Reorder point formula
Reorder point ROP = d_bar * L + Z * sigma_LT, where d_bar is mean daily demand, L is lead time in days, and Z * sigma_LT is safety stock at the chosen service level[1]. Triggers a new order when on-hand falls to ROP. Independent of EOQ Q*, which sets the order quantity once triggered.
The on-hand level at which a new order is triggered.
(Q, r) vs (s, S) policy
(Q, r): when on-hand hits r (= ROP), order Q (= Q*). Continuous review. Standard for A-class items in inventory software (NetSuite, Fishbowl, Cin7).
(s, S): at fixed review intervals, if on-hand <= s, order up to S. Periodic review. Common for C-class items where continuous monitoring isn’t justified. s plays a similar role to ROP; S - s is bounded above by Q*[2].
ROP and Q* independence
Q* answers "how much to order each time?" ROP answers "when to trigger?" They depend on different inputs (Q* on D, S, H; ROP on d_bar, L, sigma_LT, Z) and are optimised separately. A common confusion is to set ROP = Q*: that’s only coincidentally right. Set them independently and the policy works correctly.
If lead time is itself variable
Combined variance sigma_DLT^2 = L * sigma_D^2 + d_bar^2 * sigma_LT^2. Use sigma_DLT in place of sigma_LT. See lead-time variabilityfor the full derivation.