Counter and website
A POS that syncs with your ecommerce website — by not needing to
If you are searching for a point of sale that syncs with your online store, something has already gone wrong at least once. Usually it is an order for an item that sold in the shop that morning, and a message to a customer explaining that the thing they paid for does not exist.
The instinct is to look for better sync: a faster connector, a shorter interval, a more reliable plugin. That is treating the symptom. Every sync is two systems holding two copies of the same number and racing to agree, and every sync has a window in which they do not.
The alternative is to stop keeping two copies. When the till and the storefront are two interfaces onto one stock ledger, there is no interval, no queue, no failed job and no window — the number changes once, and both views were already looking at it.
What you get
What this gives you
No sync interval, because no sync
The counter and the storefront read the same stock record. A sale at the till changes the number; the next page load on the website sees the changed number. There is nothing scheduled and nothing to fall behind.
Stock reserved at checkout, not after
A web order holds stock when the customer commits, so a simultaneous counter sale cannot quietly take the same unit. The race that produces oversold orders is resolved by the database, not by whoever's job ran last.
One price change, both channels
Change a price, run a promotion, or set a price list for a customer group, and the counter and the storefront reflect it together. No more discovering that the website spent a week selling at last month's price.
Web orders in the same reports
An online order and a counter sale are the same kind of record, tagged by channel. Revenue, margin, best sellers and stock turn are one set of numbers you can split by channel — not two exports you merge in a spreadsheet.
Refunds and returns that stay correct
A return processed at the counter for something bought online puts the unit back into the same stock figure and reverses the same ledger entry. Cross-channel returns are where two-system setups break most often and most invisibly.
Nothing to reinstall after an update
There is no connector to be broken by somebody else's release. The thing that keeps your stock correct is not a third-party plugin with its own maintenance schedule.
Why 'real-time inventory sync' is usually not real time
When a product page says real-time sync, read it as 'event-driven with retries'. A sale fires an event, a queue picks it up, an API call lands, the other system updates. On a quiet Tuesday that is seconds. It is also exactly the path that degrades when it matters: at peak trading the queue is longest, the rate limits bite, and the retries back off.
The failure is also silent. A connector that drops one event in a thousand does not alarm anyone; it produces a stock figure that is quietly two units wrong, then five, then eleven, until a stock take reveals a discrepancy nobody can attribute to anything. Most retailers running a synced stack do not know their sync is lossy — they know their stock 'always drifts a bit'.
There is also a modelling problem that no amount of reliability fixes. Two systems that were designed separately represent things differently: a variant in one is a separate product in the other, a partial refund in one has no equivalent in the other, a bundle sold at the counter decrements components the webshop does not know exist. The connector has to guess, and its guesses are consistent but not always right.
How to sync store inventory with an online store automatically
If you are staying on two systems, the honest version of the advice is this: pick one as the master for stock, push in one direction only, reconcile on a schedule you actually keep, and accept a safety buffer — hold back a unit or two from what the website may sell, so a sync gap costs you a sale rather than a cancelled order. Bidirectional sync between two masters is where most of the horror stories come from.
Then check what happens on failure. Does the connector alert you, or fail quietly? Does it replay missed events or skip them? Does it handle variants, bundles, partial returns and multi-location stock the same way both systems do? These are the questions that determine whether the setup holds; feature checklists do not.
The other option is to stop needing the answer. Moving to a single system removes the whole class of problem rather than managing it, and the migration is smaller than people expect because you are consolidating two catalogues into one rather than rebuilding anything.
- One master for stock, one direction of travel — never two writers
- A safety buffer on web-available quantity if you cannot avoid a sync gap
- Alerting on sync failure, not just on success
- Explicit handling for variants, bundles, partial refunds and multi-location stock
- A reconciliation you genuinely run, on a schedule you genuinely keep
Connecting a cash register to a website
A traditional cash register cannot be connected to a website in any meaningful sense. It totals a sale and opens a drawer; it has no product records to share and no network to share them over. What people mean by the question is usually 'how do I stop having to tell the website what I sold in the shop', and the answer is to replace the register with software rather than to bridge it.
That replacement is less disruptive than it sounds. A phone or tablet running sell.ke is a working till immediately; the receipt printer and cash drawer you already own will very likely work, because 80mm thermal printers and printer-driven drawers are a commodity standard rather than proprietary hardware. The register becomes a backup you keep in the cupboard for a fortnight and then stop thinking about.
Migrating without stopping trading
Nobody can close for a week to change systems, so the sequence matters. Import the product catalogue first and check it against the shelves — this is the only genuinely laborious step, and it is work you would do anyway to fix the two catalogues you currently maintain. Set opening stock from a count rather than from the old system's figures, because those figures are the reason you are moving.
Then run both for a fortnight: new system as the system of record, old one as a fallback you stop opening. Point the storefront at the new catalogue last, once the counter side has been correct for a week. A fourteen-day trial is deliberately long enough to do exactly this before you pay anything.
