What is sales forecasting? A thorough explanation of methods, tools, and points to improve accuracy for different purposes.

Sales forecasts are important indicators that are essential for business decision-making and management stability. In particular, when formulating marketing activities and sales plans, forecasting "how much will be sold in the future" is the foundation of any strategy.

However, in reality, many people may have questions such as, "I don't really know how to make a sales forecast," or "How is it different from a goal?"

This article explains everything from the basics of sales forecasting, the differences between forecasts and targets, specific forecasting methods, data that can be used for forecasting, tools and methods, and points to further improve accuracy, in a way that is easy for non-experts to understand.

What is Sales Forecasting? Meaning and Basics

Sales forecasting is the act of estimating future sales based on past and current data.

For example,

• Sales performance for the past year and past few years
• Current number of business negotiations and customers
• Market growth rate and economic trends

The basic approach is to use data like this to predict, "If things continue like this, we expect to sell this much."

The important thing is that sales forecastsOutlook based on reality, not hopes and goalsIt is impossible to predict the future with complete precision, but making a prediction as accurately as possible can help to ensure that management and marketing decisions are made without wavering.

The difference between sales forecasts and sales targets

Sales forecasts are often confused with sales targets.
Correctly understanding the difference between these two will help make actions on the ground more persuasive.

For example, "Based on the current sales pipeline, next fiscal year's sales are predicted to be 110 billion yen."Sales forecastで す.
In response to that, the answer was "Let's aim for 120 billion yen next fiscal year in order to expand our market share."Sales target.

Sales goals should be challenging,Only by using predictions as a foundation can we design realistic goals..

Three benefits of sales forecasting

Below we will explain in detail the benefits of creating a sales forecast.

1. Business decisions become smoother

Executives and managers make decisions every day about where to invest their limited resources (people, things, and money).
At this point, the quality of your decision will change significantly depending on whether or not you have a prediction for what future sales will be.

For example, if sales are expected to grow significantly, it will be necessary to increase staff and inventory, but conversely, if sales are expected to fall, cost reduction and risk management will become urgently needed.

Seeing the future strengthens the decisions we make today.

2. It helps you set reasonable goals

If you ignore sales forecasts and set an intuitional goal such as "20% more than last year," a gap will arise with the actual results.
If goals are too unrealistic, employees will not be convinced and the plan is likely to fail.

On the other hand, if the goal is based on predictions,This allows for goal design that is easy for those on the ground to understand and highly executable.
Sales forecasts are also a "bridge that connects on-site perception with management's perspective."

3. Improved ability to explain to outsiders

Sales forecasts can also be used as persuasive material for stakeholders outside the business, such as banks, investors, and shareholders.

In response to questions such as "What is the basis for this forecast?" and "Is it really possible?"Data-based predictions allow for reliable explanations.

Sales forecasts are not only a tool for internal decision-making, but also a weapon for gaining the trust of others.

Types of data that affect the accuracy of sales forecasts

The accuracy of sales forecasts varies greatly depending on the type of data used.
Ideally, you should gather information from both inside and outside the company.

Internal data (reflecting internal facts)

• Past sales performance (weekly/monthly/yearly)
• Sales by business, organization, department, and product
• Current number of transactions (customers)
• Customer LTV (Lifetime Value)
• Average purchase frequency, unit price, repeat purchase rate, etc.

It is important that the data based on the above indicators ensures two things: accuracy and transparency so that anyone can use it. Furthermore, it would be even more ideal if each piece of data was aggregated on a daily or weekly basis.

External data (capturing signs of environmental changes)

It is also a good idea to refer to indicators and data from external factors as well as from within the company. For example, the following data can be used:

• Industry-wide growth rate
• Economic trends and price changes
• Competitor sales strategies and price fluctuations
• Legal reform, strengthening and relaxation of regulations
• Seasonality, weather, trend changes, etc.

The key to making solid predictions is to not just look at the "internal past" but also incorporate the "external future."

TypicalThree ways to create sales forecasts

Next, we will pick out and explain three methods of sales forecasting.

1. Trend prediction based on past performance

This is the most basic method. When using past sales data, sales forecasts can be calculated using the following formula.

Example:

• The year before last: 80 billion yen → Last year: 100 billion yen (25% increase)
• Assuming the same level of growth this year → 100 billion yen x 1.25 = 125 billion yen

While this method is simple and easy to understand,It is difficult to take into account changes in the external environment or internal policiesThere is a weakness.

2. Make forecasts based on sales pipeline and deal stages

This is a method of estimating "how likely it is that ongoing projects will lead to orders" by using sales activity data (sales negotiation status registered in SFA or CRM).

