I’ve been working with Shopify sales data at the SKU level and keep running into the same challenge:
daily sales are extremely noisy, especially for low-volume or long-tail products.
Some patterns I’ve seen:
-
Daily spikes that don’t represent real demand shifts
-
SKUs with intermittent sales (0, 1, 0, 2, 0…)
-
Weekly aggregation hides early trend changes
-
Monthly aggregation reacts too late for restocking decisions
I’m curious how others handle this in practice:
-
Do you smooth daily data (moving averages, rolling windows, etc.)?
-
Do you aggregate weekly by default?
-
Do you treat low-volume SKUs differently?
-
How do you avoid overreacting to short-term spikes?
Not looking for a “best” answer — genuinely interested in real-world approaches others are using with Shopify data.