Spark Inventory Blog

A reorder point can look perfectly sensible in a spreadsheet and still create the two outcomes operators are trying to avoid: stockouts on fast-moving SKUs and too much cash sitting in slow-moving ones. That is why reorder points fail for businesses selling across changing channels, managing incoming inventory, and buying from suppliers whose timing is not always reliable.

The problem is not that reorder points are inherently wrong. They are a useful control when demand is stable, lead times are predictable, and every unit is visible in one location. The problem is treating a static threshold as a purchasing decision when the operating conditions behind that threshold have changed.

A reorder point is an assumption, not a plan

The familiar formula is straightforward: expected demand during supplier lead time, plus safety stock. If a SKU sells 10 units per day, the supplier lead time is 30 days, and the business wants 50 units of buffer, the reorder point is 350 units.

That calculation answers a narrow question: when should available inventory trigger attention? It does not answer how much to buy, whether incoming inventory covers the gap, which warehouse needs the stock, or whether demand is still running at 10 units per day.

A reorder point becomes unreliable when its inputs are fixed but the business is not. Most product businesses do not operate in a fixed environment for long. A promotion changes sell-through. A wholesale order consumes stock that was expected to serve direct demand. A delayed container shifts the actual arrival date. A marketplace starts moving a SKU faster than the store does. The threshold remains unchanged while the decision it is meant to support becomes outdated.

Why reorder points fail in real operations

Demand is rarely stable at the SKU level

An average daily sales rate can hide a meaningful change. A SKU that averaged 10 units per day over the last 90 days may now be selling 4 units per day after a campaign ends, or 18 units per day after a retail account begins ordering. Reordering against the old average either adds excess inventory or waits too long.

This is particularly common with long-tail catalogs. Portfolio-level demand may appear predictable, but the per-SKU forecast can vary sharply because each item has a different sales pattern, lifecycle stage, and channel mix. A single average is not a demand forecast.

Seasonality creates the same problem on a larger scale. If a business raises reorder points ahead of peak season but leaves them elevated afterward, it can keep buying at a pace the current market no longer supports. If it does not raise them early enough, it discovers the shortage only after supplier lead time makes correction impossible.

Lead time is a range, not one number

Many teams load a standard supplier lead time into their inventory system and use it for every purchase order. That number may represent factory production time, or it may include transit, receiving, and putaway. Often it is unclear which.

The operational lead time is the period from approving a purchase order to inventory being available to fulfill demand. It can include supplier confirmation, production, freight booking, ocean or ground transit, customs, appointment scheduling, receiving, and quality checks. Any one of those stages can move.

When a reorder point assumes 30 days and the next order takes 45, the safety stock is doing work it was not designed to do. Increasing the threshold may mask the issue, but it also ties up more working capital across every normal order. A better approach is to track actual supplier lead time and treat delayed incoming inventory as a current stock-risk condition, not a reason to permanently inflate every buffer.

On-hand inventory is not the same as usable inventory

A reorder point often triggers from an on-hand balance. That is incomplete. Some of that stock may be committed to open sales orders, allocated to a wholesale customer, held for a retail location, in transfer between warehouses, or unavailable due to receiving and quality status.

The relevant question is usually inventory position: what is available now, what is incoming, what is committed, and what demand is expected before the next replenishment can be used. A buying decision based only on on-hand units can make a healthy SKU look short or a constrained SKU look safe.

This gets harder with multiple warehouses. Company-wide inventory can be sufficient while the warehouse serving the next orders has too little stock. Transferring units may be faster and cheaper than buying more, but a static reorder point does not evaluate that choice.

Incoming inventory is either ignored or trusted too much

Some reorder systems ignore open purchase orders, which leads to duplicate buying. Others count every incoming unit as available regardless of its expected arrival date, which can delay a necessary reorder.

Incoming inventory should be time-phased. An order arriving next week can cover near-term demand. An order with an uncertain arrival date three weeks after a stockout cannot. The same is true for partial shipments. If 200 units are confirmed for receipt but the remaining 800 are still in production, treating all 1,000 units as one reliable supply event gives the buyer false confidence.

The reorder point does not decide order quantity

Crossing a threshold tells a buyer that something happened. It does not produce a defensible quantity. Ordering up to a fixed maximum can work for simple, frequent replenishment, but it breaks down when suppliers have minimum order quantities, case packs, price breaks, order calendars, or limited production capacity.

A useful reorder recommendation needs to consider forecasted demand over the coverage period, current and incoming supply, desired days of supply, supplier constraints, and the next realistic opportunity to buy. The target is not simply to get above a line. It is to carry enough inventory through the next replenishment cycle without funding unnecessary stock.

Replace static triggers with a time-phased buying plan

The practical replacement for a standalone reorder point is not more complicated math for its own sake. It is a regularly refreshed buying plan built from current operating inputs.

For each SKU, start with a per-SKU forecast rather than a fixed historical average. Forecasts are still estimates, so they should be reviewed against known events: promotions, new account launches, discontinued products, stockout-distorted history, and product substitutions. The goal is not to pretend demand is certain. It is to make the uncertainty visible and manage it deliberately.

Next, calculate inventory position by date. Include usable on-hand inventory, expected receipts, open commitments, and demand across the period until new supply is available. This turns a single point-in-time balance into days of supply and a projected stockout date. Buyers can then see whether a risk is immediate, whether incoming inventory resolves it, and whether there is time to change the plan.

Finally, set the reorder quantity against an explicit coverage target. That target depends on the SKU and supplier. A high-volume item from a slow overseas supplier may justify more coverage than a low-volume accessory that can be replenished domestically in a week. There is no universal right buffer. The right buffer reflects demand variability, supply reliability, service expectations, and the working capital the business is willing to commit.

Make exceptions visible before they become urgent

A buying plan should direct attention to the decisions that need operator judgment. Not every SKU requires the same review. An item with stable demand, reliable supply, and sufficient incoming inventory may need no action. A SKU projected to stock out before its next receipt needs immediate attention, even if it has not crossed a traditional reorder point.

The most useful exception views focus on stock risk, excess inventory risk, overdue incoming supply, unusual demand changes, and recommendations that conflict with supplier constraints. These are the conditions where a buyer can change an outcome: expedite, transfer stock, adjust an order quantity, delay a purchase, or revise a forecast assumption.

This is also where a disciplined system earns its place. Spark combines sales history, on-hand and incoming inventory, supplier lead times, purchasing data, warehouse context, and inventory policies to prepare reorder recommendations and draft purchase orders. The operator reviews, adjusts, and approves the draft before anything is sent or committed.

Keep reorder points as guardrails, not autopilot

Reorder points can still be useful as a simple alert or a fallback control for predictable SKUs. They are less useful as the system that plans every purchase. The more channels, warehouses, supplier variability, and SKU-level demand changes a business manages, the more a static threshold becomes a lagging signal.

Treat the reorder point as the start of a question: will inventory cover expected demand until usable supply arrives, and what action protects availability without creating excess? A current buying plan gives the operator a better answer, while leaving the decision where it belongs: with the person accountable for the inventory and the cash behind it.