Busy or broken? Telling festive seasons and 11.11 apart from real problems
Every business in Singapore and Malaysia lives by a calendar that does not care about averages. Sales jump before Chinese New Year and drop sharply during it. Hari Raya, Deepavali and Christmas each bring their own rush. Online stores spike on 11.11 and 12.12. Payday week is busier than the week before it.
This creates a real problem for anyone watching their numbers. When everything moves, how do you tell a normal seasonal swing from a real problem? If you treat every swing as a warning, you drown in false alarms. If you shrug off every swing as "seasonal", a real problem can hide inside it.
This post is about comparing fairly.
The wrong comparison: last week
The most common comparison is "this week against last week". It is easy, and it is often wrong:
- The week of Chinese New Year will look like a disaster against the week before, which was the pre-festival rush.
- The week after 11.11 will look terrible for an online store, because 11.11 was in last week.
- A payday week will look like strong growth against the quiet week before.
None of those tells you anything about how the business is doing. They only tell you where you are in the calendar.
Better comparisons
Same weekday. Compare a Saturday with recent Saturdays, not with Friday. Weekday patterns are strong in almost every business, from restaurants to wholesalers.
Same period last year. For anything tied to festivals, compare with the same festival last year. Chinese New Year, Hari Raya and Deepavali move around the Western calendar, so "the same week last year" may land on a completely different part of the season. Line up the festival itself, for example "the 10 days before Chinese New Year this year against the 10 days before Chinese New Year last year".
Same position in the month. Payday effects mean the last days of a month should be compared with the last days of other months.
Growth-adjusted. If the business has grown 15% over the year, last year's festival is a rough baseline, not a target.
A simple way to do it yourself
If you keep your sales in a database or a spreadsheet, a useful habit is a small table of the key dates for your business, each with the matching date last year:
| Event | This year | Last year |
|---|---|---|
| Chinese New Year (day 1) | your date | your date |
| Hari Raya Puasa | your date | your date |
| Deepavali | your date | your date |
| 11.11 | 11 Nov | 11 Nov |
| School holidays start | your date | your date |
Then compare days relative to the event (5 days before, the day itself, 3 days after) instead of calendar dates. It takes an hour to set up and makes every festive comparison much fairer.
The hard part: problems that hide inside the rush
The reason this matters is not only false alarms. Real problems often happen during busy periods, precisely because everyone is busy:
- Stock-outs. A supplier who cannot keep up with festive demand. Sales of that line look "a bit down", which is easy to put down to the season, when in fact the shelf was empty.
- Till and payment problems. A card terminal going down on the busiest Saturday of the year costs more than on any other day, and is easy to miss in the noise.
- Pricing mistakes. Promotions are set up in a hurry before festivals. A wrong price or a discount on the wrong product can run for days.
- Shrinkage. Busy periods with temporary staff are when voids and refunds are least watched.
So the goal is not to ignore everything during festive seasons. It is to compare against the right normal, so that what is still unusual stands out.
What "learning the pattern" means
A system that watches your data properly should learn these rhythms by itself, from your own history:
- It notices that something keeps happening: month end is always busy, sales drop for three days at Chinese New Year, a product always sells out before Deepavali.
- It turns that into a known pattern, so that the same swing next time is expected instead of alarming.
- It keeps looking for what is unusual even inside the pattern: if this Chinese New Year's drop is twice as deep as usual for one product line, that still gets flagged.
This is how an experienced store manager thinks. They are not surprised by the festive rush, but they notice when this year's rush feels different.
How Cuedis handles it
Cuedis learns patterns like these from your own history. Things that keep happening, like a busy month end or a festival dip, become known patterns, and it stops raising alarms about them. It also knows the public holidays where your business is, so festive days are compared with festive days. Inside those patterns it keeps watching for what is still unusual, and it groups many alarms on one day into a single incident when they have one cause.
Whether you use a tool or a spreadsheet, the principle is the same: busy is not a problem, and quiet is not a problem. Different from what this time of year usually looks like is.