OneXEOS
Blog detail

Blog detail

How to Sync Inventory Across Shopee, Lazada, and TikTok Shop

Getting started

Keep stock aligned on three marketplaces: one master count, SKU mapping, buffer rules, a worked bundle example and a daily reconciliation routine.

How to Sync Inventory Across Shopee, Lazada, and TikTok Shop

You list the last twelve units of a best-selling serum on three marketplaces at once, and by lunchtime you have orders for fifteen. Nobody did anything careless; each storefront simply kept counting from its own number. The fix is to run all three storefronts from one master stock count, with every marketplace listing mapped to the same internal SKU, a buffer held back for safety, and the available quantity pushed out to each storefront whenever stock moves. Everything else in this guide is the detail of making that sentence work in a real Vietnamese seller's week.

We will cover how to decide how much automation you need, how to build the SKU map, how to size buffers, how to handle bundles and returns, a worked example with a bundle, and the reconciliation routine that catches what automation misses.

Why three storefronts drift apart

Shopee, Lazada and TikTok Shop each hold a separate quantity for your listing. Unless something ties those quantities to a single warehouse record, they will diverge. The usual causes are mundane:

  • Update lag. An order lands on one marketplace. Until the other two are told, they keep selling units that are already spoken for.
  • Cancellations and returns. Stock comes back to the shelf but is added to one storefront, or none.
  • Sales outside the marketplaces. Counter sales, chat orders, wholesale pickups and samples reduce physical stock without touching any online count.
  • Receiving and stocktakes. A delivery is shelved but not recorded, or a count corrects the warehouse sheet but not the listings.
  • Inconsistent product data. The same item has different titles, variant names and codes on each platform, so staff edit the wrong line.

Notice that none of these is a technology failure. They are failures of ownership: nobody is responsible for the number as a whole. Software helps only after you have decided who owns it.

How much automation do you need?

Match the method to your volume. The thresholds below are rules of thumb to adjust, not benchmarks.

ApproachHow it worksFits whenMain weakness
Manual editsA person opens each storefront and changes the quantity after every orderA few dozen SKUs, a handful of orders a daySlow and fragile at peak times
Master spreadsheetOne sheet holds true stock; you push updates to each storefront at set timesA few hundred SKUs, one disciplined ownerLag between updates; two editors can overwrite each other
Central stock system linked to marketplacesOrders flow into one place, stock is deducted automatically, available quantity is pushed back outMany SKUs, many daily orders, two or more marketplaces, possibly several stock locationsSetup effort; demands clean SKU data and a defined process

A useful trigger for moving up a level: cancellations for "out of stock" happen more than once a week, or someone spends a large part of a day just comparing screens. At that point it is worth looking at a multi-marketplace selling platform and judging it against the questions later in this article.

Step 1: Make one place the source of truth

Decide, in writing, that the warehouse record is the only authoritative count. What shows on each marketplace is a calculated copy. If someone must change a quantity directly on a storefront, for instance to pull a listing during a flash problem, the change has to be logged and fed back into the master count the same day.

If you stock goods in more than one place, say a back room, a rented warehouse and a shop floor, also decide which locations feed online orders. Looking at how a dedicated inventory management application handles receipts, transfers and stocktakes is a good way to list the movements your master record must be able to capture.

Step 2: Build the SKU map

An internal SKU is your own permanent code for one sellable variant. Each marketplace listing, and each variation within it, points to exactly one internal SKU. Build the map in a sheet before you configure anything:

  • One row per variant. Size M and size L are different SKUs even if they share a photo.
  • Internal codes should be short and stable. Never derive them from marketplace titles, which you will edit for search reasons.
  • Allow many listings to point at one SKU. Sellers often run two listings for the same product with different titles or pack sizes.
  • Record bundles explicitly: which component SKUs, and how many of each, one bundle sale consumes.
  • Keep retired codes in the map so an old order that comes back as a return can still be traced.

Expect this step to take longer than you think. It is also where most hidden problems surface, such as two products sharing a code or a variant nobody remembers creating. Time spent here saves far more later.

Step 3: Size your buffers and choose an allocation rule

Even a fast sync has a delay, and some stock will turn out to be damaged or misplaced. So the number you publish should be lower than the number on the shelf.

Core formulas:

  • Online available = Physical stock − Units reserved for unshipped orders − Units set aside for other channels (shop, wholesale) − Safety buffer.
  • Published per marketplace = Online available, either shared in full or multiplied by that marketplace's share.

Setting the buffer

There is no universal buffer. Base it on two things: how quickly the item sells and how long your sync takes. A workable method is to look at the most orders you have received for that item within one sync interval, then add a small margin for defects. Fast movers deserve a larger buffer; slow movers need almost none. Start cautious and adjust using the number of oversells you actually see, recorded in your reconciliation log.

Shared pool or fixed shares

  • Shared pool: every storefront shows the same available quantity. It is simple and maximizes sales, but the storefront that gets traffic first takes everything. It suits setups where updates are close to immediate.
  • Fixed shares: each storefront gets a slice. Easier to control, and it lets you protect a channel you care about or save stock for a campaign, but one storefront can sell out while another sits on spare units.

Many sellers switch between the two: shared in normal weeks, fixed shares ahead of a big campaign on one marketplace.

Worked example: a gift set built from two components

All numbers below are assumptions for illustrating the arithmetic, not market data. Substitute your own.

A skincare seller carries a serum (SKU SRM-30) and a cleanser (SKU CLN-100). It also sells a gift set (SET-A) that consumes one of each. Physical stock is 20 serums and 14 cleansers. The rules: hold back 2 of each for the retail counter, keep a buffer of 2 serums and 1 cleanser, and share availability across all three marketplaces in a single pool.