Example:

• Sales stage A (probability 30%): 100 cases x average unit price 100 million yen x 0.3 = 3 million yen
• Consultation stage B (70% probability): 50 cases x average unit price 100 million yen x 0.7 = 3.5 million yen
• Total sales forecast: 6.5 million yen

This methodIt is good for understanding short-term prospects and is used routinely by sales departments.

3. Prediction using statistical models and AI

This method involves analyzing multiple factors using formulas and algorithms to make more accurate predictions.
Here are some approaches:

• regression analysis: Express the relationship between variables that affect sales in a mathematical formula
  (Detailed article:How to forecast sales using regression analysis? Explaining the basics, steps, and points to note）
• Time Series Analysis: Forecasting based on seasonality and trend cycles
• Machine learning: Learning patterns from huge amounts of past data
  (Detailed article:Explaining the benefits, procedures, and methods of sales forecasting using machine learning）

In the field of marketing, there are "MMM (Marketing Mix Modeling)" is the technique used.

At XICA, we use these advanced models toSales forecast support that leads to decision-makingOffers.

(Detailed article:Predictive analysis of business results through marketing using MMM）

How to choose tools that can be used for sales forecasting

The accuracy, operational burden, and scope of usable data of sales forecasts vary greatly depending on the tool used. Choosing the right tool is crucial, as it aligns with your objectives and your company's data structure. Here, we outline four representative approaches.

Excel spreadsheets

This is the easiest way to get started. You input past sales data and use functions such as moving averages and linear regression to estimate future sales. Its strengths are that there are no initial costs and it is easy to use intuitively.

However, as the amount of data increases, management tends to become more dependent on individual users, making simultaneous editing by multiple people and automatic synchronization with the latest data difficult.While it can be useful for simple monthly and quarterly forecasts and as a starting point for analysis, it has limitations when it comes to highly accurate medium- to long-term forecasts.

Free download of case study

A guide to multiple regression analysis in Excel that empowers marketers
~ Understand the correlation between marketing measures and business results ~

Download Now

SFA/CRM (Sales Force Automation/Customer Relationship Management) tools

Sales force automation (SFA) and customer relationship management (CRM) systems such as Salesforce, HubSpot, and Dynamics 365 excel at forecasting based on the stage of a deal and the likelihood of closing it. They can aggregate deal data entered daily by sales representatives in real time and automatically visualize "what the expected closing date for this quarter is."

This tool is highly compatible with B2B companies and business models with clearly defined sales processes. On the other hand, its applicability is limited for startups with limited historical data, and for business models that do not involve individual sales negotiations, such as brick-and-mortar stores and e-commerce.

BI Tools and Dashboards

BI tools such as Google Lookup Studio, Tableau, and Power BI are well-suited for integrating and visually understanding data from multiple sources (sales, advertising, inventory, customer data, etc.). Rather than making predictions themselves, they function as a foundation for centrally managing and visualizing the data that underlies those predictions.

It is suitable for creating reports for management and for sharing data across departments, and is often used in combination with sales forecasting tools.

Statistical models and MMM (Marketing Mix Modeling)

The three methods mentioned above are all based primarily on "past company data." In contrast, predictions using statistical models are:We simulate future sales by simultaneously considering multiple internal and external factors such as advertising spending, seasonality, competitor activity, and economic indicators.The capabilities differ significantly.

In particular, MMM statistically breaks down how much each marketing strategy, whether offline or online, such as TV commercials, web advertising, and social media campaigns, contributes to sales, and then predicts future trends. Its key feature is that it can answer questions such as "How will sales change if we strengthen strategy A?" and "How should we allocate the budget for the next fiscal year?" based on data.

This cannot be captured by forecasts based on Excel or sales pipelines.Influence of external factorsBecause it can be incorporated, it is an effective complementary approach for companies with multiple marketing channels or those that want to create highly accurate medium- to long-term business plans.

The criteria for selecting a tool should not be "the abundance of features","Does it align with our company's challenges, data structure, and forecasting objectives?"Therefore, for many companies, a realistic approach is to start with Excel to understand the issues, and then gradually move to SFA or MMM as needed.

Approaches to Sales Forecasting by Industry and Business Model

Sales forecasting approaches vary significantly depending on the industry and business model. Understanding your own business structure and selecting the appropriate methods and data is essential for improving accuracy. Here, we'll organize our thinking based on four representative business models.

B2B (Business-to-Business Sales)

In B2B business,Pipeline forecasting based on negotiation stage and likelihood of securing a deal.This is the core of the process. By accumulating information such as "how many projects are currently underway and what the probability of each project being awarded is," we can accurately calculate short-term sales forecasts.

One point to note is that there is a tendency for variations in how accuracy is estimated depending on the person in charge. If discrepancies in perception accumulate, such as "They say 50%, but the actual order rate is 20%", the gap between prediction and reality will widen. Standardize sales opportunity data with an SFA tool,We regularly review our accuracy criteria by comparing them with past order records.This is the key to improving accuracy.

