Handling noisy SKU-level sales data in Shopify — how do you separate signal from noise?

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.