1Target specific products. Instead of focusing on a complete product line, identify the particular products you want to track. Doing this makes it easier to organize past data and forecast demand. For example, if you have an existing line of winter garments, focus specifically on gloves first instead of the entire line.
- Focus on your products that earn you the most income. For example, many entrepreneurs adhere to the 80/20 rule, which states that 20% of products or services offered by a business generally make up 80% of its revenue. Identify these products and track the demand for them.
- You may have to forecast demand for every product in your inventory, but it will be easier and more accurate if you do a few similar products at a time such as gloves, boots and winter hats.
- Consider creating a Sales and Operations Planning group that includes representatives from each department and task them with preparing a demand forecast.
2Review your marketing plans. Any marketing campaigns or sales promotions may increase the demand of your product. Look at the past data and see what was successful. See if there were special discounts or holiday sales that increased demand for your product. You want to take all of this into consideration when forecasting demand, particularly if you plan to repeat similar sales strategies.
3Review key indicators. Find out what’s behind the fluctuation in your customers’ demand. Key indicators include demographics and environmental factors. Demographics include age, gender, location, and any other set of identifying characteristics. Identifying the demand of key demographic groups helps to narrow the data pool for the forecast. Environmental factors affect demand as well. For example, a severe winter might cause a decrease in sales.
4Look at your marketplace. Analyze what competitors, customers, bankers, and other people in your marketplace are saying and doing. See if your competitors are running major sales or promotions.
5Look at the previous months. Look at both recent months and annual sales variances such as holiday time. This will help you determine annual and seasonal fluctuations. When looking at the past months, analyze the driving patterns behind the demand. Look at any price adjustments or any marketing campaigns that led to a rise in new customers. Business always increases for a reason, and a smart businessman or businesswoman will find out why. For example, you may have run a “buy one, get one free sale” in August for back to school shopping. If you choose to replicate these factors, consider that in your forecast.
6Determine your lead time. Lead time is the time between the initiation of an order and the delivery of a product. Knowing this will help you forecast demand. This will help you determine how fast you can make your product and meet demand.
- If you are purchasing your products from another company, the lead time is the time between placing your order and when it arrives on your doorstep.
- You can also determine lead time by examining the raw materials and components. Knowing your required production time will help you make a more accurate forecast demand. Focusing on a particular item helps to predict how much material you will need and the production time to make your product.
- When you have your production quantities estimated, look at the component demand of each item. For example, if you are manufacturing pencils, you will need to know how much wood, rubber, and lead to order based on your forecast.
Determining Your Approach
1Figure out which approach to use. There are four general approaches to forecasting demand. They include judgmental, experimental, relational/causal, and time series. Choose the best approach based on the history of your product. The experimental approach, for instance, is used mostly for new products that have no history data in the marketplace. These approaches are how you will gather most of your data.
- You can combine the approaches to create a more accurate demand forecast.
2Consider judgmental approaches. This method draws upon the collective market insights observed by your sales team and managers to determine demand. These people can provide somewhat or, in some cases, very accurate demand forecasts based on their own personal knowledge and experience. However, the data you gather from them might be unreliable, as it relies on your experts’ own personal views. For this reason, data derived from judgmental approaches are best used to make short term demand forecasts.
- There are several different ways of going about this, depending mainly on who you use for your panel. However, you don’t need to use them all for a proper judgmental approach. You may choose or any combination of them to achieve your goals, depending on which groups you think would provide the most accurate judgment.
3Determine if you need to use an experimental approach. This approach works best for new products, and it is not useful for existing products that have a historical demand record. This approach takes the results from a small number of customers and extrapolates the findings to a large number of customers. For example, if you contact 500 people at random in a particular city and 25% say they will buy your product within 6 months, you can assume this percentage applies to 5,000 people.
- If a small group of targeted customers loves a new technology and responds well to the test marketing, you can extrapolate that number to also forecast national demand. The problem with this approach is that it often collects more information about the customer’s preference towards your product rather than demand data.
4Consider using a relational/casual approach. This approach attempts to find out why people buy your product. The idea being that if you can understand why people buy your product, then you can create a demand forecast based on that reason. For example, if you sell snow boots, then you know the demand for your product is weather related. If the weather forecast predicts a heavy winter, you know that there will be a higher demand for your snow boots.
- These approaches include life cycle and simulation models.
