Why sales history isn’t enough: old forecasting methods no longer work

But today, that’s no longer enough. Customer behaviour has become more complex, and competition more intense. Promotions, seasonality, trends, external events, and even the weather can significantly alter the landscape. Ignoring these factors leads to problems: products are either unavailable when needed or end up sitting in the warehouse, tying up capital.

It’s time to move beyond simple templates. Let’s explore how to forecast the right way!

In this article, we’ll look at:

• Why relying solely on sales history leads to mistakes
• Which factors should be considered in forecasting
• How artificial intelligence helps retailers make more accurate predictions
• How forecast automation improves order management

Demand forecasting is one of the key tasks in retail. The accuracy of forecasts affects stock levels, product availability, sales volume, and customer satisfaction. For a long time, sales history formed the basis of calculations: if a product sold well in the past, it was assumed more should be ordered. If demand dropped, purchases were reduced accordingly.

Imagine this: a shop analyses last October’s umbrella sales and sees a spike in demand. Based on this, the system orders more for the current year. But the spike was due to an unusually rainy season that doesn’t repeat annually. The forecast turns out to be wrong, and the store ends up with excess stock.

Many retailers still rely on sales history, expecting that past performance will help predict future demand.
But in today’s environment, that’s no longer enough. Historical data shows what was sold, but not why it was sold—or why some products were left untouched on the shelves.

Relying solely on past sales means taking a gamble. To forecast accurately, it’s essential to analyse not just the numbers, but the context in which those numbers occurred. In the next section, we’ll explore which factors influence demand and how to take them into account when planning orders.

An accurate forecast takes into account not only past sales but also the wider context. This includes seasonality, marketing activity, product availability, and even local events, such as sports matches or festivals near the store. Without this information, sales history becomes not a tool for analysis, but just raw statistics.

Historical data without consideration of external factors can lead to mistakes:

• Seasonal spikes create false expectations. For example, chocolate sells well in February before Valentine’s Day, but that doesn’t mean demand will remain high in March.
• Promotions and discounts distort the picture. If a product sold well during a sale, that doesn’t guarantee steady sales afterwards.
• Out-of-stock doesn’t mean lack of demand. If a product sold out, the drop in sales doesn’t mean customers lost interest—it just wasn’t available.

The following key factors influence demand:

Demand forecasting is not just about analysing past sales.
Customer behaviour is shaped by a wide range of factors, and ignoring them can lead to errors in ordering and inventory levels.

If sales were low last month, that doesn’t necessarily mean demand was weak. It’s possible the product was simply missing from shelves or the online catalogue. Ignoring availability data leads to incorrect conclusions.

A price reduction or inclusion in a marketing campaign can temporarily boost demand. But if the promotion ends and the system continues to order large volumes, this leads to overstocked warehouses and excess inventory.

Some products only sell during specific periods—for example, sunglasses in June, winter clothing in December, or school supplies in August. Basic algorithms may average out the data, but without factoring in seasonality, they won’t produce accurate forecasts.

Sporting events, festivals, changes in weather, and even political developments can influence demand. For example, sales of televisions and satellite dishes often surge ahead of the football World Cup.

The same product can sell differently depending on the region. In southern cities, demand for winter clothing is lower than in northern ones, while ready-made meals are more popular in business districts than in residential areas.

To ensure forecasts are accurate, retailers must consider more than just past sales—they need to factor in product availability, the impact of discounts, seasonal shifts, local specifics, and external influences. In the next section, we’ll look at how to collect and analyse data to forecast demand as accurately as possible.

If you rely solely on sales history and ignore these factors, forecasting becomes little more than guesswork. That’s why modern analytics systems use not only past data, but also information on current trends, external events, and customer behaviour.

To make accurate forecasts, it's important to consider:

Demand forecasting requires not only the right algorithms but also high-quality data.
If the source information is inaccurate or incomplete, even the most powerful AI system won't be able to provide a reliable forecast. Errors in ordering, overstock, or product shortages—all of these are the result of poor data collection.

Stock levels and deliveries
It’s not just about what was sold, but also how much was available in stock and at the points of sale. For example, if a product didn’t sell, it may not be due to low demand—it might simply have been out of stock.

