firsthand. But I’ve also seen organizations assume that the tool itself would “fix” forecasting—this is the perfect recipe for failure. The failure is not because of the tool, but because the decisions, the thinking and behaviour to adopt continuous forecasting methodology is never adopted..
Technology amplifies behaviour; it does not replace judgment.
When planners blindly accept system-generated numbers, errors propagate faster. When every forecast is overridden without discipline, the system never learns. The balance lies in augmented decision-making—where technology provides insight, and humans provide context.
Reframing Forecasting as a Strategic Capability
One of the most useful mindset shifts I’ve observed is when organizations stop asking, “How accurate is our forecast?” and start asking, “How well does our forecast help us make better decisions?”
Accuracy matters—but forecast value-add matters more.
A forecast that is slightly less accurate, but allows earlier intervention, better inventory positioning, or faster response to demand shifts is often more valuable than a technically superior model that arrives too late to act upon.
From a CIO’s perspective, this reframing changes everything—from how tools are selected to how success is measured.
Why Demand Forecasting Must Be a Joint CIO–Supply Chain Decision
One important lesson I’ve learned—sometimes the hard way—is that demand forecasting should never be seen as a purely supply chain initiative. Nor should it sit entirely with the CIO as a technology-driven transformation.
In reality, forecasting succeeds or fails in the space between these two roles.
Supply chain leaders bring something no system can infer on its own: a deep understanding of real demand drivers, market behaviour, promotional dynamics, distributor realities, and where execution typically breaks on the ground. Without this context, even the most advanced model remains theoretical.
CIOs, on the other hand, need to ensure that the forecasting solution and the tool rests on a credible and clean data foundation with historical inferences and current repositories, integrates with the enterprise landscape, scales with the business, and remains sustainable.
When these two perspectives come together early—during solution evaluation and design—the organization is far more likely to adopt a platform that is not only analytically sound, but also operationally usable. Planners trust it. Sales engages with it. Leadership listens to it.
I’ve also seen the opposite. Solutions chosen in isolation—whether driven solely by supply chain requirements or by technology ambition—often struggle to deliver long-term value, despite strong intent and investment. The gap is rarely technical, but rather CIOs should also be knowledgeable to understand the theoretical aspect of forecasting as it will aid in better partnership with the various tenants and users of the solutions.
What CIOs Should Focus On
Start with Decisions, Not Data
Before selecting tools or designing models, I encourage CIOs to ask:
What decisions will this forecast influence?
Who will act on it?
How quickly must those decisions be made?
Clarity here prevents over-engineering and ensures the forecasting system aligns with business reality.
Design for Volatility, Not Stability
Forecasting models that assume smooth patterns fail spectacularly. Systems must be designed to detect early deviations, incorporate external signals, and allow rapid scenario evaluation.
In my experience, the most effective planners were not those with the most complex models, but those who could quickly ask, “What if this assumption is wrong?”
Balance Granularity with Actionability
More detail is not always better. Forecasting every SKU at every location may look impressive, but if planners cannot act meaningfully at that level, it becomes noise.
Successful organizations deliberately operate at multiple levels—strategic, tactical, and operational—based on where decisions truly need to be made.
Tool Selection: What Really Matters
CIOs are often asked to evaluate demand planning tools as technology projects. In reality, they are organizational transformation projects.
From experience, success depends less on features and more on:
Ease of adoption
Seamless integration
Transparency of logic
Flexibility as the business evolves
The ability to learn from outcomes
What problem this solution is going to solve for my organisation to bring efficiency
The CIO’s Evolving Role
Today, the CIO has evolved far beyond the role of systems integrator; they are becoming a change agent in the transformational journey. In demand forecasting, they serve as the integrated pillar across functions, the sole of data integrity, and the catalyst for insightful conversations.
The real success of forecasting is the outcome of lower stockouts, lower costs of inventory and improved P&L and real time changes in forecasting impacting the manufacturing line usage. in meeting rooms—when discussions shift from “Whose number is right?” to “What are we seeing, and how should we respond?”
That shift is cultural. Technology is merely enabling it!.
Closing Reflections
After decades in FMCG and many conversations with peers, one thing is clear: there will never be a perfect forecast.
But there can be better decisions, faster responses, and fewer surprises.
Demand forecasting is not about predicting the future. It is about preparing the organization to adapt when the future refuses to behave as expected.
For CIOs embarking on this journey, my advice is simple:
Embrace the inherent complexity without oversimplifying it.
Invest in people and culture change as much as platforms.
Design systems that thrive amid disruption.
And always remember: forecasting fuels decisions, not just spreadsheets.
