Inventory Prediction for WordPress Stores: How the Forecast Is Actually Calculated
Not a black box: here is the exact math behind KasbPro’s stock forecast for WooCommerce stores — velocity, trend, days of cover — and the one assumption it makes for every product today.
"AI-powered inventory forecasting" is a phrase used loosely enough that it is worth showing the actual arithmetic instead of asserting it. This is exactly what KasbPro computes from a WooCommerce store's sales history, in plain terms, including the one simplifying assumption it makes for every product.
The starting point: velocity, not a unit count
WooCommerce shows you a stock number. That number means nothing on its own — 15 units is two weeks of runway at one sale a day, and gone by tomorrow at ten a day. The first thing the forecast computes is daily velocity: units sold over the last 30 days, divided by 30, per product.
Trend: is it speeding up or slowing down
Velocity alone treats every product as flat. The forecast also compares the last 30 days against the 30 days before that, producing a trend ratio — above 1 means demand is growing, below 1 means it is declining. That ratio is capped between 0.5× and 2× so one unusually large or quiet week does not swing the forecast to an extreme.
Turning that into a forecast
The 30/60/90-day forecast is: daily velocity × number of days × trend, rounded. It is deliberately simple — a straight-line projection adjusted for recent direction, not a machine-learning model trying to detect seasonality or external events. It cannot know about a planned promotion, a competitor closing, or a first-time holiday spike, because none of that exists in the sales history it is computing from.
Days of cover and the stockout date
Days of cover is current stock divided by daily velocity — literally, how many days until you run out at the current pace. From that, a projected stockout date is calculated directly (today plus days of cover). This is the single most useful number on the dashboard, because it converts an abstract unit count into a date you can act on.
The reorder quantity, and the one fixed assumption
The recommended reorder quantity accounts for three things: demand during your supplier's lead time, demand during a review cycle (how often you typically reorder), and a safety buffer of half a lead time's worth of demand, minus what you already have in stock.
Here is the honest limitation: today, lead time defaults to 7 days and review cycle to 14 days for every product, uniformly. There is no per-product or per-supplier lead time setting in the app yet — a product from a supplier that takes three weeks to deliver is forecast with the same 7-day assumption as one you can get overnight. If your suppliers have meaningfully different lead times, treat the reorder quantity as a reasonable default to sanity-check against what you actually know about that specific supplier, not a number tuned to them individually.
What "dead" and "healthy" actually mean
The forecast also classifies each product — reorder, low, healthy, overstock, or dead — from the same numbers. A product with no sale in 90+ days and stock still on hand is flagged dead; one with far more days of cover than you would realistically sell through is flagged overstock. This classification system is covered in more depth in the companion post below, because it answers a different question than forecasting does: not "how much will I sell" but "what should I stop ordering."
Where this genuinely helps, and where it does not
It replaces guesswork with arithmetic over your own real sales data — that is a real improvement over "it feels like we're low." It is not a substitute for knowing your business: a founder who knows a promotion is coming next month should order ahead of what the forecast says, because the forecast has no way to know that yet.
What is free
The velocity calculation, trend, forecasts, days of cover, and reorder quantities are all on the free plan — this is arithmetic over data KasbPro already has, with no cost to compute. The Pro plan adds an LLM-narrated weekly summary of these same numbers in plain English inside wp-admin; the underlying numbers themselves cost nothing extra.
Related reading: the AI behind your inventory classification and from forecast to purchase order.