Your Channels Don’t Talk. Your Inventory Pays.

A multi-channel sync tool copies counts between channels. It does not decide which record is true. For standard SKUs that is enough. For unique units, consignment, pair-level pricing, and drops, it moves the oversell problem instead of solving it. The fix is one inventory record that every channel reads.

Amazon does not know what Shopify sold. The warehouse does not know what eBay listed.

The retail industry’s standard answer is a sync plugin. Connect the channels, set a polling interval, and let the tool push counts back and forth. Many sellers run this way, and plenty of them should.

This article argues against that default for a specific kind of seller. If your inventory is not standard, a sync tool does not fix the space between your channels. It adds a third opinion about what you have.

What does a sync tool actually do?

It copies quantities from one channel to another on a schedule or on an event.

That is a useful job. It is also a narrow one. A sync tool assumes three things:

• Every item is a SKU with a quantity.
• Every channel describes the item the same way.
• A count that is right on one channel is right everywhere.

When those assumptions hold, sync works. When they break, the tool keeps copying numbers faithfully while the numbers drift away from the shelf.

Why do sync tools still oversell?

Because they synchronize copies. They do not own the truth.

Each channel keeps its own record. The tool tries to keep those records in step. Between updates, or when a rule does not fit, two channels can each believe the last unit is available.

The illustrative sequence looks like this:

1. The last unit sells on one marketplace.
2. The other channels still show it in stock.
3. A second order arrives for a unit that is gone.
4. Next day: a manual refund, an apology, and a marketplace ding.

That last step is the real cost. Marketplaces measure cancellations, and a seller rating is slow to rebuild.

Where do sync tools break first?

At the exceptions. For many sellers, the exceptions are the business.

Your inventory	What a sync tool assumes	What actually happens
Unique units (one of one)	Quantity is a number you can add and subtract	Each unit has its own condition, size, and price
Pair-level or unit-level pricing	One price per SKU per channel	Two identical products carry different prices
Consignment	You own everything you list	Ownership and payout differ per unit
Shows and pop-ups	Channels are online storefronts	Sales happen offline and reach the system late
Drops and upcoming releases	Listed means in stock	You list items you do not own yet

Each row is a place where someone on your team steps in by hand. When off-the-shelf tools fail, you are not only choosing software. You are choosing who owns the manual exceptions.

Isn’t more sync just the price of selling everywhere?

That is the status quo, and it is worth challenging.

The standard advice is to add channels for reach and add a tool to hold them together. The hidden assumption is that reconciliation is permanent overhead: a weekly tax you pay for being on more than one marketplace.

It is not permanent. It comes from a design choice: letting every channel hold its own copy of the inventory. Change the design and most of the reconciliation disappears.

This is the same pattern that sits behind the Complexity Ceiling. More software is often how a company hits the ceiling faster. The growth systems engineering guide explains the broader principle: make the tools agree instead of adding tools.

What is the alternative to syncing copies?

One inventory register that every channel reads.

The register is the single record of what you own. Channels do not hold their own truth. They display the register’s truth in their own format.

Three rules make it work:

1. A sale writes to the register once. Every channel reads the new count from the same place.
2. Channel rules live on top of the record, not inside copies. Pricing, fees, titles, and margin floors differ by channel without duplicating the item.
3. Anything that does not match goes to a queue, not a customer. A mismatch becomes a task for a person, not an oversell.

The brand site then becomes the public layer of that register, not a second store someone has to keep honest.

The retail systems page shows the register in practice: inventory synchronization, pricing logic, order flow, and exception handling as one operating layer.

Has this worked for a real retailer?

Yes. HEAT4YAFEAT is a collector in Indianapolis selling through eBay and StockX seller portals.

Demand was real. Infrastructure was not. There was no owned catalog, no owned checkout, and no record of a pair that a second channel could read.

Conjai built the operating layer first:

• One inventory register for every pair he owns.
• eBay and StockX connected to that record.
• A drops pipeline that publishes upcoming releases, with listings, images, and a take in his voice, without treating those rows as stock.
• A pair becomes inventory only when he acquires it.
• The brand site is the public layer of the register, not a separate catalog.

The published results: 650+ pairs placed, 99.1% positive on eBay, selling since 2016. Demand was never the problem. Agreement was.

Why do drops break most sync tools?

Because a drop is a promise, not a unit.

Sneaker, streetwear, and collectibles sellers often publish upcoming releases before they own the product. Customers want to see the calendar. The seller wants the listing ready.

If the calendar and the inventory share a table, a pair you do not own can sell. Most sync tools cannot tell the difference between “coming soon” and “in stock,” because both look like a listing.

Drops are not stock. The register keeps them in a separate state until acquisition, so the calendar can publish freely without inflating what you can sell.

What should you check before adding another channel?

Every new marketplace adds reach and one more place where inventory can disagree. Before you connect it, answer five questions:

1. Does the new channel read from the register, or keep its own copy? If it keeps its own copy, you are adding a second opinion.
2. Can it express your pricing rules? Some channels cannot hold unit-level prices or margin floors, so someone ends up repricing by hand.
3. How does it report a sale? A channel that reports orders late widens the window for an oversell.
4. What happens to returns and cancellations? A returned unit has to go back into the register with its real condition, not as a fresh copy.
5. Who owns the exceptions? If the answer is “whoever notices,” the channel will cost more time than it saves.

If most answers are clean, add the channel. If they are not, fix the register first. Reach you cannot fulfill accurately is not growth. It is a larger reconciliation tax, collected every week.

When should you just buy the sync tool?

When your catalog fits it.

Buy the $200/month sync tool if most of these are true:

• You sell standard SKUs with interchangeable units.
• One price per product per channel is fine.
• You own everything you list, and it is on the shelf.
• Oversells are rare enough that fixing them by hand is cheaper than engineering.

Many businesses need custom engineering only once their channels, pricing logic, inventory rules, or exceptions move beyond the standard case. We build for the part the platform will not do, and we will tell you plainly when the platform is enough.

How much are the mismatches costing you?

Use your own numbers, not a benchmark.

Count oversold orders per month and multiply by the cost of each: refund fees, shipping, and marketplace penalty. Then count the hours per week spent fixing listings and counts by hand, and multiply by the loaded cost of those hours.

Add the two. That monthly figure is the reconciliation tax. The retail systems page has a calculator that runs the arithmetic, and the overselling diagnostic walks through the symptoms.

FAQ