Connecting Your POS to Your Online Store
If your shop counter and your website sell the same physical stock, they are two authorities on one shelf — and that is guaranteed to produce overselling. Connecting them is mostly a data-hygiene job, and the decision that matters is which system is allowed to be right.
The moment a Malaysian retailer starts selling online as well as over the counter, a specific and predictable problem appears: two systems both believe they know how much stock exists, and they are both wrong. A sale at the till does not tell the website. A sale on the website does not tell the till. The shelf, meanwhile, has one true quantity that neither system is tracking correctly. Connecting them removes the problem — and the work is less about technology than about one decision made up front and a catalogue tidied once.
The decision that comes first
Which system is the source of truth for stock?
Answer that before you look at any integration. Everything else follows from it, and the failure mode of not answering it is two systems overwriting each other in a loop, which is worse than no integration at all.
| Source of truth | Choose this when | The trade-off |
|---|---|---|
| The POS | Most of your sales are in-store; staff work in the POS all day | The online store depends on POS uptime and its data quality |
| The online store | Most sales are online; the shop is a secondary channel | Shop staff may have to work in a second system for stock |
| A separate inventory system | You sell on marketplaces too, or have several outlets | One more system to run — but the only sane answer at multi-channel scale |
If you sell on Shopee, Lazada or TikTok Shop as well, the third option is usually the right one, because a POS is rarely designed to be the authoritative stock source for three marketplaces. See syncing inventory across channels.
The unglamorous prerequisite: one SKU list
This is where these projects actually fail, and it has nothing to do with integration.
- Every product needs one SKU code that is identical in both systems.
SHIRT-BLU-Min one andBlue Shirt (M)in the other will not match, and no integration can fix that for you. - Variants must be distinct SKUs. Size and colour are different products for stock purposes, always.
- Bundles need defining. A gift set that consumes three component SKUs has to decrement all three. Get this wrong and you oversell the components — see how overselling happens.
- Retire duplicates before you sync. Two records for the same product become two records in both systems, and the mess compounds.
- Barcodes should match too, where you use them.
Budget real time for this. On a catalogue of a few hundred SKUs it is typically a day or two of work — and it is a day or two that decides whether the integration works or produces a year of mysterious discrepancies.
What actually needs to sync, and which way
| Data | Direction | Notes |
|---|---|---|
| Stock levels | From the source of truth to everything else | The whole point. One way, always. |
| Products and prices | From wherever you maintain the catalogue | Decide once; maintaining in two places guarantees drift |
| Online orders | Into the POS or inventory system | So the shelf count is right and the day's sales are complete |
| In-store sales | Into the inventory system | Same reason, opposite direction |
| Customers | Optional | Useful for loyalty; check personal data obligations first |
| Returns | Both ways | The most commonly forgotten flow, and a reliable source of variance |
Prices deserve a deliberate decision. Many retailers legitimately price differently online and in store — delivery costs, marketplace fees and local competition all differ. Decide whether prices sync at all, and if they do not, decide where each is maintained so nobody has to remember.
Multi-outlet: stock by location, not in total
If you have several shops, a single total stock figure is actively misleading. You need stock by location, and then a rule for what the online store may sell:
- All locations pooled — maximum availability, but the order may need transferring between outlets before it ships.
- One fulfilment location only — simplest and most predictable; the other outlets' stock is invisible online.
- Nearest-outlet allocation — best customer experience, most complex to run.
Start with one fulfilment location. It is the option least likely to produce a cancellation, and cancellations are what this whole exercise exists to prevent. The wider operational picture for chains is in the multi-outlet maintenance SOP.
Testing before you go live
Do all five, with real data:
- Sell one unit at the till. Confirm the online figure drops within the sync interval.
- Sell one unit online. Confirm the POS figure drops.
- Process a return in store. Confirm stock goes back up in both.
- Receive new stock. Confirm it appears everywhere.
- Sell a bundle. Confirm every component decrements.
Test five is the one that gets skipped and the one that most often fails.
After go-live: three things to watch
- Stock accuracy against a physical count, weekly at first. See stock take for omnichannel sellers.
- Sync failures. Most tools log them; nobody reads the log until something breaks. Check it weekly for the first month.
- Cancellations caused by out-of-stock. This should fall. If it does not, your source of truth or your buffer is wrong.
And plan the accounting side rather than discovering it: online and in-store sales both have to reach your books without double counting. Marketplace-to-accounting automation covers that flow.
Frequently asked questions
My POS has no integration. What are my options?
Three: a middleware or sync platform that supports both systems, a scheduled CSV export and import, or replacing the POS. In that order of cost. A daily CSV routine is unglamorous but genuinely workable for a small catalogue, and much better than nothing.
Can I just update stock manually twice a day?
For a very small catalogue with low volume, yes, and plenty of shops do. It stops working when volume rises or when a marketplace is added, because the sync gap is where overselling comes from and a manual gap is measured in hours.
Do online and in-store prices have to match?
No, and there are good reasons for them to differ. Just be deliberate about it, make sure customers are not surprised, and ensure the price a customer is shown is the price they are charged — required under the Consumer Protection (Electronic Trade Transactions) Regulations 2024.
What about SST on online versus in-store sales?
The tax treatment follows the supply, not the channel. If you are SST-registered, the same rules apply to both — see SST and accounting basics and confirm anything unusual with your tax agent.
How long does this take to set up?
The technical connection is often hours. The catalogue cleanup beforehand is days, and it is the part that determines success. Plan for the cleanup as the project and the integration as its final step.
Sources: this guide is retail systems practice rather than a statement of law. The requirement that the price a customer is shown is the price payable arises under the Consumer Protection (Electronic Trade Transactions) Regulations 2024, administered by KPDN. SST treatment per RMCD, as covered in our SST guide. Integration availability differs by POS and store platform — confirm supported connections with each vendor before committing.
Estimate it in under a minute with our free calculator.