  1. Online available for each component. Serum: 20 − 2 − 2 = 16. Cleanser: 14 − 2 − 1 = 11.
  2. Standalone listings. Every marketplace shows 16 serums and 11 cleansers.
  3. Gift set availability. The set can only be as large as its scarcest component: the smaller of 16 and 11, which is 11 sets. But the same cleansers also serve standalone sales, so the set and the cleanser compete for one pool of 11.
  4. Decide the priority. The seller decides sets should never exceed 6, so they cannot swallow the stand-alone cleanser stock. SET-A publishes 6.
  5. An order arrives. One marketplace receives 2 gift sets and another receives 1 cleanser. Serums reserved: 2, cleansers reserved: 3.
  6. Recalculate. Serum online available becomes 16 − 2 = 14. Cleanser becomes 11 − 3 = 8. Gift set publishes the smaller of 14, 8 and the cap of 6 minus the 2 already sold, so 4.
  7. A return comes in. One gift set is returned and inspected. The cleanser is sealed and resalable; the serum box is damaged. Cleanser goes back to stock (available 9), the serum goes to a defective bin and does not return to the available pool.

Two lessons here. First, a bundle must always be calculated from its components rather than counted separately, or you will promise the same bottle twice. Second, a return is two decisions, not one: what comes back and in what condition.

Movements that must update the count

EventWhat should happenCommonly missed
New orderReserve units immediately; deduct on dispatchDeducting only on dispatch leaves a window for resale
Cancellation before dispatchRelease the reservationReleased in the warehouse record but never republished
ReturnRestock sellable units; route damaged ones elsewhereAdding everything back without inspection
Goods receivedAdd after checking quantity and conditionStock sits on the shelf, unrecorded, while listings show zero
Off-platform sale or transferDeduct from the source locationCounter and chat sales are never written down
Stocktake adjustmentCorrect physical stock and republishWarehouse sheet fixed, storefronts not

Put a name beside every row. Any event without an owner is a future discrepancy.

Questions to ask before you pick a tool

If you decide to use software, resist choosing from a feature list. Ask for a trial with your own SKU file and a week of real orders, and use these questions:

  • Does it only import orders, or does it also push available quantity back to each storefront?
  • What is the measured delay between an order on one marketplace and the updated number appearing on the others?
  • What happens when the connection to a marketplace drops: visible alert, queued catch-up, or silent failure?
  • Can the SKU map be imported in bulk, and does it handle bundles, variations and several listings per SKU?
  • Is there a change log showing who altered stock, when, and why?
  • Can you set buffers and per-marketplace shares at SKU level?

To see which applications OneXEOS currently offers in this area, read the page on marketplace integrations and bring your answers to the list above to any conversation with the team.

Reconciliation routine

Automatic sync reduces errors; reconciliation finds the ones that remain. A rhythm that stays manageable for a small team:

  • End of each day, about ten minutes: compare the published quantity on each storefront with the system's calculated figure for your top ten movers. Log every difference, even one unit.
  • Weekly: physically count a rotating group of SKUs, for example skincare one week and accessories the next. Correct the record and write down the reason.
  • Monthly: review orders cancelled for lack of stock, express them as a share of all orders, and decide whether buffers should rise or fall.
  • After every large campaign: audit all participating SKUs, because that is when sync is under the most stress.

When you find a gap, classify it: sync delay, missed return, off-book sale, or mapping error. After a few weeks one or two categories will dominate, and that tells you which part of the process to fix first.

Common mistakes

  • Editing quantities directly on a storefront and forgetting the master count, so the next sync overwrites the edit.
  • Running without a buffer on the products that sell fastest.
  • Syncing orders into the warehouse record but never pushing quantities back out.
  • Restocking returns without inspecting them.
  • Letting the SKU map rot: duplicate codes, orphaned listings, bundles with no recipe.
  • Launching a campaign without changing allocation, so one marketplace drains the whole pool in minutes.
  • Treating software as the end of the project, with no owner for daily checks.

Frequently asked questions

Do I need software if I sell on only two marketplaces?

Not necessarily. With a small catalogue and low daily order counts, a well-run master spreadsheet can be enough. Start considering software when comparison time grows, or when oversells begin to cost you cancellations and rating damage.

Does sync have to be instant?

Faster is safer, but what matters more is knowing your real delay and covering it with a buffer. A slightly slower sync with a well-chosen buffer beats a fast one with no way to handle exceptions.

How do I treat stock held in a marketplace-run fulfilment programme?

Record it as a separate location from your own warehouse, and calculate availability according to where each order will actually ship from. Programme terms vary and change, so read the marketplace's current rules instead of assuming.

What if one product has several listings?

Map all of them to the same internal SKU and either share the available quantity or split it by percentage. The essential point is that every listing deducts from one number.

What should I do when I oversell anyway?

Contact the buyer quickly, offer a clear choice such as waiting for restock, an equivalent item or a cancellation with an apology, then trace the cause through your stock log and adjust the buffer or the process.

Next steps

  1. Export every listing and variation from each marketplace and map them to internal SKUs.
  2. Write down that the warehouse record is the source of truth and who may change it.
  3. Set starting buffers for your twenty fastest products and pick shared or fixed allocation.
  4. Start the daily ten-minute check and the weekly rotating count, noting the cause of each gap.
  5. If you move to software, trial it with your own file and real orders against the questions above before committing.
POSCRMInventoryKDSStaffReportsAI Menu

Ready to upgrade the operating system for your business?

You are ready to upgrade the operating system for your business in about 5 minutes. You create an account, pick your industry and switch on the apps you need, with the core plan free forever. No credit card and no long-term contract. You can add a branch at any point. Everything you turn on shares the same 1 data layer.

Sync Inventory Across Shopee, Lazada and TikTok Shop