MRP vs DRP: Planning Production vs Distribution Inventory

The plant hit its production schedule every week of the quarter, yet the Guwahati depot ran out of stock twice while Bhiwandi's central warehouse had pallets of the same SKU to spare. The production planners pointed to their MRP run, which showed every component order on time. The distribution team pointed to their spreadsheet of depot reorder points. Neither plan was wrong on its own terms; the two simply never talked to each other.

Material requirements planning and distribution requirements planning use the same arithmetic but point it in opposite directions. Once you see where one hands off to the other, you can decide which a business needs, size lots without starving the next level, and price the hidden costs of moving stock between depots, including the GST that Indian inter-state transfers attract.

• What this guide covers: the direction of each plan, the records each reads and writes, the shared netting logic, worked DRP and MRP grids, lot-size effects, and how to choose an operating model.
• What it adds: a two-depot DRP grid in units, a cost ledger for the same transfer plan in a US network (USD) and an Indian network (INR), and a checklist for setting up either plan.

MRP explodes a bill of materials; DRP pulls demand up a network

MRP works downward from a finished-goods schedule through the bill of materials to components; DRP works upward from the points where customers buy, through the distribution network, to the warehouse or plant that supplies it. The arithmetic is identical; the tree it walks is different.

MRP starts with a master production schedule: how many finished units the plant will make in each week. It multiplies that schedule by the quantity of each component in the bill of materials, nets the result against component stock and open purchase orders, and offsets the remainder by supplier or production lead time. The outputs are planned purchase orders and work orders, each with a due date and a release date.

• Kind of demand: MRP plans dependent demand, calculated from a parent schedule; DRP plans independent demand, forecast from customers at each stocking point.
• Structure walked: MRP follows the bill of materials; DRP follows sourcing rules that say which location supplies which.
• Lead time used: MRP uses purchasing and manufacturing lead times; DRP uses lane transit time, including loading, dispatch and receiving.
• Output document: MRP proposes purchase and work orders; DRP proposes transfer orders and, at the top of the network, purchase or production requirements.

DRP starts with demand at the edge: each depot's forecast and open customer orders. It nets that against depot stock and transfers already in transit, offsets the shortfall by the transit lead time of the lane that supplies the depot, and produces planned transfer orders. Those planned transfers become the demand of the supplying warehouse, and the process repeats one level up until it reaches a plant or a supplier.

Records each plan reads and who maintains them

Both plans fail quietly when their master data is stale, and the fields they depend on belong to different teams. Assign each record an owner before the first run, or the planning engine will turn yesterday's assumptions into tomorrow's orders.

MRP depends on engineering and purchasing data: the bill of materials, scrap factors, supplier lead times and order minimums. DRP depends on network data: which location supplies which, how long each lane takes, the pallet or truck multiple for a transfer, and the safety stock held at each depot. A single wrong transit time can matter as much as a wrong bill of materials quantity.

The netting arithmetic MRP and DRP have in common

Every week in the planning horizon, both plans project the stock balance, compare it with safety stock, and create a planned receipt when the balance would fall short. The planned receipt is then moved back in time by the lead time to give a release date.

Planners usually lay this out as a grid with one column per week. Gross requirements are the demand to be met; scheduled receipts are orders already released, such as transfers on the road; the projected available balance is what would be on hand at week end; planned receipts fill the gap; planned releases are those receipts shifted earlier by lead time. The code block defines each row.

-- Weekly netting for one item at one location (DRP depot or MRP component)
-- t        : week number, 1 = current week
-- GR[t]    : gross requirement (depot forecast + open orders, or parent schedule x BOM qty)
-- SR[t]    : scheduled receipt already released (transfer in transit, open PO)
-- SS       : safety stock;  Q : lot size multiple;  LT : lead time in weeks
-- PAB[0]   : current on-hand balance

for t in 1..H:
  PAB_pre  = PAB[t-1] + SR[t] - GR[t]
  if PAB_pre < SS:
      PR[t] = Q * ceil((SS - PAB_pre) / Q)    -- planned receipt in lot multiples
  else:
      PR[t] = 0
  PAB[t]   = PAB_pre + PR[t]
  PRel[t - LT] = PR[t]                        -- planned release; t - LT < 1 means past due

-- DRP roll-up: GR_central[t] = sum over depots d of PRel_d[t]
-- MRP explosion: GR_component[t] = PRel_finished[t] * BOM_qty

The only differences between the two uses are where gross requirements come from and which lead time is applied. For a depot, gross requirements are its own forecast and lead time is the lane transit time. For a component, gross requirements are the parent item's planned releases multiplied by the bill of materials quantity, and lead time is the supplier's or the plant's.

