How to Prevent Overselling: Multichannel Inventory Sync
Why overselling happens across Amazon, Kaufland, Zalando, eBay and Shopify, and how to build a multichannel inventory sync that stops it, step by step.
- Written by
- Md AshikSenior Software Engineer, Artisan Dojo
- Published
- Reading time
- 8 min read
Quick summary30-second read
- One source of truth should own available-to-sell stock and push changes to every channel as soon as an order or stock movement happens.
- Sync available stock, not on-hand stock, so units already promised to open orders are never sold twice.
- Event-driven delta updates keep channels fast; a scheduled full reconciliation corrects updates that were missed or never applied.
- Safety buffers on low-stock items absorb marketplace processing lag; the faster your sync, the smaller they can be.
- Measure oversell rate and sync latency per channel to know whether your inventory sync actually works.
To prevent overselling across Amazon, Kaufland, Zalando, eBay and Shopify, you need one system that owns available-to-sell stock, pushes changes to every channel as soon as an order or stock movement happens, and holds back a small buffer for the seconds or minutes when channels disagree. Multichannel inventory sync rarely fails because a number is wrong. It fails because the right number arrives too late.
Here is why it happens, how a sync that holds up is built, and how to tell whether yours is good enough.
Why does overselling happen when you sell on several marketplaces?
Every channel keeps its own copy of your stock. Overselling happens in the gap between a sale on one channel and the moment the others learn about it. The usual causes:
- Polling intervals. If you check for new orders every 15 minutes, a sale on Amazon stays invisible to Kaufland and Shopify for up to 15 minutes.
- Marketplace processing lag. Sending a quantity update is not the same as the update being live. Amazon accepts listing changes for processing and applies them afterward, and bulk feeds can take longer. Your system says 0; the marketplace still sells.
- Wrong stock figure. Syncing physical stock (on hand) instead of available stock means units already promised to open orders get sold a second time.
- Bundles and kits. A bundle has no stock of its own. If its quantity is not recalculated whenever a component sells, the bundle keeps selling parts you no longer have.
- FBA mixed with merchant stock. Amazon holds FBA stock in its own warehouses. Counting it as shared stock, or pushing a merchant quantity onto an FBA offer, creates numbers that do not match reality.
- Multiple warehouses. If warehouses feed channels separately, or one is left out, totals go wrong.
- Returns and cancellations. Uninspected returns booked as sellable, or cancellations that never release stock, skew every channel.
- Two systems writing stock. Whichever writes last wins, and it is often the wrong one.
What causes overselling, and how do you fix each cause?
| Cause | Fix |
|---|---|
| Orders fetched on a slow schedule | React to order events (webhooks, notifications) and keep polling only as a fallback |
| Quantity updates applied late by the marketplace | Hold a safety buffer on low-stock items so lag does not turn into a sale |
| Physical stock synced instead of available stock | Sync available-to-sell: on hand minus reserved, committed and blocked units |
| Bundle stock maintained by hand | Calculate bundle quantity from components on every change |
| FBA and merchant stock mixed | Keep FBA quantities separate; only sync merchant-fulfilled offers from your system |
| Several systems updating the same channel | One source of truth; every other system reads, never writes stock |
| Missed or failed updates | Scheduled full reconciliation that corrects drift |
| Returns restocked before inspection | Book returns into a non-sellable state until checked |
How should a multichannel inventory sync be built?
The architecture is mostly about who owns the number and how fast changes travel:
- One source of truth for available-to-sell. Usually your ERP or warehouse system. It calculates a single available quantity per SKU from all sellable warehouses, minus open orders and reservations.
- Events in. Every new order, cancellation, goods receipt and stock correction is an event. Each one triggers a recalculation for the affected SKUs, and for any bundles that contain them.
- Allocation per channel. From the available quantity, compute what each channel may show: the full amount, a capped amount, or the amount minus a safety buffer.
- Delta updates out. Push only the SKUs that changed, as soon as they change. Queue them per channel and pace them to that channel's rate limits.
- Reconciliation on a schedule. Once or several times a day, compare every SKU's quantity on every channel with what it should be, and correct the difference. This catches what the event path missed.
- Monitoring. Alert when the update queue backs up, when a channel rejects updates, or when no orders arrive from a channel for longer than usual.
The event path keeps you fast. Reconciliation keeps you correct. You need both, because webhooks get lost and marketplaces sometimes accept an update and then fail to apply it.
Top tip
When an order comes in on one channel, the first job of your integration is not to import it into the ERP. It is to reduce available stock and push the new quantity to every other channel. Order import, invoicing and picking can follow a few seconds later.
What do Amazon, Kaufland and Shopify offer for stock updates?
