Retail online-offline inventory software is the mechanism behind shared stock: what counts as sellable, which locations each channel may draw on, when a unit is reserved, how quickly availability reaches each channel, and what happens when two channels commit the same unit. With Pentoggle, a retailer can describe how their channels should share stock and generate the starting application around it.
This page is the mechanics. The case for running channels as one business is Omnichannel Retail Software. Your own site is Retail E-commerce Management Software and marketplaces are Marketplace Order and Payment Management Software. The underlying stock record with its states and movements is Retail Inventory Management Software.
Most retailers already run an accounting system such as QuickBooks, Tally or Xero, a POS and a web platform, each with a view of stock. Those stay where they are. What is often still managed outside them is the arbitration between them: which figure is authoritative, who may promise it, and what happens when they disagree.
Key takeaways
- Sellable is a definition before it is a number. Reserved, damaged, display, in-transit and awaiting-inspection units are not sellable, and including them is a common cause of oversells.
- Every channel needs an explicit rule about which locations it may promise from. Left implicit, the rule is whatever the integration happened to do.
- Reservation timing is a trade-off. Reserving at order protects the customer who committed; reserving at pick keeps stock available longer. Neither is right for every channel.
- Sync speed sets the size of the collision window. Faster is better and no speed removes the window entirely, so an oversell process is still required.
- A useful number is the oversell rate per channel, read alongside how much stock is being held back as buffer.
The spreadsheet is often not the problem
A store that updates its website stock once a day from the shop system is running a workable arrangement, particularly where online volume is small and items are rarely down to the last unit.
The trouble starts at identifiable points.
When the last unit is the one that sells
Most items have plenty of cover and never collide. The collisions happen on the item that is down to one or two, which is also the item most likely to be selling well in both channels at once.
When the buffer becomes the policy
After a few oversells, someone sets the website to hide the last two of everything. The oversells stop and the store now never sells the last two of anything, across the whole range, all year.
When each channel has its own quantity
The site has a number, the marketplace has a number, the POS has a number. When they disagree, there is no authoritative figure to settle against, only whoever updated theirs last.
When an oversell has no process
The order was accepted, the stock is not there. What happens next is invented by whoever notices, and the customer's experience depends on who that was.
What online-offline inventory software holds
Sellable definition
Which stock states count as sellable, per channel where they differ, so the published figure excludes reserved, damaged, display, in-transit and awaiting-inspection units.
Channel-location rules
Which locations each channel may promise from: warehouse only, specific stores, all locations, or a weighted set.
Buffers
Any quantity held back per channel, per item or per category, set deliberately with a reason rather than applied across the board.
Reservation policy
When a unit is committed: at order, at payment, at pick, or after a review step, defined per channel.
Availability publishing
What is sent to each channel and when, whether a quantity or a simple in-stock flag, with a record of what was last published.
Change events
Every movement that alters sellable stock, sale, receipt, transfer, return, adjustment, triggering an update to the channels that care about it.
Oversell handling
A defined process when committed demand exceeds available stock: which order has priority, who is contacted, what is offered, and how it is recorded.
Reconciliation
Periodic comparison between what each channel believes and what the record holds, with differences listed.
Sellable is a definition, not a count
The figure a channel publishes should be what the business can actually deliver, and that is narrower than what it owns.
Units reserved for a customer order are committed. Damaged units are not sellable until repaired. Display units may or may not be sellable depending on policy. Units in transit between locations belong to neither end yet. Returned units awaiting inspection are not sellable until someone looks at them. Each of these is a stock state, as described on Retail Inventory Management Software, and each has to be excluded from the published figure or it will be sold.
The definition can legitimately differ by channel. A website that ships from a warehouse may publish warehouse sellable stock only. A store's own click-and-collect page may publish that store's sellable stock. A marketplace with strict cancellation consequences may publish a more conservative figure than your own site, which is a deliberate choice rather than a buffer applied by habit.
What matters is that the definition is written down and applied by the system, rather than living in an integration nobody has examined since it was set up.
Reservation timing is the real trade-off
Once a channel has accepted an order, the units behind it have to be protected from the rest of the business. When that protection starts is a choice with consequences either way.
Reserving at order protects the customer who has committed. The unit leaves the sellable pool immediately, and the counter cannot sell it. The cost is that abandoned or unpaid orders hold stock out of sale until they expire, so a short expiry is needed.
Reserving at payment is a middle position, common where payment follows the order by a short interval.
