demand forecasting 1024x576 1

How Can SMEs Overcome the Myths of Demand Forecasting?

First published in May 2023. Updated November 2026.

Demand forecasting for small and mid-sized manufacturers

Every business owner wishes they had a crystal ball to see changes in customer preference, industry growth and new product success. No such crystal ball exists, but you still need to plan. That is where demand forecasting comes in.

Demand forecasting is a familiar supply chain tool, often sold with an alluring promise:

“What if we could predict what our customers will order far enough in advance to obtain materials and manufacture the goods in time? Then delivering on time and in full would be easy, because the goods would already be available when the order arrived.”

Companies spend heavily on software, computing power and specialists trying to fulfil this promise. Yet most small and mid-sized enterprises (SMEs) in India do not run an effective forecasting process. Four common myths hold them back, and the result is usually very high inventory alongside poor customer service levels.

Myth 1: “The forecast is always wrong, so why do it?”

The forecast will be wrong, but there are degrees of error, and a consistent, well-thought-out forecast is still the best available view of future demand. Many factors can move demand that no model captures, such as shifts in customer behaviour or unexpected events like a pandemic. Choosing three parameters to suit your business goes a long way towards keeping forecast error manageable:

  • Length (horizon): the further ahead you forecast, the less accurate the forecast will be.
  • Time interval: longer buckets are more accurate. Monthly demand is much easier to forecast than weekly or daily demand.
  • Level of detail: forecasting every individual SKU accurately is much harder than forecasting product families or value streams.

Myth 2: “Our business can’t be forecast, and forecasting is only for large companies”

Demand forecasting matters for businesses of every size. Small businesses use it to plan inventory, production schedules and marketing.

If you run a pure make-to-order business whose customers accept long lead times and sometimes uncompetitive prices, you may not need to forecast. Most businesses do not have that luxury. At a minimum they need a rough-cut forecast to secure better material prices and delivery performance.

Even in a semi make-to-order environment, much of demand planning can be forecast. It usually works best to forecast at the level of a whole value stream. That gives the required production rate for each value stream and allows capacity, raw material and labour to be planned through the year.

Myth 3: “Forecasting needs statistics experts and expensive software”

Forecasting can be complex, but plenty of tools and resources help non-experts get the basics right, and many software packages have simple interfaces for entering data and generating forecasts.

When choosing forecasting software, pick a system robust enough for the volume of data you will enter. The best results come when technology is fitted to the needs of the people using it, rather than people having to fit the technology.

Number crunching matters, but what makes or breaks a forecast is how well you know your customers and your market. That needs a sales team that talks to customers and shares what it learns honestly with the rest of the company.

Myth 4: “Collaborative forecasting is more pain than it’s worth”

Research consistently shows that the best forecasts combine a statistical baseline, built from time-series history, with a review in which several stakeholders can adjust it.

The hard part is usually getting stakeholders such as Sales to take on the discipline and accountability of producing a company-wide forecast at a level of detail planners can use.

Ideally, the monthly forecast review is part of your sales and operations planning (S&OP) process. Each month, review and adjust the forecast, roll a new month onto the end of the 12-month horizon, and adjust for known, significant changes in demand.

How SMEs can put demand forecasting into practice

  1. Collect historical sales data to understand demand patterns and identify trends, seasonality and other drivers.
  2. Analyse market trends, including changes in customer behaviour and the competitive landscape.
  3. Choose a forecasting method that suits your business and capability, such as time-series analysis, regression, or AI and machine-learning models.
  4. Choose the right tool: an Excel template, a cloud platform or an AI-based tool, depending on your data and team.
  5. Decide the horizon, time interval and level of detail for the forecast.
  6. Validate the forecast against actual sales and adjust.
  7. Translate the forecast into raw material requirements for short and long lead-time suppliers, and into internal capacity plans, with enough safety margin to absorb forecast error.
  8. Review the forecast every month as part of S&OP.

A good forecast also changes how much stock you need. See how forecast error drives safety stock in our guide to ABC-VED analysis for critical parts, and why every extra day of cover costs more than you think in the real price of “just in case” stock.

Frequently asked questions

Is demand forecasting worth it for a small manufacturer?

Yes. Even a simple monthly forecast at product-family or value-stream level helps you buy materials earlier at better prices, plan capacity and labour, and hold less stock while improving delivery.

How accurate does a forecast need to be?

It never needs to be perfect. It needs to be consistent, reviewed every month and backed by enough safety stock or capacity to absorb the expected error.

Which software should an SME start with?

Start with what fits your data and team. Many SMEs begin with a structured Excel model and move to a cloud or AI-based tool once the monthly review process is working.

At 7th Mile Shift, we help mid-sized manufacturers set up practical demand planning and S&OP routines. See our inventory management services or talk to us about your forecasting process.