Economic Order Quantity: Calculate It, Then Test It Against MOQ

The economic order quantity formula is Q = √(2DS ÷ H). D is annual demand in units, S is the cost of placing one order and H is the cost of holding one unit for a year. In Excel, =SQRT(2*D*S/H) returns the order size that minimises yearly ordering plus holding cost.

Ford W. Harris developed the model in 1913 (Wikipedia: Economic order quantity). It assumes demand is constant, each order arrives in full when stock reaches zero, the cost per order is fixed whatever the quantity, and each order covers one item.

The logic is simple. Ordering cost per year is D ÷ Q × S, which falls as orders get bigger. Holding cost per year is Q ÷ 2 × H, because stock falls from Q to zero and averages half of Q. The total is lowest where the two are equal.

Fill in five inputs and the sheet returns the EOQ, how often to order and what the order costs you per year. The example below uses a stock-keeping unit (SKU), meaning one distinct item you stock, with illustrative inputs of ours.

Our example: demand D is 12,000 units a year, landed unit cost is ₹250, the holding rate is 24% of unit cost, and the cost per order S is ₹1,800.

At the EOQ, ordering cost is 14.14 × ₹1,800 and holding cost is 424.3 × ₹60. Both come to about ₹25,456, which is why the total is ₹50,912. Rows B10 to B14 are the MOQ test. A 1,500-unit minimum raises yearly cost from ₹50,912 to ₹59,400, an extra ₹8,488 (16.7%), which a discount of 0.28% on ₹30 lakh of purchases would cover.

Cost per order is everything that happens once per purchase order, whatever its size. Holding cost per unit is unit cost times a yearly rate covering capital, storage and losses. Estimate each from your own books, because the formula only repeats back what you feed it.

Cost per order (S). List the steps of one purchase and cost each. Our illustrative ₹1,800 is: buyer time ₹350, approval ₹150, the fixed part of freight and unloading per delivery ₹800, receiving and the goods receipt note (the record that confirms what arrived) ₹300, and invoice matching and payment ₹200. Freight that rises with quantity does not belong in S, and the landed unit cost (price plus duty, freight and handling) belongs in B2. Our landed cost calculation page shows how to build it.

Holding rate (H ÷ unit cost). Our illustrative 24% is capital 10% (₹25), storage 8% (₹20) and insurance, shrinkage and obsolescence 6% (₹15), which sums to ₹60 per unit. Take the capital rate from your own cash credit or working capital limit, as stated in the sanction letter (the bank's written approval of your limit and rate). A cash credit limit is a bank facility you draw on as needed. For context, the Reserve Bank of India raised its policy repo rate by 25 basis points to 5.50% on 7 October 2026, and the rate you actually pay is set by your bank.

To check storage, multiply rent by the floor area each unit occupies, including its share of aisles. Weighted average transacted warehouse rents in 2025 ran from ₹18.6 per sq ft per month in Ahmedabad to ₹28.0 in Pune, with Bangalore at ₹23.2 and Hyderabad at ₹21.5 (Knight Frank, India Warehousing Market Report 2025). These are weighted averages of transactions Knight Frank recorded in 2025, so your godown may differ. At Hyderabad's ₹21.5, a year costs ₹258 per sq ft (21.5 × 12). A unit that takes 0.08 sq ft costs about ₹20.6 a year to store, which matches the ₹20 above.

Demand (D). Use the units you expect to buy over the next twelve months. If the item is growing or fading, a forecast is better than last year's consumption. Use history only when demand has been steady.

Not very. Total cost at any order size Q, relative to the minimum, is ½ × (EOQ ÷ Q + Q ÷ EOQ). This is an algebraic identity, so you can verify our table with a calculator.

The EOQ moves with the square root of S and H. If you misjudge either one by a factor of two, the order size is off by about 1.41 times and total cost rises by only 6.1%. A factor of four wrong adds 25.0%.

We also recalculated the example with different inputs. Holding rates of 15%, 24% and 35% give EOQs of 1,073, 849 and 703 units. A cost per order anywhere from ₹600 to ₹5,000 gives an EOQ between 490 and 1,414 units. An eight-fold range in ordering cost moves the answer by less than three times.

Our recommendation is to measure S and H to within a factor of two and stop refining them. The flat curve is a property of the formula. NC State's tutorial makes a related point (2011): EOQ gives a good indication of whether current order quantities are reasonable even when its assumptions hold only roughly. Recalculate when demand, supplier terms or your borrowing rate change by a lot. A 20% shift in order size costs under 2%.

Our rule of thumb: order the MOQ if it is within about 1.3 times the EOQ, because the penalty is then about 3.5% or less (see the table of order sizes above). Beyond that, the extra holding cost grows quickly, so compare it with the discount you are being offered. In our example, 1.3 × 849 is about 1,104 units, where the penalty is still under 4%.

The table below holds the example's inputs fixed: 12,000 units, ₹250 landed unit cost, ₹60 holding cost, ₹1,800 per order, and ₹30 lakh (₹30,00,000) of yearly purchases. The last column is the discount on all 12,000 units that would pay for the extra stock.

The break-even is the extra cost divided by yearly purchases. It is approximate, because cheaper stock also costs a little less to hold. If the supplier offers more than the break-even and you have the cash, shelf space and shelf life for the extra stock, take it. If not, negotiate. Options are a smaller minimum, or a blanket purchase order, where one agreement covers deliveries released over time, if the supplier agrees.

Round the final quantity to the carton, pallet or truck. If cartons hold 50 units, order 850 (17 cartons), not 849.

Real orders face limits the formula ignores, such as the supplier's production run, your credit limit, a design revision mid-run and truck capacity. Treat each as a constraint on Q, then cost the nearest feasible quantity in the sheet.

The EOQ stops working when its assumptions fail: demand is not steady, the price changes with quantity, several items share one order or truck, or stock can expire. The model assumes constant demand, full-lot delivery and one item per order (Wikipedia).

Our approach in each case:

• Erratic or seasonal demand: use a forecast for D and recalculate each season.
• Expiry dates: cap the order at what you sell before the date, even if the EOQ is higher.
• Price breaks: cost each break quantity in the sheet, as in the MOQ table above.
• Timing: the EOQ says how much, not when. Use the reorder point formula, which includes safety stock (extra stock held against delays or demand spikes), for when.

We handle the step after the sheet: raising the right orders without double-ordering. In Arka Inventory you can plan purchases from real orders, forecasts and stock, and schedule purchase requisitions or purchase orders, with stock across locations and inventory on order in view.

Low-stock and back-order alerts come from our Basic plan upwards. Landed cost tracking, which gives the landed unit cost your sheet needs, is on our Enterprise plan. Arka Inventory is cloud software built natively on Salesforce. As of October 2026, Basic costs $199 a month and Advance $499 a month, billed yearly, and Enterprise pricing is flexible; see Arka Inventory pricing.

To see your purchasing in Arka Inventory, request a demo on our contact sales page and ask which plan fits.