Many distributors still rely on legacy sales and demand forecasting methods, leveraging a combination of spreadsheets, ERP-based forecasting and simple statistical models. These methods limit distributors and lead to inefficiencies, increased costs and lost opportunities.
Here’s why:
Static, History-Based Sales/Demand Forecasting Models are Inadequate
Many distributors still apply a number based on an expectation they’ll have linear growth across the board.
“We sold 100 widgets last year, so we expect 110 this year based on a 10% increase.”
But this approach:
- Oversimplifies trends and ignores real-world disruptions.
- Can’t predict economic downturns, competitor actions and shifts in customer behavior on its own.
- Ignores randomness; every forecast will have an element of unpredictability that can’t be projected based on last year’s sales alone.
Legacy forecasting also assumes slower, more gradual changes. By the time distributors realize demand is going up rather than down like they expected, it is often too late.
Static Sales/Demand Forecasting Doesn’t Incorporate Uncertainty
Plans and forecasts can’t be hard-and-fast. They need to account for risk and a range of possible outcomes. The problem: Static models can’t “see” events or shifts that haven’t happened yet – the definition of uncertainty. They assume past patterns will repeat, when today’s market is anything but predictable. Static models also are not updated in real time, lagging behind actual change. This means businesses are reacting, not proactively adjusting to uncertainty.
Traditional Sales/Demand Forecasting Methods Don’t Account for External Market Considerations
They focus only on internal structured data from their ERP and other systems. Most legacy forecasting methods do not factor in external conditions, such as:
- Macroeconomic conditions such as interest rates and unemployment trends
- Regulatory and trade policies, such as tariffs, environmental regulations or import restrictions
- Industry-specific indicators, such as housing starts, consumer spending and automotive sales
- Competitive factors, such as new market entrants
- Weather and seasonality, such as unusual weather that may not align with past seasonal trends
For example, a building products distributor we worked with discovered that demand for their product correlated with light-duty truck sales – which was unrelated to their internal sales data. This was uncovered when they analyzed more than 4,000 potential external factors and their influence on sales. Narrow forecasting methods often miss these connections, which can lead distributors to misallocate inventory, overstock or suffer stockouts.
Legacy Tools Can’t Measure the Extent Factors Will Have an Impact on the Forecast
Even if a model acknowledges external factors, it may fail to quantify the impact of those factors on demand. For example, just because it’s snowing does not mean every household will immediately rush out to buy shovels. The extent of the snow’s impact will be determined by other factors, such as:
- How many customers already own a shovel
- Customers that prefer a snowblower to a shovel
- Customers who hire someone else to do the work rather than buy their own shovel
Forecast Models Often Put too Much Weight on Sales Reps’ Input
Distributors frequently rely on sales reps’ input to inform their product-level forecasts. Sales forecasts from the field are an important data point and should be considered in the big picture. After all, these reps are having real conversations with customers. But they must be balanced with other internal and external factors.
Distributors must also account for potential inconsistency in these forecasts. A sales rep may lean very optimistic or extremely conservative. Some are also more reliable than others.
Let’s say a sales rep predicts a 20% increase in orders from a major customer based on past conversations. The distributor increases their inventory to meet the expected demand, but the customer ends up ordering 10% less than last year because of budget cuts. The distributor ends up with surplus inventory that takes months to clear.
The Costs of Static Sales/Demand Forecasting Models
Excess Inventory Costs
Everyday a product sits unsold in a warehouse, it incurs carrying costs, including:
- Warehousing costs, such as rent, utilities and storage.
- Financing costs
- Depreciation and obsolescence, as products lose value over time. This matters more in some markets than others, such as seasonal goods and electronics.
Let’s say a fasteners distributor forecasts that sales will increase 5% based on historical data. Then, a new competitor enters the market and takes away market share. Static forecasting fails to adjust for this, which results in the distributor over-ordering inventory. As sales fall, they’re stuck with that slow-moving stock tying up space and capital.
Another example: A building products distributor orders too much lumber based on outdated demand forecasts. When the market shifts, and new projects slow:
- The inventory sits unsold while storage costs increase.
- The distributor pays interest on the unsold inventory.
- Lumber prices drop, so they are forced to sell below cost, losing money on every unit.
Lost Revenue
When distributors forecast at a category level, they may see overall growth in a category but fail to identify which SKUs are driving that demand. For example, a distributor may forecast an increase in demand for hand tools.
But if screwdriver sales are expected to go up – while wrench sales fall – the distributor will end up with too many wrenches and not enough screwdrivers. That means they’ve lost sales on the higher-demand screwdrivers due to stockouts and wasted capital on the wrenches due to excess inventory.
Labor and Productivity Losses
Manual spreadsheet-based forecasting – which requires distributors to pull data from different sources, apply adjustments and update projections regularly – can be time-consuming and error-prone. This cost in time is often overlooked; it can take weeks to get a single forecast figure. Other ways time is wasted include:
- Warehouse teams dealing with last-minute inventory shifts
- Executives wasting time in constant review cycles, wondering whether to adjust their forecasts
- Procurement teams making last-minute corrections, placing emergency orders or negotiating with suppliers
Inability to Forecast at Scale as a Distributor Grows
Many distributors lack the manpower to handle forecasting at the scale required for effective inventory management. With thousands (or even millions) of SKUs, it’s impossible to manually manage demand forecasting without making errors. This means that as a company grows, inefficiencies will multiply.
What’s the Solution?
“Good enough” sales/demand forecasting is not good enough. Distributors that continue to rely on static forecasting will struggle with inefficiencies and higher costs. Distributors need AI and machine learning to analyze and incorporate real-time external factors to identify correlations.
Machine learning and advanced analytics can also analyze and present potential outcomes under different conditions to account for unpredictability. That allows distributors to maximize profits while keeping on eye on risk associated with supplier delays, geopolitical events, weather and economic shifts.
Distributors now have the ability to make dynamic adjustments when markets change – refining forecasts based on new data, rather than using fixed assumptions.
For example: Before COVID, a decking, trim and home renovation products distributor found that interest rates correlated negatively with demand. In other words, as interest rates increased, demand decreased – and vice versa. Post-COVID, that flipped. When interest rates went up, demand for their products went up. People were no longer taking out home equity loans for renovations. AI detected this in the data, but traditional tools could not.
Are your forecasting tools costing you money? It’s time to move beyond outdated methods and embrace intelligent forecasting.
