Product Demand Forecasting: How to Forecast Demand for Retail and eCommerce
In the ever-changing world of retail and e-commerce, adaptability is everything — and 2020 is clear proof of that. With the pandemic rolling on, businesses who were able to adjust their pricing strategy, stocks, and staffing in advance have been able to stand the storm relatively well.
The ability to forecast sales and demand, in this case, is an essential practice, without which it would have been almost impossible to adapt and answer the question ‘Will this product sell?’. In this article, we will try to shed some light on the subject and provide more information on what demand forecasting is. We will also show you how to apply it in your own retail or e-commerce businesses, regardless size.
What is demand forecasting?
By definition, demand forecasting is a type of systematic and analytical approach that uses various historical sets of data such as past sales, stock volumes, and market prices to predict future consumer demand.
This process is normally applied to obtain results that cover definite periods of time (a quarter, a year, five years, etc.) or to estimate future product success within a particular group (popularity of smart home technology among middle-class households, for example).
Aside from answering the will-it-sell question, the demand forecasting process helps us understand:
- What the gap between demand and supply of a product is.
- How wide and long a product line should be.
- If sales targets are achievable.
- What minimum/maximum inventory levels should be.
- Whether the price policy is adequate or not.
- If there are any overstaffing issues, etc.
Ignoring demand forecasts, on the other hand, can lead to poor inventory management, ineffective production, and poor product launch campaigns. In 2013, Walmart came across some serious stock issues in their stores due to warehouse mismanagement — the company had enough product, in terms of volume, but didn’t have enough staff to deliver it to their retail stores. This resulted in billions of dollars in losses in potential sales, and negative feedback from their customers.
In 2001, Nike developed and started to use supply chain software without proper testing and demand analysis. They ended up oversupplying excess amounts of low-selling models, while supply of the very popular Air Jordans were at a minimum. According to a company press release, this inadequate supply amounted to $100 million in losses.
Accurate demand forecasting, done in advance, would have helped to avoid such pricey problems, to say the least.
Importance of demand forecasting and benefits for retail and e-commerce
Demand is one of the most important metrics for any business, and the foundation for any activities and strategies to be built upon. Predicting staffing needs, actual production capacity, material supply, machine utilization — this is hardly a complete list of objectives that should be assessed while forecasting a demand.
Helps to forecast sales and demand for certain items
Demand forecasting helps make it much easier for sales departments to plan and carry out all their daily operations by setting tangible and achievable targets. This process also reduces the risk of performing actions that don’t do anything to positively affect the business.
Based on past sales data, information about general market conditions, and upcoming campaigns, managers can more accurately predict sales and total revenue.
Accurate sales forecasting reports, marketing budgets, production, and pricing policies are all especially important for seasonal products or those that have long life spans.
Increases stock turnover rate
Overstocking is an inefficient investment, and a highly volatile strategy, since you can’t immediately access your money if the market goes down. In this case, your budget is literally laying on the shelves, and can’t be reinvested into other activities like marketing campaigns, business development, etc.
Forecasting stock helps to increase turnover rates, which boosts the revenue generated over a specific period of time (hour/day/month). It also is a way for retailers to better understand how effectively they are using their warehouses by estimating revenue per square foot.
Helps to reduce wastage
A high rate of perishable losses has always been a major issue for the retail industry. According to Researchgate, loss figures in the ‘producer-middlemen-customer’ supply chain can amount to up to 25% of total production. This is due to poor cooperation between different links in the chain, an inability to supply the right volumes of product at the right time, inventory build-ups generated due to spikes in demand, etc.
By forecasting demands effectively, managers can remove the uncertainty of how much product to provide, mitigate the ‘ripple effect’ of overstocking, and increase turnover rates. This all, in turn, decreases production wastage, which is especially important for goods that have a short shelf life.
Staffing becomes more effective
Struggling to manage human resources effectively leads to financial losses. Staffing issues become obvious during market fluctuations and sudden changes in demand for specific products (like the face mask demand during the COVID outbreak).
Analysis of past employment needs helps HR departments plan future staffing requirements and avoid unnecessary costs. It also helps to assess desired staffing levels in different departments of a company, and the number of jobs needed to produce or supply goods, thus preventing staffing shortages to ensure that a company is compliant with all legal requirements in that regard.
Demand planning aspects and demand forecast metrics
A Goldilock principle — holding only a required amount of a certain product — perfectly describes the idea of demand planning. The demand forecasting process includes a set of different techniques applied across a wide range of supply chain components; however, the main goal and metrics always stay the same.
Smart financial decisions
To be able to create accurate plans, businesses should base their analysis on five core financial metrics:
- Revenue. The income that the business generates.