Distribution requirements planning nets each depot's weekly forecast against on-hand stock, in-transit transfers and safety stock, rounds any shortfall up to the transfer lot multiple, and offsets the release by lane transit time; a Guwahati depot with a two-week lane releases in week 1 for a week 3 need.

A worked DRP grid for two depots and one central warehouse

Two depots with steady forecasts produce a central demand pattern that is anything but steady. The worked grid below shows why, using pallet-sized transfers and two lanes with different transit times.

The network is a central warehouse at Bhiwandi supplying a depot at Nagpur and a depot at Guwahati. Nagpur forecasts 120 units a week, holds 260 on hand, keeps 100 as safety stock and receives transfers in lots of 240 units with a one-week lane. Guwahati forecasts 90 units a week, holds 300, keeps 60 as safety stock and receives lots of 180 with a two-week lane. All figures are illustrative.

Customers buy 210 units a week across the two depots, every week. The central warehouse, however, sees 420 units in weeks 1 and 3, nothing in weeks 2, 4 and 6, and 240 in week 5. That pattern is created entirely by lot rounding and lead-time offsets, and it is what the central location must plan against.

Guwahati's grid also shows why the lane time matters. Its week 3 receipt has to leave Bhiwandi in week 1, so the transfer order must be released this week. If the transit time in the master data said one week instead of two, the same need would appear as a week 2 release, and the depot would fall below its safety stock before the truck arrived.

Where DRP hands its demand to MRP at the central level

The central warehouse's lumpy gross requirements become the input to the master production schedule, and from there MRP explodes them into components. Lot sizes chosen at the depots therefore travel all the way to the purchasing desk.

Suppose the Bhiwandi warehouse is supplied by an adjacent plant with a one-week production lead time, a batch size of 1,000 units, 700 units on hand and safety stock of 200. Netting the central gross requirements gives 280 at the end of week 1 and again in week 2. In week 3 the 420-unit requirement would leave minus 140, so a 1,000-unit batch is needed in week 3 and must be released to production in week 2.

• Central projection: 700 on hand, then 280 in weeks 1 and 2, 860 in weeks 3 and 4 after the batch arrives, and 620 in weeks 5 and 6.
• Master schedule: one batch of 1,000 units released in week 2 for receipt in week 3.
• Carton explosion: each unit needs one printed carton, so MRP sees a gross requirement of 1,000 cartons in week 2, when production starts.
• Carton netting: 600 cartons are on hand, leaving a shortfall of 400; with a two-week supplier lead time, the purchase order should have been released in week 0.

The result is a past-due planned purchase order on the first run, even though every depot is comfortably above safety stock. Nothing in the depot plan was wrong; the combination of a pallet-sized transfer at two depots, a 1,000-unit batch and a two-week carton lead time created a need the supplier cannot meet in time.

The trace below shows how that exception appears and how a planner resolves it without overriding the arithmetic.

1. Mon 05:30, DRP run: planned transfers TR-88231 (240 units to Nagpur) and TR-88232 (180 units to Guwahati) are created with week 1 release dates.
2. Mon 05:50, master schedule: the central projection falls to minus 140 in week 3, so planned production order PP-41907 for 1,000 units is proposed for week 2 release.
3. Mon 06:10, MRP run: the explosion of PP-41907 needs 1,000 cartons; with 600 on hand and a two-week lead time, planned purchase order PR-77310 for 2,000 cartons, the supplier minimum, shows a release date in the past.
4. Mon 09:15, buyer review: the carton supplier confirms it can deliver 2,000 cartons in ten working days, which lands in week 3, not week 2.
5. Mon 11:40, planner decision: PP-41907 is split and firmed as 600 units in week 2 and 400 in week 3; the central balance in week 3 is 460, above safety stock, and PR-77310 is released today for week 3.

Lot sizes and lead times behave differently in each plan

In DRP, lot sizes are set by transport economics such as pallet layers, full pallets and truckloads; in MRP, they are set by changeovers, supplier minimums and batch sizes. Both kinds amplify variation as they move up the chain, so they should be chosen together.

• Transfer lot size: a bigger transfer lowers freight per unit but raises depot stock and makes central demand lumpier; it should rarely exceed two to three weeks of depot demand.
• Production batch: a large batch lowers changeover cost per unit, but it converts lumpy central demand into even larger component orders.
• Transit lead time: DRP lead times are usually short and fairly stable, but a one-week error on a long lane moves every release on that lane.
• Supplier lead time: MRP lead times are often longer and less reliable, which is why the past-due exception in the trace appeared at component level.
• Time fences: firm planned orders inside the near-term horizon, so a fresh forecast does not reshuffle trucks and batches that are already being prepared.

A useful test is to ask what the central warehouse would see if every depot ordered weekly in exact forecast quantities. If the answer is a smooth 210 units a week and the actual pattern swings from zero to 420, the lot sizes, not demand, are driving the variation. That is usually cheaper to fix at the depot than at the plant.

