demand forecasting

In this blog, we take a deep dive into demand forecasting, why it is critical for ecommerce businesses, the types of demand forecasting, and best practices you need to consider when planning your demand forecasting strategy.  

Let’s begin by understanding demand forecasting with a simple example. 

Have you observed how retail chains plan their merchandise for Thanksgiving or Black Friday sale every year? 

They invariably have a list of products with steep discounts or early-bird offers. It could be household electronics like ovens, washing machines and television sets, or devices like smartphones, laptops, iPads, or more basic items like lawn signs and bath mats.  

While, as a customer, you simply walk-in and purchase what you need, what these retail stores are doing at the backend is planning and estimating well-in-advance what their customers are more likely to purchase, and how many customers are likely to purchase each product. This is known as demand forecasting. 

Simply put, demand forecasting is using insights from past purchase trends to determine what the likely demand for a product or service in the future. Having said that, demand forecasting needn’t necessarily be based on past purchase behaviour. If a brand is launching an entirely new product, it can still forecast demand by holding focus groups and conducting surveys to gauge customer interest, and draw insights from past demand of related products.  

Why is demand forecasting necessary for ecommerce businesses? 

Estimate demand and supply 

Typically, ecommerce businesses have warehouses where they stock the products they offer on their platforms. Demand forecasting is a necessary step for them to understand the demand and supply at all times, and to manage inventory better. They don’t want to be left with a high stock of products that are simply lying around, going to waste. 

Arriving at the right pricing strategy 

Let’s say a fashion ecommerce brand is planning its forecast for every year. The most obvious insights they can draw is that sweaters will be on demand during winter and shorts will be in demand during summer. But, not all forecasting is as straightforward. Sometimes, there may be products that may suddenly increase in demand during an unexpected time – like, when there are sudden rains and customers need an umbrella.  

In such a scenario, you have two options; as an ecommerce retailer with heavy stock of umbrellas, you can either offer the umbrellas at a discounted price so that the stocks are depleted quickly, or if you know your umbrellas are worth the money, you can hike up the prices and earn higher revenue on this product.  

Determining the right pricing strategy to adopt in such a scenario is possible with demand forecasting. In the past, what have your customers had to say about your umbrellas? What have they been willing to pay? What price are your competitors offering it at? All these statistics put together help you estimate the likely demand and pricing for your product. 

Retaining customers 

We’ve all been in this situation where we have liked a product but it has been out of stock. Now, if we’ve repeatedly had the same experience with the brand, where their products often go out of stock, we would look to alternative brands to meet our needs. In such a scenario too, demand forecasting plays an important role in helping businesses estimate the likely demand for a product based on past queries and purchases, to help supply out of stock products immediately and retain customers with the brand. 

4 Types of Demand Forecasting 

Usually demand forecasting is done based on two parameters; quantitative analysis which is based on hard data from past sales and customer demand, and qualitative analysis which is based on market research, expert opinions from the industry, and prevailing economic conditions. 

Having said that, demand forecasting can broadly be categorised into four types; 

Macro-Level Forecasting 

Like the name suggests, macro-level forecasting incorporates data from a wide perspective. It extracts data from the industry the ecommerce business is operating in, data from competitors and data from related sectors. Businesses can use macro-level forecasting when launching new products or exploring newer avenues or markets to estimate demand, inventory, and likely return on investment.  

Micro-Level Forecasting 

This method is limited only to comparing your business’ performance in the past to what can be estimated for the future. For example, micro-level forecasting can be used to determine how each of your products have performed in the past, to plan inventory, cash flow required, and likely returns you will make in the future. 

Long-Term Demand Forecasting 

This kind of forecasting is done for a longer period which can be anywhere between 12 to 48 months. Long-term forecasting help businesses to identify sales patterns in their products, which can in turn help them understand how they need to alter their supply chain. For example, in the example we saw above on the demand for umbrellas – if the demand is likely to spike in the next 24 months owing to climate change and unpredictable weather, businesses need to use long-term forecasts to find out if their current suppliers can meet that demand, or if they need to onboard more suppliers, or move to a larger supplier.  

Short-Term Forecasting 

This forecasting is done between 3-12 months. Short-term forecasting estimates the demand for a product or category of products based on seasons, trends and events. For example, during the football season, a clothing retailer brand may see a spike in demand for football merchandise like t-shirts, shoes and collectibles.  

It helps businesses plan in advance and answer critical questions like; is the demand likely to spike a week before the football season, or during the season? Which of the football merchandise is likely to see a greater demand? What quantity of each set of merchandise needs to be stocked in the warehouse and so on. 

A Merit expert says, “A few best practices you can consider when performing demand forecasting are; ensure that you use both quantitative and qualitative data to plan demand and supply for your products. Secondly, collaborate with your sales teams to ensure that you get complete, accurate data on sales trends from the past. Lastly, always rely on industry experts to give you a broader perspective into how the industry and market is performing and how it is likely to change in the future.” 

Merit’s Expertise in e-Commerce Data and Intelligence  

Our state-of-the-art eCommerce data harvesting engine collects raw data and provides actionable insights  

  • Three to four times faster than standard scrapers  
  • At lower cost  
  • With Increased accuracy (up to 30% compared to standard scrapers)  

Our powerful, new scraper engine can gather massive data sets from multiple sites and geographies in real-time so you can stay informed on customer behaviours and market trends.  

Merit’s eCommerce data engine provides a high degree of confidence in insights generated from analytics – thanks to confidence in the data quality and access to enriched data.  

To know more, visit: 

Related Case Studies

  • 01 /

    AI Driven Fashion Product Image Processing at Scale

    Learn how a global consumer and design trends forecasting authority collects fashion data daily and transforms it to provide meaningful insight into breaking and long-term trends.

  • 02 /

    Construction Materials and Project Contacts Mining Using NER

    A leading UK construction intelligence provider, part of a £350m global information business, required detailed coverage of all current and upcoming UK construction projects, with accurate and full data at every stage of the project.