- COGS. Cost of goods sold. This is exactly how much money it takes to make a product.
- Gross profit. The profit that a retailer makes after subtracting the COGS.
- EBITDA. Earnings before interest, taxes, depreciation, and amortization. This metric is normally used to identify the general financial profitability of a company before non-operating expenses.
- P&L. Also known as a ‘profit and loss statement,’ this metric summarizes all the revenue, costs, and other expenses that occurred during a specific period (quarter, half-year, year).
By utilizing these metrics, it is possible to measure the effects of discounts on pricing strategy, and estimate product costs, expenses, and revenue growth.
Production forecasts are seen as an estimated input for all resources, raw materials, and associated costs. Here are some of the main metrics used to make effective predictions:
- Production volumes. Actual volumes of products that a company’s factory can produce in a day, month, or year.
- Capacity utilization. Metric that shows how close a company is to full production capacity.
- Maintenance costs. By evaluating equipment costs, it’s possible to determine whether any production units are overusing the company’s resources.
- Return on assets. Helps to understand how efficient the company’s assets are, and how much profit is generated per equipment unit.
Better understanding of the production chain allows company management to adjust production plans and make more informed decisions when it comes to new investments or business expansion.
Cost-effective inventory management
Retailers and e-commerce store owners won’t be able to meet customer demands if they don’t know what is going on in their warehouses. An effective inventory management strategy is part of a demand forecasting process that can be developed using the following metrics:
- Carrying costs. A combination of all expenses involved in the inventory management process. These are storage, administration, handling, obsolescence, theft, and damage costs.
- Items fill rate. A ratio of received volumes to ordered volumes of product. By utilizing this metric, managers can also measure order fulfillment performance.
- Inventory turnover. A main indication showing how frequently a company sells its products. A low turnover rate would normally mean overstocking or that products are not being sold as expected. High turnover, on the contrary, would mean understocking and potentially lost sales.
- Pick, pack & ship accuracy. Helps to understand to which degree of effectiveness the inventory is performing at the moment.
Not only does an effective inventory management system give more accurate forecasts, but it also reduces operational costs and working time for every employee involved in the process.
In a highly competitive market, demand directly depends on the number of competitors that are ‘on the map.’ These are existing companies and newbies that have just rolled out their new products. In such a dense environment, pricing and the ability to bring it in line with customer expectations means a lot.
Pricing strategy is largely based on information from prior analysis of market conditions, inventory effectiveness, production plans, etc. The full picture, in this case, comes together as a comparison of different parameters such as:
- Profit margins to sales volume (for a specific period)
- Levels of inventory to sales volume
- Market demand to production capacity
- Production costs to equipment effectiveness
- Inventory turnover to carrying costs
- Order cycle time to customer satisfaction level
These are a few examples of those relationships that should be assessed in order to come up with a precise pricing tactic.
Better marketing strategies
In the demand forecasting process, marketing is intended to balance the ‘before’ and ‘after’ states.
- ‘Before’ — accurate predictions help to come up with better and more client-focused marketing campaigns.
- ‘After’ — once a forecast is complete and retailers start to follow it, marketing campaigns allow for adjustments in demand, to keep it closer to the forecasted model line.
Types of demand forecasting
All demand forecasting activities are generally divided into two major groups, each of which has their own objectives.
- Macro-level. On this level, only general conditions are taken into account (health of the economy, national income, paradigm shifts, etc).
- Industry-level. Industry forecasts are normally prepared by the corresponding trade associations based on statistical data. These also provide a broad overview of industry and government regulations.
- Firm-level. Done by companies to assess product popularity, sales volumes, consumer demand and preferences.
- Short-term. Estimates are carried out for 3–12 month periods and normally include seasonal demand predictions. This type of demand forecasting helps to establish a production policy, pricing strategy, organization, and distribution policy.
- Long-term. Compared to short-term estimates, these are less accurate; however, they cover longer periods of time (5–10 years). Long-term forecasts are used to decide on the launch of new products or business expansion.
Practical demand forecasting models and methods
There are dozens of models and methods of demand forecasting that businesses use today. Which technique to pick depends on each particular case, so we strongly recommend you start forecasting only after you understand which data you want to assess.
Here are the most commonly used models and methods in product demand forecasting:
Demand forecasting models
- Qualitative demand forecasting. This model is useful when you don’t have enough data to work with. Here, the analysis is done by interviewing a group of experts or a focus group, or by using various surveys. This model is best suited when demand forecasting for new products, or for businesses that are just starting out.
