Can AI flag dead stock before it ages?

Mostly, yes, and the best place to flag it is before it is bought. Dead stock gives early signs in data a lumber yard or millwork business already holds. Weeks of cover rise because sales slow while the stock stays. All recent sales come from one customer or one job. The item was bought or made for a special order and the leftovers went to stock. A product line is being replaced. A check on those signals at the purchase order stops some dead stock from arriving. A monthly ranked list catches the rest months before the ageing report would, while it can still be returned, transferred or sold.

What the usual answer says

The usual answer says AI inventory tools analyse sales history, seasonality and demand patterns to catch slow movers, and recommend markdowns, transfers or buying changes. The better guides add ageing buckets and act early: one, from Marquis Data, says "Acting at the 91-day mark rather than the 365-day mark preserves both the carrying cost savings and the liquidation optionality", and puts a purchasing hold on items in the watch bucket. That is sound. It leaves out the signals particular to a lumber or millwork business.

What a late flag looks like

An ageing report flags an item once it has been dead for a while. From Quarri's own work with a lumber and millwork manufacturer: 34%+ of stocked SKUs had not sold in over twelve months.

At that point the options are few: write down, scrap, or sell at a loss.

Stop it at the purchase order

The cheapest dead stock to deal with is the stock never bought. In lumber and millwork, much of it starts with a purchasing decision: a full unit bought to fill one order, a mill-run lot, a special order's surplus put into stock. The signals are visible when the order is raised. When a buyer raises a replenishment order for an item whose cover is already rising, or whose recent demand came from one finished job, the system can say so before the order goes out. A special order can carry a rule too: buy the quantity the customer ordered, and put nothing extra into stock unless someone chooses to.

The early signals, for a lumber business

For stock already on hand, a model can rank every item monthly on four signals. Cover rising is on-hand quantity divided by the last three months' sales, in weeks, climbing while sales fall. One customer means all of an item's recent sales went to one customer or job, so the demand ends when the job does. Special-order origin means the item first came in for a special order, and in our experience that surplus is a frequent source of dead stock in millwork and specialty lumber. Replaced lines are items whose profile, finish or supplier has been superseded in the item master.

The list is only as good as the action taken on it: return to supplier, transfer to a branch that sells it, offer to the customer who bought it, or price to clear.

What predicts stockouts, and why it matters here

Research on the reverse problem shows which data carries the signal. Yang Liu and colleagues, in a 2025 study of a retailer's data covering over 1.6 million SKUs, found that "current inventory levels, short-term demand forecasts (three months), and recent sales data are the most influential factors in predicting stockouts". They add that "recent demand forecasts and sales data have greater predictive power than longer-term projections (six and nine months)".

Dead stock is the other side of the same balance: too much inventory against too little recent demand. We would expect the same near-term signals to point to it, but the study didn't test overstock, and it comes from retail rather than lumber.

Count before writing down

Some stock that looks dead may not be there at all. Yacine Rekik and colleagues, in a 2025 study of grocery stores, found that an inventory audit lifted store-wide sales 11%, all of it from items where the system showed more stock than there was. Grocery is not a lumber yard, and the effect was stronger on perishables. But the direction is a warning: before marking down or writing off an old item, count it.

When it doesn't apply

Stock held on purpose, for a seasonal peak or a contracted supply, will show high cover and is not dead. Very slow but steady sellers, such as specialty mouldings sold a few times a year, need a longer window than fast lines.

Quarri for wood products is built around how a wood products plant runs, where the order book meets real capacity.

Quarri is an AI-native data platform for the timber supply chain. It connects buying, production, sales and inventory for forest management, sawmill, wood products and pulp, paper and packaging operators.