Calculate demand using time series approaches. Time series approaches attempt to mathematically calculate demand using past figures and trends as a guide. Specifically, you can use moving averages, weighted moving averages, and/or exponential smoothing to attempt to accurately predict your demand. These approaches will give you harder numbers than other approaches, but must be combined with other, subjective approximations to account for the effects of future changes in the market or business plan.
Using Judgmental Approaches
1Form a jury of executive opinions. Gather a small group of high-level managers in your company and have them estimate demand. Each member of this group can provide valuable insight based on their experience with the market. They can also help in selecting quality material vendors and marketing campaigns. This approach is inexpensive and not as time consuming as other judgmental approaches. The downside is that these projections are based on the opinions of the experts who may be biased and pushing their own agendas.
2Create a sales force composite. Ask each salesperson to project their sales. The sales team is closest to the marketplace and is knowledgeable about the desires of the customer. Combine these projections at each level of sales by city, state, and region. The upside to this approach is its low cost and the ease of collecting data. The downside to this approach is that it’s based on consumer opinions, which can easily change. Also, the salesperson may inflate the numbers to help ensure his or her job security.
3Hire individual market experts. Market experts watch for industry trends and consult with your sales force to predict demand. These could include trade magazine writers, economists, bankers, and professional consultants. An individual can only gather a limited amount information, however, so it is recommended that you assemble a team of market experts to gather as much data as possible.
- These individuals can provide you with insight about the markets that is at a higher level than your own sales team may be able to provide. However, being outsiders to your company, they have less of a grasp on the demand for your individual products. You should use these people to forecast market demand and then estimate how well your company may fare within that market using internal judgments.
4Use the Delphi Method. First, create a panel of experts. This can include a group of managers, selected employees, or industry experts. Ask them individually for their estimate of demand. Have them answer questionnaires in two or more rounds. After each round, present the findings of the previous round anonymously. Encourage the experts to revise their answers with the previous findings in mind. The goal is that the group will eventually start to agree on the forecast.
- Use a pre-defined stopping place such as a certain number of rounds, consensus, or stability in results.
Using Experimental Approaches
1Survey your customers. You can collect information from them in several ways: telephone or e-mail surveys, statistical reviews of customer order history, and market trends. Ask them about their purchasing plans and projected buying behavior. Use a large pool to help generalize results. Ask them how likely they are to buy your products and tally the results.
- Customers are in the best position to know the demand for a product. The danger from surveys is that they often overestimate actual demand. While a customer may show interest in your product, actually buying it is a different thing altogether.
- Keep in mind that conducting surveys can be expensive, difficult, and time consuming. Surveys rarely form the base of a successful demand forecast.
2Use test marketing. Use this during the early stages of your product development. Find a small, isolated, area that has your targeted demographic. Roll out every stage of your marketing plan including advertising, promotion, and distribution plans. Measure product awareness, penetration, market share and total sales. Fine tune your market strategy based on the information you receive so that you will run into fewer problems when you launch your product nationally.
3Host consumer panels. Gather a small group of potential customers in a room and let them use your product and discuss it. The customers are usually paid a small amount for participating. Panels are similar to surveys in that they are more useful to analyze the product rather than forming the basis for a demand forecast.
4Use scanner panel data. Find a large set of household customers to agree to participate in an ongoing study of their buying habits at grocery stores, for example. Have these customers agree to submit information such as the size of their households, their ages, their household income, and any other information you find relevant to your product. Whenever they buy groceries, their purchases are recorded and analyzed. This data can be collected when they use their store grocery card. This creates a rich database to create statistical models and see relationships in data.
- As with other types of experimental approaches, it can be difficult to apply these results to demand forecasts.
Using Relational/Causal Approaches
1Examine previous years’ sales for monthly or seasonal trends. Look over sales figures for past years to determine which times in the year account for the higher percentage of your sales. Are they constant? Do you experience higher sales in winter or summer? Measure the increase or decrease in sales during these times. Was the change higher or lower in certain years? Then, think about why this might be the case. Use what you’ve learned and apply it to the current year’s forecast.
- For example, if you sell snow boots, you might have experienced a particularly large boost in sales in a cold winter. If this year is forecasted to be a similarly cold winter, you should increase your demand forecast accordingly.
