A wholesale order takes the last 30 units of a fast-moving SKU.
The warehouse has already allocated them. Shopify still shows the product as available. By the time someone spots the mismatch, more orders have come in—and the team is trying to work out which customers they can still fulfill.
Every system may be connected. The operation is still out of step.
This is where the difference between integration and synchronization becomes practical. Integration moves information from one system to another. Synchronization is about keeping relevant data aligned across systems as inventory changes. Effective synchronization also requires clear rules for handling updates that arrive late, out of order, or in conflict.
The simplest way to put it is:
Integration asks: Can these systems exchange data?
Synchronization asks: Do they agree on the state, timing, ownership, and correction logic needed for this decision?
A connection is the starting point. Operational agreement is the goal.
Being connected is not the same as agreeing
Most ecommerce tech stacks were built one tool at a time.
Shopify ran the store. Then came a marketplace, a 3PL portal, accounting software, perhaps an ERP—and usually a spreadsheet to patch the gaps between them.
Connecting those tools can save hours of rekeying data. But the connection itself does not decide:
how quickly a stock change needs to appear everywhere;
which update should win when two systems disagree;
whether a unit is still sellable after it has been reserved elsewhere;
where an inventory correction should be made; or
who needs to act when an update fails.
Those are operating rules, not connector settings.
Without them, data can move successfully while the team still has to stop and ask, “Which number do we trust?”
The same SKU can have several correct numbers
An inventory number is only useful when you know what it represents.
Shopify separates inventory into on hand, available, committed, unavailable, and incoming. A warehouse might report everything physically on the shelf. A sales channel, meanwhile, needs to know what is still available after orders, reservations, damaged stock, and safety stock have been taken into account.
Both numbers can be correct. They are simply answering different questions.
If those inventory states are causing confusion, our guide to on-hand vs. available inventory explains what each number represents and why they can differ without either one necessarily being wrong.
Location matters too. A product can be sitting in one warehouse and still be unavailable to an online customer because that location is not set up to fulfill the order.
So when someone says, “Shopify doesn’t match the warehouse,” the mismatch may be caused by a delay. But it may also mean the team is comparing physical stock with sellable stock, one location with another, or inventory before and after a reservation.
When the disagreement is specifically between Shopify and a fulfillment partner, our guide to why Shopify and your 3PL can show different inventory numbers walks through the inventory-state, timing, location, and synchronization differences worth checking first.
Good synchronization does not force every system to show the same number. It makes the differences understandable.
Where coordination breaks
The following problems are ordinary. What matters is whether the business has agreed on how to handle them.
Timing
A short delay is not equally risky for every SKU. It may be harmless for a slow-moving item with plenty of stock, but costly during a promotion or when only two units remain.
The useful question is not, “Is this real time?” – It is, “Is this fast enough for the decision we are about to make?”
Update sequence
That means the last message received is not automatically the latest truth. The process needs a way to confirm the current state instead of trusting arrival order.
Reservations and allocations
A unit can exist physically while already being promised somewhere else. Wholesale allocations, marketplace orders, draft orders, damaged stock, quality-control holds, and safety stock can all reduce what is actually available to sell.
If one system sends the physical count and another works from available inventory, they can appear to disagree even when both are behaving as designed.
Reservations can make that distinction even harder to see. Our guide to how inventory reservations affect what you can actually sell explains how commitments and holds can change availability without changing the physical stock on the shelf.
Manual corrections
Inventory corrections often create a second problem. Someone adjusts the warehouse count, while someone else—trying to fix the same issue—changes Shopify.
Neither person is necessarily wrong. The problem is that the business has not made it clear which system owns that correction.
When an inventory discrepancy appears, the first step should be to investigate the underlying inventory movement and process before correcting the record. Shopify’s inventory adjustment history records what changed, when it changed, what caused the adjustment, and who or what made it—context the team can use to understand the discrepancy.
Failed updates
Some updates will fail. They may be delayed, rejected, duplicated, or never processed.
When that happens, the team should not have to hunt through every system to discover the problem. They need to know which SKU or order is affected, which systems now disagree, whether a customer promise is at risk, and who owns the next step.
The hard part is not moving data when everything works. It is deciding what happens when it does not.
Why replacement is not the only question
When reconciliation becomes a daily burden, replacing the whole stack can feel like the cleanest answer.
Sometimes it is. A new ERP or inventory platform may be the right move. But “Which system should we buy?” is usually not the best first question.
Start by asking what is actually broken. Is the current tool missing a needed feature? Are updates too slow? Are teams using different definitions? Is ownership unclear? Or are exceptions simply invisible until a customer order is affected?
A new platform will still need decisions about inventory states, locations, reservations, transfers, overrides, and conflict handling. It will also bring migration, configuration, training, and workflow changes.
If those operating decisions are still unresolved, a larger system may move the uncertainty into one place without removing it.
Six questions to ask before changing the stack
Before committing to a replacement, take one real SKU and follow it from receipt to sale. Use that journey to answer six questions.
1. Which system owns each kind of inventory information?
Identify where physical stock, available-to-sell inventory, commitments, returns, transfers, damaged units, and safety stock are first recorded.
One system does not need to own every type of information. The important point is that ownership is explicit.
2. How fresh does the information need to be?
A campaign launch, replenishment decision, customer promise, and month-end reconciliation do not carry the same risk. Set the acceptable delay around the decision, not around what the software happens to offer.
3. What happens when systems disagree?
The right answer may depend on the data. The warehouse may own the physical count, while Shopify or a marketplace owns its open commitments.
Avoid naming one universal “source of truth” until you have defined which truth each decision needs.
4. Can a correction be explained—and undone?
The team should be able to see who changed the number, when they changed it, what evidence they used, and how to reverse the adjustment if it was wrong.
5. Who owns the exception?
A mismatch that appears on everyone’s dashboard but belongs to no one is unlikely to be resolved quickly.
Decide who investigates it, who can approve a correction, and when the issue needs to be escalated.
6. What new work will the change create?
Look beyond implementation cost. Consider configuration, training, maintenance, workflow disruption, and any manual reconciliation that will remain after launch.
The best option is not simply the one that connects the most tools. It is the one that leaves the team with fewer unexplained decisions.
What operational confidence looks like
Operators do not need another dashboard that simply repeats the disagreement. They need to know what changed, when it changed, which system owns the relevant information, why the numbers differ, and which exceptions need a person to step in.
At SuppliFlex, this is one of the problems we’re focused on understanding: how teams can make inventory differences more traceable and explainable while continuing to work with the systems they already use.
If your connected systems still disagree, which coordination rule is missing?