Seasonality is a characteristic of time-series where the data has predictable and somewhat regular fluctuations that repeat year over year. It is safe to assume that any pattern of data changes over one-year periods represents seasonality. It is usually driven by weather or commercial seasons.

We have to differentiate the term from cyclical data, as the latter can span over various times periods, shorter or longer than one year.

Please, keep in mind that I am writing this article the intention to avoid getting into statistics. I am instead trying to look at seasonality from a financial analysis and modeling perspective.

Understanding Seasonality

Lots of data sets are affected by the time of year, and it’s essential to adjust for seasonality. This way, we can achieve a more accurate comparison between periods. When we adjust for seasonality, we are effectively evening out the swings in the observed time-series so we can draw proper conclusions from our analysis.

Retail sales are the obvious and most common example. They tend to be higher in the holiday season. If we compare margins and revenues, we might think the market is growing around Christmas as compared to Autumn. However, if we adjust for seasonality, it might turn out that’s not the case.

In sales, changes in customer behavior drive seasonality, and the business rarely has control over that. What we can do is analyze the trends and patterns and use the results to plan accordingly.

Examples of Seasonality

During the summer, there’s a high demand for ice cream, which falls in the winter. Knowing that we can forecast the market for ice cream pretty well as it follows a similar pattern every year. Therefore we can assume that if we analyze the performance of an ice cream company, we will notice strong seasonality.

When reviewing the expense accounts of a manufacturing company, we notice heating costs are highest in the winter and non-existent during the warm months. We can assume this trend will repeat over the years, which shows seasonality.

Seasonality in Financial Analysis and Modeling

We define seasonality as periodic fluctuations and cycles in specific areas of the business, following some pattern. It can be a calendar one like summer and spring or an economic one like the Christmas season.

Most often, we use the concept in the analysis of stock prices and market trends. Businesses can use seasonality to improve working capital management.  As an example, if we usually have lower sales in the second quarter of the year, we can decrease our inventory purchases to avoid overstocking.

Whenever we analyze a company, we must check for and consider seasonality in its performance. Higher sales in a specific season might be misleading for investors if they don’t account for the seasonality of the business. Therefore analysts have to consider past performance as well to understand how low seasons even out such high seasons over time.

As we analyze our business, it is paramount that we take a look at our clients as well. If we provide services to a company with specific seasonality, we might experience the same patterns. It is vital to analyze our clients’ behavior and incorporate the findings in our plan for the future.

A common way to decrease the effect of seasonality is to diversify the company’s product portfolio. For example, if we sell wafer cones for ice cream during the summer, we can start selling chocolate-coated wafers in the winter, and thus even out the seasonal pattern of our sales revenue.

High Season

Analysts need a comprehensive understanding of the seasonality of the business and prepare for the busy season in advance. Imagine we are in charge of a beach towel business. If we wait for the demand to start rising at the end of spring, it might be too late to meet it. We will face issues trying to supply that many products if we only react to the market. Therefore, we need to plan and ramp up production in the previous months. On the other hand, our materials suppliers also need to analyze our seasonality and prepare to meet our increased demand for fabrics.

Low Season

It is essential to focus on the off-season periods as well. When we have lower sales and less pressure on the team, it’s an excellent opportunity to focus on the company’s support activities. Such times are great for various undertakings, like:

• Implementing a new CRM or other systems;
• Obtain ISO certifications for our production facilities;
• Raise brand awareness through different marketing campaigns;
• Participate in social responsibility programs;
• Develop new products;
• Stock-up for the upcoming busy season;
• Train employees and recruits;
• Fine-tune processes within the company.

Example Seasonality Model

The concept we are looking at applies to many more scenarios within financial modeling and analysis, and not only sales. We might identify seasonality in various other data sources that are still relevant to the business.

The following is one of the simplest ways to prepare estimations while considering the seasonality of the time-series. However, it is fast to set up and extremely useful when preparing overall budgets and performance analytics.

Imagine we are running a bath towels manufacturing company, and we are looking to expand into beach towels. To plan our beach towels marketing spending, we need to analyze its seasonality. Google Trends is an excellent source of information in such circumstances. Let us go ahead and download all the available data for the worldwide interest towards the keyword ‘beach.’

And it’s seasonality from the first look.

Here’s the data in Excel, showing average monthly interest in the keyword. To facilitate our analysis, we have added a few additional columns. These are the name of the month, a period formatted as a ‘date’ field, and a period number.

Let’s start analyzing the data. The easiest way to forecast in Excel is by using the linear regression model, which we can calculate with the SLOPE and INTERCEPT functions.

Now that we have these coefficients let’s use a linear regression function of type f(y) = ax + c to calculate a forecast for the same periods.

When we plot those two, we can see that our forecast captures the slightly declining trend but misses out on the seasonality aspect of the data set.

To correct our estimates for seasonality, we will calculate the Seasonality Index for each month. We do so by averaging out all values for a specific month of the year and see what their ratio is over the average of all data points.

Now that we have our Seasonality Indexes, we can use them to adjust our linear forecast.

Let’s plot the last year and a half and our forecast to the end of 2021 (January 2019 to December 2021). We immediately see that our prediction is performing much better when compared to the existing values for 2019.

However, it completely misses the latest developments in the first half of 2020. As we can see, the global pandemic leads to enormously lessened interest in the ‘beach’ keyword, as people were facing other more pressing issues.

It is paramount to understand that this is one of the simplest ways to forecast seasonal time-series, and it can’t handle significant one-off events. However, we see interest in the keyword is resurging in June 2020, so our forecast might end up being appropriate for our plans for marketing spending in 2020 and 2021.

Conclusion

The concept refers to fluctuations in performance due to external factors that have a similar occurrence pattern every year. These fluctuations are generally predictable and repetitive. The harder seasonality hits the business, the easier it is to identify this pattern and anticipate it with the appropriate preparations.

It is one of the most popular statistical methods we use to improve performance forecasts. We must look for data behavior patterns and identify seasonality whenever we analyze time-series data.

Please, show your support by sharing this article with colleagues and friends.

Also, don’t forget to download the example Excel model below.