Seasonality is the presence of periodic, repetitive and regular patterns and variations in your data. Seasonality may be caused by various factors external to your business, such as the weather or holidays and can impact different parts of your business in different ways. For example, a clothing retailer may sell more items in December due to the holiday period, however particular items such as sunglasses or coats may see different spikes in demand.

Understanding periods of seasonality can help inform decisions about staffing, marketing campaigns, inventory and more during busy periods. Understanding seasonality also ensures changes are put into context - i.e. identifying whether a drop in sales is expected or unusual for the period.

If you’re interested in further reading about seasonality and seasonal adjustment, please refer to this article by the Australian Bureau of Statistics.

What types of seasonal analysis does Anna perform?

1. Analyses monthly averages to understand variances and identify peaks and troughs

Note: This visualisation appears when you use the word ‘seasonality’ in a question or in the secondary insights when you ask a ‘trend’ question.


Anna calculates the monthly average based on the entire time range of the dataset. This helps you identify if sales are consistent across months or if there are particular spikes. 

In this example, by analysing the average sales per month using the 30 months of data available, Anna has identified that there is significant variance in the average sales between different months and that the peak each July this year accounted for approximately 10% of sales.

Note: This analysis is triggered in the secondary insights when the word ‘seasonality’ is used in a question. For example: Seasonality of Sales for the Pharmaceutical department


Anna uses a STL model to decompose your data into three separate components:

  1. Fitted trend - this is the underlying, longer-term direction of your data 

  2. Fitted seasonality - periodic, regular, repetitive variations in your data

  3. Remainder - the ‘leftover’ component once trend and seasonality are removed from your data. The remainder is the irregular component of your data.

This model allows Anna to understand if your business is seasonal or not. We will explore each component of the visualisation above in more detail:

1. The fitted trend is the underlying, longer-term direction of your data  In this example, you can see there is a steady increase over time in the fitted trend. The fitted trend shows seasonally adjusted sales data, which allows for accurate comparisons between two periods that would otherwise not possible to compare (e.g. retail sales in July 2018 and December 2018).

2. The fitted seasonality is the periodic, regular, repetitive variations in your data. In this example, there are significant peaks in sales in July of each year. You can also see increases in sales in September 2017 and September 2018. This could help you forecast sales for September 2019, and prepare you to increase your stock/staffing for this period.

3. The remainder is the ‘leftover’ component once trend and seasonality are removed from your data. The remainder is the ‘irregular component’ and shows us the anomalies in the data. In this example, Anna has identified that sales spiked in April 2019. This could be due to external factors, like a new product launch.

3. To highlight anomalies in your data that are not systematic, predictable, or in line with previous behaviour. 

Note: This analysis can be found in Discover and Missions.

Please refer to this article for information on anomaly detection.

Related articles
Unexpected changes

If you have any questions, please contact us at

Did this answer your question?