A few details worth knowing:
- Amazon SP-API. For seller-fulfilled offers, quantity lives in the
fulfillment_availabilityattribute of the listing. The Listings Items API (patchListingsItem) updates individual SKUs, and Amazon's own guidance favors it for ongoing changes to smaller numbers of items. For large catalogs,JSON_LISTINGS_FEEDhandles bulk submissions, but feeds are processed asynchronously and are not guaranteed to be processed in order. On the event side,ORDER_CHANGEnotifies you of important order changes,LISTINGS_ITEM_MFN_QUANTITY_CHANGEtells you when the available quantity of a seller-fulfilled listing changes, andFBA_INVENTORY_AVAILABILITY_CHANGEScovers FBA stock. Rate limits apply to all of it, so pace your calls (here is how to handle 429 errors). - Kaufland Seller API. Stock is the
amountfield on a unit. You can update one unit withPATCH /units/{id_unit}or up to 150 units per request with the bulk endpoint, which reports a status per unit. Push notifications such asorder_unit_newandorder_unit_status_changedtell you about sales without polling. If your endpoint is unreachable, Kaufland retries for about 12 hours and then disables the subscription, so events from that period have to be fetched through the API. More on this in our Kaufland Odoo integration guide. - Shopify. Shopify tracks several inventory states per location, including
on_hand,available,committed,reservedandsafety_stock. If your system is the source of truth,inventorySetQuantitiessets absolute values and supports a compare-and-set check, so an update fails instead of overwriting a quantity that changed in the meantime. Shopify itself recommends this mutation only for systems that are the inventory source of truth, andinventoryAdjustQuantitiesfor everyone else.
Zalando and eBay have their own APIs and limits. The pattern is the same: listen for orders, push deltas, reconcile.
How do you set safety stock and channel buffers?
A buffer turns timing gaps into unsold units instead of oversold ones. Some practical rules:
- Buffer low stock, not everything. Two units held back from 500 cost nothing. Two units held back from three is a real decision.
- Size the buffer by sales speed and channel lag. A fast-selling item on a channel with slow updates needs more buffer than a slow item on a channel that updates quickly.
- Use a threshold rule. For example: show full stock above a threshold, and below it show stock minus the buffer, or zero.
Buffers hide sellable stock, so the faster your sync, the smaller they can be.
How do you measure whether your inventory sync is working?
Track two numbers at minimum:
- Oversell rate. Orders cancelled or delayed for lack of stock, divided by total orders, per channel. Tag these cancellations with a reason so they can be counted.
- Sync latency. Time from a stock change in your source system to the new quantity confirmed on each channel. Watch the slowest cases, not the average.
Also count how many differences each reconciliation run corrects. If it keeps finding many, the event path is leaking.
Is multichannel inventory management software enough, or do you need custom work?
Multichannel inventory management software is often the right answer. If it supports your ERP and marketplaces, handles your bundles and warehouses, and updates fast enough, use it.
Custom work makes sense when the tool cannot see the real available stock (for example, your warehouse system is not supported), when it cannot model your bundles or allocation rules, when its sync interval is too slow for your volume, or when two tools end up writing stock to the same channel. Often a small custom service feeding the correct number into the existing tool is enough. We compare the options in custom integration vs connector.
Step by step: how to stop overselling in an existing setup
- List every system that writes stock to each channel. If there is more than one per channel, fix that first.
- Confirm which figure is synced. It must be available-to-sell, not on hand.
- Measure sync latency per channel with a test SKU.
- Replace polling with events where the marketplace offers them, keeping a slower poll as a fallback.
- Add buffers to low-stock items on the slowest channels.
- Add a reconciliation job and log every correction it makes.
- Track oversell rate per channel each week and look at every oversold order's cause.
Common mistakes
- Treating "update sent" as "update live". Check processing results and listing status, not just the API response.
- Full syncs only. Pushing the whole catalog every hour is slow, burns rate limits and still leaves an hour-long gap.
- No reconciliation. Event-driven sync without a safety net drifts quietly until a busy day exposes it.
- Forgetting cancellations and returns. Stock leaves on orders but never comes back, or comes back too early.
- Ignoring rate limits. Throttled updates queue up, so stock is most out of date at peak.
- One buffer for every item. Uniform buffers hide stock on bestsellers and do too little on items about to run out.
FAQ
What is multichannel inventory sync?
It is the process of keeping stock quantities consistent across every sales channel. When stock changes anywhere, from a sale, return, receipt or correction, the new available quantity is sent to every marketplace and shop that sells the item.
Can you prevent overselling completely?
Not with certainty, because every marketplace applies updates with some delay. You can make it rare: event-driven updates, available-to-sell instead of physical stock, buffers on low-stock items and regular reconciliation.
How often should inventory sync between channels?
Changes should go out as soon as they happen, not on a timer. A full reconciliation can run nightly or several times a day to correct anything missed.
Why does Shopify oversell when my ERP shows the right stock?
Usually another app also writes stock to Shopify, the ERP syncs on-hand instead of available stock, or updates are delayed. Check which apps write inventory.
Losing orders to overselling?
We build and maintain marketplace ERP integrations with event-driven stock sync across Amazon, Kaufland, Zalando, eBay and Shopify, and we keep them running as the APIs change. Run through our integration checklist first, or describe your channels and systems below and book a free 30-minute call. We will tell you where your stock sync is losing time.