Look for customer reactions. This refers to situations where a change in your product or its market resulted in higher or lower sales. Create charts of your historical sales for the product and mark important dates, for example a price increase or the introduction of a competing product. This can also be broader, like a reaction to the shifting economy or changes in consumer spending. Read relevant trade journals and newspaper articles to gather this information. Having all of this data at hand can give you a better idea of what might affect your future demand.
3Create a life cycle model. A life cycle refers to the “life” of your products, between when it was first introduced and the present day. Look at the sales of your product at various stages. Examine the nature of customers who buy the your product during these stages. For example, you will have early adopters (those who love the latest technology), mainstream buyers (people who wait for product reviews and referrals), laggards (they only buy when the product has been out for a long time), and other types of consumers. This will help you determine your product’s life cycle trends and the demand patterns for your product.
- The industries that use this model the most include high technology, fashion, and products facing short life cycles. What makes this approach unique is that the cause of the demand is directly linked to the product’s life cycle.
4Use a simulation model. Create a model that simulates the flow of components into manufacturing plants based on your material requirement planning schedules and the distribution flow of your finished goods. For example, calculate the lead time to receive each component including shipping time no matter where it is sourced in the world. This will give you insight on how fast you can make your product to meet the demand.
- These models are known to be difficult and cumbersome to create and maintain.
Using Time Series Approaches
1Use the moving averages method. This is a mathematical technique used if there are little to no trends present in your data. This method will provide an overall impression of data over time. Find out the actual demand for the previous three months. Once your have the total, divide that by four (accounting for the next month). The formula will be F4 = (D1 + D2 +D3) ÷ 4. In this equation ‘F’ represents the forecast and ‘D’ correlates with the month.This equation works well for steady demand.
- For example, forecast = 4,000 (Jan.) + 6,000 (Feb.) + 8,000 (March) /4 = 4,500.
2Determine the weighted moving average (WMA). If you have fluctuating demand, use this formula,which takes variation into consideration. The formula is WMA 4 = (W * D1) + (W * D2) +(W * D3). The ‘D’ stands for demand and the number correlates with the month. ‘W’ is the weighted constant, which is normally a number between 1 and 10 and is based on past history.
- For example, WMA = (4 * 100) + (4 * 250) + (4 * 300) = 2,600.
- Use a greater weighted constant number for more recent data and a lesser number for older data. This is because more recent data has a stronger influence over the forecast.
3Determine exponential smoothing. This technique is an averaging method that considers recent changes in demand by applying a smoothing constant to the most recent data. This is a useful technique if the recent fluctuations are the result of an actual change such as a seasonal pattern (holiday time) instead of random changes. 
- Find the prior periods’ forecast. This will be represented as (Ft) in the formula. Then, find the actual demand for product during that time period. This will be represented as (At-1) in the formula.
- Determine the weight being assigned to it. This will be represented as (W) in the formula.This ranges between 1 and 10. Assign the lower number for older data.
- Put your data into the formula Ft = Ft-1 + W * (At-1 – Ft-1) or for example, Ft = 500 + 4(W) * (590 – 500) = 504 * 90 = 45,360.
1Compile your results. Once you have collected your data, create a chart or graph that shows the demand forecast. Do this by crossing your product demand quantity with the upcoming months. For example, if you create a line graph, put the months on the horizontal axis and product demand quantity on the vertical axis. If you forecasted that you will need 600 units in October and 800 in November, then place those points on the graph. Draw a line between the points. You can also plot past data on the graph to compare your research data with historical data.
2Analyze your results. You now have your results tabulated or displayed in an easy to read form, but what do they mean? Look for trends, like growing or declining demand, and cyclicality, like busy seasons or months. Compare your data to that of previous years and see how it stacks up as far as volume and pattern. Look for evidence in the data that your marketing plans are working or have worked in the past.
- Additionally, go back and determine how exact you believe your forecast to be. Have you been optimistic with your forecast? How large of a margin of error do you expect?
3Display and discuss your forecast. Show your forecast to the appropriate people in your company and discuss it with them. Gather input from sales and marketing, finance, production, and all other managers and then revise your forecast. When everyone agrees on the forecast, they can plan a better business strategy.
4Monitor and modify your forecast. As you gather new data, modify the forecast to reflect this. You want to use all information as it comes to you. If you do not constantly monitor and update your forecast, you can make costly mistakes and it will affect your financial sustainability.
Ask a Question
If this question (or a similar one) is answered twice in this section, please click here to let us know.
Thanks to all authors for creating a page that has been read 100,153 times.