DRP lot sizing at depots sets the shape of central demand: two depots forecasting 120 and 90 units a week, with 240 and 180 unit transfer lots, produce central gross requirements of 420, 0, 420, 0, 240 and 0 over six weeks, which the master production schedule then turns into a single 1,000-unit batch.

US and Indian networks: the same transfer plan, different costs

The DRP arithmetic is the same in both countries, but the cost of acting on it is not. In India a transfer between GST registrations is a taxable supply with its own invoice and e-way bill, while a US company moving its own stock between states raises no federal tax document.

Under Section 25(4) and 25(5) of the CGST Act, each GST registration, and establishments in different States or Union territories, are treated as distinct persons. Schedule I paragraph 2 makes a supply between distinct persons in the course of business a supply even without consideration. Rule 28(1) values it at open market value, but its second proviso deems the invoice value to be open market value where the recipient is eligible for full input tax credit.

• Bhiwandi to Nagpur: both in Maharashtra; if they share one registration, the transfer is not a supply, but a movement above ₹50,000 still needs an e-way bill under Rule 138(1), which covers movement for reasons other than supply.
• Bhiwandi to Guwahati: Maharashtra to Assam, so a tax invoice with IGST is raised by the Maharashtra registration and the Assam registration takes the input tax credit.
• E-way bill validity: Rule 138(10) allows one day for the first 200 km and one more for each further 200 km, so a lane of roughly 2,700 km carries 14 days of validity.
• United States: there is no federal value-added or goods and services tax, so an inter-state transfer needs shipping documents but no federal tax invoice; state tax questions are outside this example.

The ledger applies the six-week transfer plan from the DRP grid to an Indian network and to a matching US network, Dallas to Phoenix on a one-week lane and Dallas to Seattle on a two-week lane. Unit costs, freight rates, the 24% annual carrying rate and the 18% IGST rate are illustrative assumptions, not quotes or a statement of any product's tax rate.

The IGST on inter-state transfers is cash-neutral for the company only if the receiving registration has enough output tax to absorb the credit. A depot that mainly supplies exempt or export customers can accumulate credit it cannot use, which turns a planning decision about where to hold stock into a working-capital cost. Our analysis of inventory turnover by warehouse shows how transfer invoices also distort depot-level stock ratios.

Inter-state stock transfers in India between distinct GST registrations are supplies under CGST Act Schedule I paragraph 2 and Section 25(4), valued under Rule 28(1); a ₹1,53,000 Bhiwandi to Guwahati transfer at an illustrative 18% carries ₹27,540 IGST, credited to the Assam registration.

Choosing between MRP, DRP or both for your operating model

The right answer depends on whether you make things, move things, or both. A manufacturer with depots needs both linked; a distributor needs DRP only; a single-site make-to-order shop needs MRP only.

• Single plant, ships direct: MRP from a master schedule is enough; there is no network for DRP to plan.
• Distributor or retailer without production: DRP plans branch transfers and rolls them up to purchase orders on suppliers at the central warehouse.
• Manufacturer with two or more stocking depots: DRP should generate the master schedule's demand, and MRP plans components from that schedule.
• Third-party logistics operator: the 3PL usually executes transfers that the client's DRP plans; it needs the planned transfer feed, not its own DRP.
• Few depots, short lanes: where every depot is a day from the central warehouse and reviewed daily, a reorder point may work as well as a full DRP grid.

Cost thresholds are a matter of judgement, not rule, but a pattern holds: DRP earns its setup effort when lane lead times exceed the review period, when depots number more than two or three, or when transfers are lot-sized by pallet or truck. If none of those applies, simpler depot replenishment is usually adequate. For buffer sizing at each depot, see our guide to distributor safety stock calculation.

Grid calculations to automate and allocation calls to keep human

Netting, offsetting and lot rounding should run as fixed calculations on every planning cycle. Decisions that trade one depot or customer against another, or override a forecast, belong to a named planner who can explain the call in the planning meeting.

Our evaluation of inventory planning software for VMI, safety stock and DRP replenishment looks at those tools from a buyer's angle, including consignment and supplier-managed stock. This article stays with the concepts and arithmetic you need before reading any vendor's demonstration.

Setup checklist before running linked MRP and DRP plans

A first linked run needs clean lanes, agreed lot sizes and a review routine more than it needs new software. Work through the checklist with the owners named earlier before switching off the spreadsheet reorder points.

Most of the benefit comes from making the depots' real needs visible to the plant, not from planning precision. More comparisons and worked planning guides are collected in our logistics and supply chain library. If you are selecting a planning system or connecting DRP to an existing ERP and want a second opinion on the scope, you can tell us about your project.