Trends and shifting preferences
Product popularity can change under the influence of social media, influencers, and global trends. For instance, growing interest in eco-friendly or gluten-free products can shift demand in a particular category.

Sales history
It remains the foundation for calculations but should be used in combination with other factors. Data must be cleaned of anomalies caused by short-term promotions, supply chain disruptions, or unexpected events.

Competitor activity
If a nearby competitor launches a major promotion, it may affect the sales of similar products in your store. Analytical tools can track price changes and marketing campaigns of your competitors.

Economic conditions
Changes in consumer income, inflation, or financial crises can influence demand across various product categories. During tough economic times, for example, consumers tend to choose mid-range products over premium brands.

Customer behaviour
Analysing in-store customer journeys, online sales data, reviews, and search queries helps predict future demand. If users frequently search for a certain product online, it’s a sign of likely demand in the near future.

Price dynamics
Changes in supplier prices or fluctuations in exchange rates can impact a product’s final retail price and its demand. For instance, a sharp rise in the cost of imported goods can lead to increased sales of more affordable alternatives.

Staff performance
Salespeople and merchandisers also affect demand. Poor product placement, a lack of consultants on the floor, or low staff motivation can reduce sales—even if customer interest is high.

Product shelf life
For items with a limited shelf life, it’s important to factor in not only demand but also the risk of write-offs. Products with short expiry dates require more precise forecasting and agile inventory management.

Without quality data, forecasting turns into guesswork. For accurate calculations, it’s essential to consider not only sales history but also current stock levels, the impact of promotions, external events, and customer behaviour. In the next section, we’ll look at how to analyse this data and make forecasts more accurate.

Demand forecasting is not just about analysing numbers—it’s a complex data-driven process that requires modern analytical tools.
Without automated systems, it’s difficult for companies to take into account all the factors affecting sales: seasonality, promotions, customer behaviour, and stock levels. It is analytics systems that enable retailers to see the full picture and make accurate decisions.

Most retailers still build forecasts using Excel reports and manual analysis.
This approach may work for small shops, but in large chains and complex supply networks, it quickly becomes inefficient. Automated analytics systems:

Unlike manual analysis, analytics platforms work with data in real time, identifying potential issues and offering solutions.
The system analyses demand trends, evaluates the impact of promotions, seasonal fluctuations, and competitor offers, and then adjusts forecasts accordingly. If a product is losing popularity, the algorithms automatically reduce order volumes, helping to avoid overstock. When sales rise after a promotion, the system assesses whether this is due to genuine demand growth or a short-term effect, and suggests the best procurement strategy. If demand for a product varies depending on the day of the week, analytics adjusts the supply schedule to prevent shortages or surpluses.

This approach allows retailers to respond quickly to changes, reduce forecasting errors, and optimise warehouse operations.

To get the most out of analytics, it's important to start by assessing current processes: identify which data is already being collected and how it’s being used. Next, choose a platform that can take into account not only stock levels and sales but also external factors such as promotions and seasonal fluctuations. The system is then integrated with internal and external data sources to enable more accurate forecasting.
It’s essential to set up automated reports and alerts so that staff can respond to changes quickly without wasting time on manual data handling. The final stage is staff training, as the effectiveness of analytics depends directly on how well the team uses the data provided.

This approach makes order management more predictable and resilient, allowing the business to adapt more quickly to changes in demand and reduce risks.

Modern retail simply cannot operate effectively without analytics. Automated systems allow businesses not just to analyse past sales, but to see the real picture of demand, identify trends, and manage processes in real time. Companies that use such solutions operate faster, more accurately, and more efficiently than competitors who still rely on manual analysis.

Traditional forecasting based solely on sales history no longer reflects the realities of modern retail.
Demand is shaped by a wide range of factors: seasonality, promotions, local events, customer behaviour, trends, and competition. Ignoring these aspects leads to inaccurate forecasts, overstocking, or shortages of in-demand products.

Retail is a fast-moving environment where outdated, template-based methods no longer work. Companies that embrace analytics and digital tools are achieving better results by improving forecasting accuracy, order efficiency, and customer service.

A precise forecast is more than just a way to avoid stockouts or surplus—it’s a strategic asset that helps businesses grow, reduce risk, and increase profitability.

Tilda Publishing