Reading return across Chinese New Year, Ramadan, and Deepavali weeks

A year in Malaysia is not a smooth sine wave with a dip in December. Chinese New Year travel, Ramadan and Hari Raya, Deepavali, state holidays, and the exam calendar for anything adjacent to schools will move a habit more than a new onboarding screen will.

We ask teams to keep a small calendar beside the canvas: the weeks they already know will distort return. Then we cut those weeks as their own cohorts instead of blending them into a monthly average. A meditation product that looks “down 18% in January” may be looking at people who were in hometowns with different mornings, sleeping in other rooms, sharing bathrooms, and not failing the habit in any sense that a product meeting should punish.

Ramadan is the sharpest example for anything tied to eating, prayer, sleep, or evening social life. If your product asks for a 5:30 run, you already know the story. If your product asks for ten minutes of quiet in the afternoon, you may see the opposite — more completion, not less. Average them together and you will ship a reminder change that helps nobody.

Deepavali weeks in some of the Tamil-language drill products we have read show a pattern closer to a festival of visits than to a holiday off: short sessions survive, long sessions do not. That is a ritual change, not a retention failure. Name it.

You do not need a special method. You need the honesty to put the calendar on the same page as the table. Teams that refuse, because a global template does not have a column for “Raya week,” will keep rediscovering Malaysia every year.

If this note names the argument in your team, send the ritual and the week people vanish.