Introduction to Sales Forecasting

Traditionally, forecasting is managed using tools such as Microsoft Excel. This can work fine for smaller brands.

But, as your brand grows, you might find that working in Excel becomes more of a hindrance than a help. If you find that’s the case, there are a number of forecasting tools available today that can keep up with your brand’s growth.

But for now, let’s ensure we properly understand sales forecasting!

What is sales forecasting?

Sales forecasting is the process of estimating future revenue by predicting how much of a product or service will sell in the next week, month, quarter, or year. At its simplest, a sales forecast is a projected measure of how a market will respond to a company’s go-to-market efforts.

Sales forecasting adds value across an organization. Finance relies on forecasts to develop budgets for capacity plans and hiring, and production uses sales forecasts to plan their cycles. Forecasts help sales operations with territory and quota planning, supply chain with material purchases and production capacity, and sales strategy with channel and partner strategies.

What are the various components of sales forecasting?


For a given CPG company, we have sales data, which is a detailed breakdown of how many products we sold and how much money we made. For example, sales data may look like this:

We refer to this data as actuals. “Actuals” simply means this is what actually happened in the past. Actuals are one of our most important sources of truth; they serve as the basis for our forecasts.

Seasonality and baseline

Products often have seasonal trends that occur consistently every year. For example, candies tend to sell more around holidays such as Valentine’s Day and Halloween, and soups tend to sell more during the winter.

Seasonality values are multipliers that are applied to the sales value for a particular week. A value of 1.0 means “normal” or no change; less than 1.0 means the product trends poorly this week compared to normal, and greater than 1.0 means the product trends favorably this week. For example, the seasonality information for a particular snack bar could look like this:

What this means is: during the week of 2022-01-09, sales are expected to be normal; during the week of 2022-01-02, seasonal trends will cause the product to sell only 80% as well as it normally does; during the week of 2022-01-16, seasonal trends will cause the product to sell 120% as well.

What is the baseline?

To determine seasonality, we have to work backward from the baseline. Conceptually, the baseline is just a set of data that serves as a starting point. The baseline is whatever we say it is. However, we want it to be as useful and accurate as possible.

Surprisingly, the baseline does not exist in the real world — it is an idealized concept. We are saying: in a vacuum, our product would be selling this much, all else equal, with no incremental sales included from activities such as promotions. The ideal situation means we must ignore all factors that may affect our sales, which include both seasonal and non-seasonal factors. Non-seasonal factors are things such as supply chain shortages, political situations, pandemics, cultural shifts, and so on.

But the truth is that real-world data is messy, and chaotic, and is the result of all these factors. Thus we have to work backwards from real-world data to get a “clean” picture of what the baseline is.

Determining the baseline

We can get a baseline as follows: we take our historical sales data, then factor in both seasonal and non-seasonal events, and work backward from there. Let’s say this was our sales data:

There was a sharp drop in sales in the week of 2021-01-24. Let’s say we find out that there were supply chain issues during that week, and we estimate that it reduced our sales by 600 units. We adjust our data as follows:

Then we proceed to seasonality. Seasonal trends tend to be stable from one week to the next, with occasional spikes, such as for holidays like Halloween. In this case, let’s say we know our product sells better in the winter, and sales decrease sharply as weather gets warmer. Then our seasonality might look like this:

Then, we can assess whether or not these values make sense by dividing our sales values by the seasonality multiplier, to get our baseline:

Now, this would be our baseline. However, we feel that it doesn’t make much sense, because we expect sales to go down as the weather gets warmer. So, we adjust our seasonality as follows:

Now we get a more sensible-looking baseline:

To summarize: we adjust our seasonality values in order to get an accurate baseline.


Aside from seasonal factors, sales can also be affected by promotions (hopefully, in a positive manner). A company runs promotions so that more units get sold. In exchange, they have to reduce the price per unit, as well as pay retailers to run the promotions. Hence, we not only have to look at sales values, but also revenue.

Each promotion has a lift value associated with it, which refers to how much more we expect to sell compared to if we didn’t run the promotion. For example, if we run a promotion on chips by reducing the price from $4.99 to $3.99, that would be a 20% discount; however, we expect 40% more people to buy it because the retailer will display it more prominently on their shelves. Therefore, we could determine that 0.4 is an appropriate lift value.

Thus, the forecast during the time period where we run that promotion should have 40% higher sales. This by itself is not enough — we would also forecast the revenue, which takes into account both the discount and the increase in sales. Hence, revenue would be roughly 1.0 * 80% * 140% = 112% of the usual amount.

Keys to successful sales forecasting

Improving the accuracy of your sales forecasts and the efficiency of the forecast methodology depends on multiple factors, including strong organizational coordination, automation, reliable data, and an analytics-based process. Ideally, sales forecasts should be:

  • Collaborative. Leaders should synthesize input from a variety of sales roles, business units, and regions. Frontline sales teams can be of great value here, providing a perspective on the market you hadn’t considered before. 
  • Data-driven. Predictive analytics can reduce the impact of subjectivity, which is often more backward-looking than forward-looking. Using common data definitions and baselines will foster alignment and save time. 
  • Produced in real-time. Investing in the real-time capability to course-correct or forecast allows sales leaders to quickly gain insight so they can make more informed decisions. This enables them to quickly and accurately update the forecast based on demand or market changes.
  • Single-sourced with multiple views. Generating the forecast as a single source of data gives you great visibility into rep, region, and company performance, and helps align different business functions across the organization.
  • Improved over time. Use the insights provided by an improved sales forecasting process to create more refined future forecasts where accuracy improves overtime against a set of accuracy goals.


Why Sales Forecasting is Essential for CPG Brands

Own Your Future With Intelligent Forecasting.

Forecasting in Excel is tedious, time-consuming, and error-prone.

Take the guesswork out of forecasting. Vividly automates the entire sales forecasting process and updates any changes in real-time.

See how industry leaders like Liquid Death and Perfect Snacks benefit from Vividly’s intelligent forecasting.