Compare
Sync approaches, and what each costs you
| What you are comparing | Manual updates | Connector / plugin sync | One system (sell.ke) |
|---|---|---|---|
| How current is web stock? | Hours to days | Seconds to minutes, when healthy | Current — same record |
| Overselling at peak | Frequent | Most likely exactly then | Not possible |
| Failure mode | You forget | Silent event loss | None to have |
| Variants, bundles, part-refunds | Handled by you | Handled by a guess | One model, both channels |
| Multi-location stock online | Impractical | Usually flattened to one number | Per location, shown honestly |
| Ongoing maintenance | Daily | Connector upkeep and monitoring | None |
Questions
POS that syncs with your website — questions
Which POS systems sync with an ecommerce website?
Most mainstream POS products offer some path to it — a first-party ecommerce add-on, an official connector to a major platform, or a third-party sync app. They differ in whether the link is first-party or bolted on, whether sync is one-way or bidirectional, how failures are surfaced, and how variants and returns are modelled. Those four questions predict how well it holds up far better than the feature list does. sell.ke takes the other route: the storefront is part of the POS, so there is no link to evaluate.
How fast is the stock update between the shop and the website?
There is no update to time. The counter and the storefront query the same stock record, so the website is not told about a counter sale — it was already reading the number that the sale changed. The only latency is the page load itself.
Can the website show stock per location?
Yes. Because locations hold their own stock rather than contributing to one pooled figure, the storefront can show availability honestly per location and route an order to the one that can actually fulfil it — including click-and-collect from a specific branch. Synced setups usually flatten multi-location stock into a single number because that is all the connector can carry.
I already have a WooCommerce or Shopify store. Do I have to abandon it?
No, and you should not decide that today. sell.ke integrates with WooCommerce, so you can keep the storefront you have and use sell.ke as the system of record behind it — which already removes the manual stock work even if you never switch storefronts. Whether to move the storefront itself is a separate decision about themes, plugins and SEO that is worth making on its own merits later.
What happens to an online order placed while the shop is offline?
Nothing unusual — the storefront is hosted and stays up regardless of the shop's connection. What pauses is the counter's own connectivity, and offline mode covers that: the till keeps selling and reserving stock locally, then reconciles when it reconnects. The narrow case worth knowing about is the same unit being sold offline at the counter and online at the same moment, which reconciliation surfaces as a conflict for a human to settle rather than resolving silently.
Does it keep working when the internet drops?
The counter does. Offline mode keeps taking sales, printing receipts and reserving stock while the connection is down, then reconciles everything when it returns — so a bad afternoon costs you connectivity, not revenue or a gap in the record. The parts that are inherently online stay online: card and mobile-money authorisation, tax-authority submission, and the storefront itself all need a live link.
Can I run more than one location?
Yes, from the Basic plan up. Locations share one product catalogue but hold their own stock, so moving goods between them is a recorded transfer rather than a re-count at both ends. Reports run per location or across all of them, and staff permissions are set per location — a branch supervisor does not need to see another branch's margins.
How long does setup take?
A single-counter shop is usually selling the same day. The work is the product list, not the software: import a spreadsheet, scan barcodes, or let Amina build products from a supplier invoice or a photo of your price list. A multi-location business with thousands of SKUs and opening stock counts should plan about a week and run the old system alongside for a fortnight. The 14-day trial is long enough to do exactly that.
How much does sell.ke cost?
Plans run from USD 12/month for a single-location till to USD 115/month for unlimited scale, with the online store on your own domain included from USD 23/month. In Kenya the same plans are KES 1,499 to KES 14,999/month, billed in shillings. Every plan starts with a 14-day trial and no card, and there is no per-terminal licence and no commission on your sales. There is no hardware to buy either — sell.ke runs on a phone, tablet or laptop you already own.
Read next
The other pages that answer the question you are probably about to ask.
All-in-one POS, inventory & online store
One system for the counter, the stock and the website — not three that sync.
Omnichannel retail management
Every channel drawing on one inventory, one customer record and one ledger.
Multi-location inventory
Per-location stock, recorded transfers, and reports that split or combine.
Ecommerce website builder
An online store on your own domain, built from the catalogue you already sell from.
Inventory management software
Stock that updates itself as it moves, instead of a spreadsheet you maintain.
Shopify vs WooCommerce vs all-in-one
Three different purchases, compared on architecture rather than feature counts.
Go deeper on a feature
Selling at the counter
Split payments, held orders, manager-gated discounts and a drawer count that reconciles.
An online shop off the same stock
A branded storefront on your own domain, reading the same products, stock and prices as the till.
Stock you can trust
Batch and expiry, serial numbers, reorder levels, stock takes and the formulas behind them.
See it for your trade
Clothing boutique
Size-then-colour variants, per-variant stock, the online shop as a second branch.
Electronics & phone shop
IMEI and serial tracking, warranty by unit, consignment stock, high-value write-offs.
Hardware shop
Sell by length or weight, wholesale versus retail lists, quotations, credit accounts.
Try it on your own products
Fourteen days, no card. Import your product list or let Amina build it from a photo of your price list, and run it alongside whatever you use now.