How to Find Trends in Sales Data (Seasonality, Growth, Mix)

To find real trends in sales data, you need to separate three things that are usually tangled together: underlying growth, seasonality (the pattern that repeats every year) and mix (changes in what you sell or who buys it). Compare periods like with like, smooth out the noise, and break totals into their parts. This guide shows how, with simple methods you can do in a spreadsheet.

Start with the right shape of data

You need transaction-level or at least daily data with a date, amount, and the dimensions you care about (product, category, branch, channel, customer). Two or more years is ideal for seasonality; at minimum, 12 months. Clean it first: duplicates and wrong dates create fake trends. The data cleaning checklist covers what to fix.

1. Compare like with like

Month-on-month comparisons mislead in seasonal businesses. December is almost always bigger than November in retail; January is often weak. Year-on-year (this March vs last March) removes most seasonality.

In this illustrative example, month-on-month says March was a great month. Year-on-year tells a different story: growth has been slowing from about 9% to under 1%. That is the trend worth talking about.

2. Smooth the noise with moving averages

Daily and weekly sales jump around. A moving average shows the direction. A 7-day moving average removes the weekday pattern; a 3-month or 12-month rolling total shows longer direction. In Excel, if daily sales are in column B starting at B2, the 7-day average in C8 is =AVERAGE(B2:B8), filled down.

A 12-month rolling total is particularly useful: each point is the sum of the last 12 months, so seasonality is fully cancelled and you see pure growth or decline.

3. Measure seasonality explicitly

A simple seasonal index tells you how each month compares to an average month. For each month, divide that month's sales by the average monthly sales for the year, and average across years if you have several.

• If average monthly sales are R350,000 and December is R525,000, December's index is 1.5.
• If January is R280,000, its index is 0.8.
• To judge whether this January was good, compare actual sales with expected sales: average month × 0.8.

This also makes targets fairer. A flat monthly target sets people up to fail in January and coast in December.

4. Break growth into volume, price and mix

Revenue = number of orders × average order value. Or, per product: units × price. When revenue grows 10%, ask which part moved.

• Volume: more orders or more units sold.
• Price: the same items at higher prices (including inflation).
• Mix: a shift toward more expensive (or cheaper) products, customers or channels.

Example: say revenue rose from R1.2 million to R1.32 million (+10%). Orders went from 4,000 to 3,800 (−5%), but average order value rose from R300 to about R347 (+16%). Growth came from bigger baskets, not more customers. If your price increase was 6%, the rest came from mix: customers buying more premium items. That is a very different business story from ‘sales are up 10%’.

5. Slice by segment to find where the trend lives

A flat total can hide one segment growing fast and another shrinking. Look at the same trend by branch, channel (online vs in-store), product category and customer type. Rank segments by absolute change in rand, not just percentage; a 60% jump on a tiny product line matters less than a 5% drop in your biggest category.

6. Separate trends from one-off events

A single large order, a promotion, a stock-out or a disruption such as load shedding can look like a trend in a short window. Mark known events in your data (a ‘promo’ or ‘event’ column helps) and check whether the change persists afterwards. For one approach to measuring disruption, see analysing load-shedding’s impact on sales. Outlier days are covered in spotting anomalies in business data.

Watch margin, not just revenue

A revenue trend can look healthy while profit quietly erodes. If your data includes cost of sales, run the same year-on-year and mix analysis on gross margin. Say discounting grew your top line by 7% but pushed gross margin from 38% to 33%: on R4 million of sales, that is roughly R150,000 less gross profit than you would have made at the old margin and the new volume. Trends in margin by category often reveal more about the health of the business than trends in sales alone.

Putting it together

1. Clean the data and confirm totals reconcile.
2. Chart a 12-month rolling total to see the true direction.
3. Compare year-on-year by month to see if growth is speeding up or slowing.
4. Split growth into volume, price and mix.
5. Find which segments drive the change, in rand.
6. Rule out one-off events before calling it a trend.

Summarix does the computational part of this automatically for uploaded files and connected sources, and you can ask follow-up questions like ‘which category drove the growth?’. See features, and our guide to sales KPIs for which numbers to track each month.