Furthermore, in B2B transactions, where there is a time lag between receiving an order and recording revenue, it is necessary to clearly define the **timing of revenue recognition (contract signing, delivery, acceptance, etc.)** before making forecasts.

B2C/EC (Business-to-Consumer E-commerce Websites)

In B2C and e-commerce, instead of individual negotiationsChanges in purchase frequency, average transaction value, and number of customersWe build our forecasts around this. The basic formula is "Sales = Number of Customers × Average Purchase Price × Purchase Frequency," and we start our forecast by deciding which of these elements to increase and how.

Another characteristic is that it is highly susceptible to seasonality and external factors. For example, understanding past seasonal fluctuation patterns as time-series data, such as the Christmas shopping season, year-end sales, and back-to-school season, leads to more reliable predictions. Advertising spending and competitor price fluctuations also directly impact sales,A mechanism to incorporate changes in marketing strategies into predictions.Is important.

SaaS/Subscription

In monthly and annual recurring billing models,MRR (Monthly Recurring Revenue) and Churn Rate (Customer Cancellation Rate)This will be the central indicator for sales forecasting.

The basic idea is a cumulative calculation: "This month's MRR + New customer acquisition MRR - Churn MRR = Next month's MRR". We incorporate both how to increase the number of new customers through advertising and marketing measures and how to reduce churn among existing customers as separate drivers in our forecast.

Even a slight change of just 1-2% in the churn rate can make a huge difference in the annual sales forecast, so the accuracy of churn management and customer success directly impacts the accuracy of sales forecasts. (Based on LTV (Customer Lifetime Value))Cohort analysisCombining it with this is also effective.

Retail and store business

In retail businesses with physical stores,Number of customers, average transaction value, and purchase rateThese three indicators form the basis of the forecast. In addition to these, external factors such as location characteristics (surrounding population, trends of competing stores), weather, and local events directly affect sales, so it is necessary to incorporate these into the forecast.

Companies with multiple stores segment their customers based on attributes such as area, store size, and customer base.We estimate the sales of the new store based on the performance of similar stores.The method is effective. Furthermore, to quantitatively understand the impact of offline measures such as flyer distribution and TV commercials on the number of customers visiting a store, an integrated analysis method such as MMM is suitable.

When a business model is multifaceted (e.g., e-commerce + physical stores, SaaS + implementation support, etc.), it is more accurate to forecast each revenue stream separately and then build up the overall company sales forecast. Attempting to forecast the whole with a single formula is likely to result in figures that deviate significantly from reality, so caution is advised.

Three key points to improve the accuracy of sales forecasts

Don't just make a prediction once and leave it at that.Check the difference (error) between the forecast and the actual resultsAnd it is important to continue to improve it.

Furthermore,

• Monitoring in short spans such as daily or weekly
• Reflecting the impact of new products and initiatives in forecasts
• Incorporate field information from sales and marketing departments
• Consider changes in external factors (policies, climate, SNS buzz, etc.)

Combining these factors increases the practicality of predictions.

A highly accurate sales forecast is a forecast that can be used. It is only valuable if the field can make use of it.

Sales forecast simulation: Examining "what if" to support decision-making

In sales forecasting,Simulation is a powerful tool to support decision-makingBased on past data and trends, you can hypothesize various "what if" scenarios and visualize the impact on sales.

For example,

• "What will happen to sales if we increase advertising costs by 20%?"
• "How much will sales drop if our primary channels become unavailable?"
• "If we release new products A and B, how will sales growth differ?"

To such questions,The value of simulation is in obtaining a well-founded forecast based on numerical values.
In particular, by utilizing MMM,Prediction under complex conditions with multiple variables moving simultaneouslywill also be possible.

In practice, when considering budget allocation and investment decisions, "visualizing the sales impact" has the secondary effect of making it easier to reach consensus within the company.
Simulation can be said to be an essential element in achieving both "confidence" and "accountability" in decision-making.

Summary

Sales forecasting is not just about making up numbers. It is a very practical tool for thinking that connects the "current reality" with the "future we are aiming for."

It serves as a foundation for designing goals realistically, a basis for allocating resources, and a persuasive tool for winning trust both inside and outside the company.
That is why, rather than feeling or hope,Data-driven outlookHaving this will give you strength in business.

Starting with the handy Excel tool is a good choice, but if you need more precise analysis, you should consider more advanced methods such as statistical models, AI, and MMM.

In particular,Companies that want to visualize the sales impact of their marketing activitiesFor this reason, MMM is a very effective approach.
At XICA, we support "predictions that directly lead to results" centered on MMM. Why not turn data into "power that leads to decision-making" rather than just "collecting it and leaving it at that"?

Author information

XICA Co., Ltd. Marketing Team