- Quantitative demand forecasting. This model, on the other hand, deals with concrete initial information. By analyzing historical data, managers are able to generate forecasts based on the patterns sourced from that data. This works well when you need to model seasonal patterns or identify changes in demand trends.
Demand forecasting methods
- Causal method. This technique implies that the parameter that is being assessed depends on, or has a cause-effect relationship with, other parameters in a group. The causal method is used to build forecasts by a particular product or product category.
- Projective forecasting method. This method deals not with parameters, but patterns in general. It examines past trends (in sales, for example) and projects them into the future. This technique is used for setting mid-term and long-term goals.
- Product life-cycle forecasting. Here, predictions are done to figure out 1) which stage of the life-cycle a product is currently at, and 2) the possible length of each stage. Any product normally goes through five lifecycle stages: development, introduction, growth, maturity, and decline.
How to perform a seasonal demand forecast
Seasonal demand is a particular period of time that has a certain and, in most cases, predictable pattern. Seasonality often means that retailers need to adjust their inventory to meet market changes: reduce stock before the quiet months and increase it right at the moment when demand starts to rise.
Here’s a short checklist that will help you better optimize for seasonal demand:
- Take advantage of peaks in demand.
When the market is rising, make sure everything is ready: inventory, warehouse, sales team, marketing campaigns — should all be up and running.
- Prevent excess stock levels.
Make sure you know exactly when the seasonal trend begins to decline, to prevent overstocking and avoid carrying expenses.
- Keep suppliers informed and time orders and operations properly.
Inform suppliers about your seasonality in advance to make sure they will be able to deliver the right amounts of products at the right time. Ensure that your customer service team is able to handle all orders during the peaks in demand.
- Optimize inventory in advance.
Be prepared to enter a season. If the demand is about to rise, make sure you have space in your warehouse to allocate new stock. If the demand is about to go down, find ways to get rid of the excess volumes of product.
How to forecast product demand and predict market changes
#1 Define objectives
In the initial stage, it is important to understand what objectives your forecasts will be built upon. The final goal, of course, stays the same (demand forecast), but the ways of achieving that goal may be different: via pricing strategy adjustment, sales department optimization, inventory management improvement, etc.
#2 Collect data
Before making any forecasts or predicting market demand changes, make sure your historical data are relevant and correct. Otherwise, as the GIGO principle states, ‘nonsense on the input will produce nonsense on the output,’ and there will be no use in such forecasts.
#3 Analyze information
Once you have all the required data on hand, analyze it using one of the statistical methods for demand forecasting. Pick the one that suits your business best.
#4 Adjust budget
Based on your predictions, make budget adjustments to best fit the forecasted trend line. By reviewing the data and demand forecast metrics, you can also improve your supply chain, eliminate excess costs, rethink your approach to staffing, etc.
#5 Stick to what works best
If you have made business decisions in the past that proved to be excellent ones, stick to them. If something works well, use it. If you notice that the real-world market situation goes in a direction opposite to the one projected by your forecast, adjust your strategy. This is usually safer than waiting until things go really bad.
Methods of forecasting demand — keep your business in focus
Truth to be told, there’s no ultimate or one-size-fits-all solution of how to forecast demand or what retail demand forecasting methods to use for your business. That’s why we always recommend staying true to your business and its needs, letting it guide you when deciding between quantitative and qualitative demand forecasting methods.
Demand forecasting examples
Over the course of our work at FAYRIX, we have managed many retail and e-commerce projects.
Predicting demand in food retail
One of our clients, Grupo Bimbo — a globally known baking company — had us develop a model for predicting customer demand at POS’s. We worked through their statistical data and introduced a machine learning algorithm that helped to decrease returns of unsold products by 40%!
Predicting demand in fashion industry
For Milavitsa, the biggest lingerie producer in Eastern Europe, we developed a forecast model that helped predict dealer demand for a product, using prior purchase plans and two years of actual sales data. The result was a 150–200% decrease in slow-moving stock.
Predicting customer operations
Here’s one more demand forecasting example: for Sberbank, the largest bank in Russian Federation, we built a product demand forecasting model that allowed them to forecast the total and individual spendings of their customers.
Predicting market demand. Final thoughts.
Demand forecasting is what helps businesses set the right trajectory for all of their future moves. Whether it is a new product launch, insufficient sales or over-sufficient inventory, the answer always sits somewhere in your customers’ minds and not anywhere else. That’s why it’s crucial that you don’t overlook the importance of demand forecasting.
We hope this article gave you an idea of what the main purpose of demand forecasting is and helps you clearly identify all the benefits of demand forecasting. If you would like a free consultation for your retail or e-commerce project, we want to hear from you! Send us a message or fill out the form below and our customer support team will get back to you ASAP.