Reserving at pick keeps stock available for as long as possible, which suits stores where online volume is low relative to the counter. The cost is that the picker sometimes finds the unit gone, which is an oversell discovered late.
Adding a review step, for fraud checks or manual approval, changes the picture again, since the order exists but is not yet certain.
The workable answer is usually different per channel. Reserve at order for marketplaces where a cancellation carries consequences beyond the customer. Reserve at pick for a low-volume own site where the counter should have priority. Whatever is chosen should be set as a rule, with an expiry where reservations can go stale, rather than emerging from how each integration behaves.
Sync speed sets the collision window
Between a unit selling at the counter and every channel knowing about it, there is a window. Anything sold in that window is an oversell waiting to be discovered.
Shortening it is the main lever. Updating channels on the event, when a sale, receipt, transfer or adjustment happens, rather than on a schedule, reduces the window from hours to seconds for most movements. Where a channel only accepts scheduled updates, the frequency is set by what that channel allows, and the store should know which channels are slower and treat their stock more conservatively.
Publishing a flag rather than a quantity is worth considering for some channels. Where a channel only needs to know whether an item is available, sending in-stock or out-of-stock avoids constant quantity updates and reduces the noise, at the cost of some information for the customer.
No sync speed removes the window entirely, which is why the oversell process is not optional. It has to define which commitment takes priority, usually the earlier one or the one from the channel with the highest cancellation cost, who contacts the customer and how quickly, what is offered, a unit from another location, a substitute, a delivery date or a refund, and how the event is recorded so the rate can be measured.
Handled well, an oversell is a phone call and an alternative. Handled badly, it is a cancellation the customer discovers themselves.
Buffers should be deliberate and small
A buffer is stock hidden from a channel so that a collision cannot happen. It is a legitimate tool and an expensive habit.
Applied across the whole range, a buffer of two units means the store never sells the last two of anything online, all year, across every item. For a store with thousands of items, that is a large amount of stock permanently withheld from sale to prevent an occasional apology.
Applied selectively, a buffer makes sense: on the fastest-moving items where the collision risk is real, on channels where cancellation carries a penalty, or on items picked from a busy shop floor where the counter is likely to win the race.
The way to keep it honest is to measure both sides. Oversell rate says what the buffer is preventing. Buffer-held stock, in units and value, says what it is costing. Reviewed together every few months, the buffer can be tightened where the risk turned out to be lower than feared.
Where shared stock looks different by business type
- D2C Brand Software, where several marketplaces and a website draw on one warehouse and collisions between channels are the main risk.
- Apparel Retail Software, where single units of a size and colour are common and the last-unit problem is the normal case.
- Grocery Store Software, where online orders are picked from shelves customers are shopping, and substitution rather than reservation is often the practical answer.
- Electronics Retail Software, where serialised units mean a specific unit is committed rather than a quantity.
- Multi-Store Retail Software, where the channel-location rule decides how much of the chain's stock the website can see.
- Furniture Retail Software, where lead times are long and availability is often a date rather than a quantity.
Why retailers choose Pentoggle for shared inventory
One sellable figure, defined
States excluded explicitly, per channel where they differ, so what is published is what can be delivered.
Channel rules you set
Which locations each channel may promise from, and any buffer, decided deliberately rather than inherited from an integration.
Reservation timing per channel
Reserve at order where a cancellation is costly, at pick where the counter should have priority, with expiry on stale reservations.
Updates on the event
Availability pushed when stock moves rather than on a schedule, with slower channels identified and treated accordingly.
An oversell process, not an apology
Priority, contact, alternatives and recording defined in advance, because no sync speed removes the window entirely.
A useful number for shared inventory
Oversell rate per channel, read alongside the stock being held back as buffer.
Alone, either can be optimised in a way that damages the business. Zero oversells achieved by hiding a quarter of the stock is not a success; full availability with a steady stream of cancellations is not either.
Read oversells by item. They concentrate on fast movers and single-unit variants, which is where a targeted buffer or a faster sync is worth the effort.
Read the cost of each oversell by channel. A cancellation on your own site costs a customer's goodwill; on a marketplace it may affect the account, which is why the reservation and buffer rules for that channel can reasonably be tighter.
Ready to build retail online-offline inventory software?
You publish a stock figure to your website and your marketplaces.
You may not be able to say what that figure excludes, which locations it draws on, or how long ago it was last true.
Describe your channels, your locations and what each may promise to Pentoggle in plain English and generate a working first version in hours, then refine it around your process.