A third lens: Predictability
Two products can both show “demand rising.” One has sold steadily for months. The other sells a lot one week, nothing the next, then a lot again. Should you treat these two signals the same way? I don’t think so.
Most inventory tools look at two things. The first is demand direction: is the product gaining or losing interest? The second is stock coverage: how many days of stock are left? Both are useful, but both depend on one thing they don’t show, which is how reliable the sales pattern is.
I use a consistency score to check how steadily a product has sold across four consecutive periods (7, 14, 21 or 28 days). Based on the score, each product gets a label:
Predictable Sales: the pattern is steady, so the signals can be trusted.
Unpredictable Sales: the pattern is erratic, so the signals are less reliable.
New Products: there isn’t enough history yet to judge.
Predictability is a confidence score, not a separate metric
I don’t treat predictability as another number to watch. It changes how much you should trust the other two.
A rising demand signal on a steady seller can mean restock now.
The same signal on an erratic product should mean watch it closely.
The signal is the same, but the right action is different, because the confidence is different.
Why this matters in daily work
Without this, every signal looks equally urgent. You may over-order a product that had one lucky week, or ignore a real trend on a product you could have trusted. A confidence label helps you decide where to act fast and where to wait for one more period of data.
It won’t give exact answers. It tells you which signals deserve action first.
How I use this in StockSight
I built this idea into StockSight, a Shopify app I made for inventory and demand analysis. For each product it shows:
Demand direction: whether interest is increasing or decreasing, compared with the product’s expected demand.
Stock coverage: how many days your current stock will last.
Predictability: a label (Predictable, Unpredictable or New Product) based on a consistency score, plus the sales quantities of the last four consecutive periods (7, 14, 21 or 28 days), so you can see the pattern behind the label.
The goal is not to forecast the future with certainty. It helps you decide which products need action now and which ones need a closer look first.
You can try StockSight on the Shopify App Store: https://apps.shopify.com/stock-levels
A question for you
How do you handle products with erratic sales? Do you treat them differently from your steady sellers, or look at every product the same way? I’d like to hear how other merchants decide.
