Salon POS, Stock and Commission: Getting the Three Numbers to Agree

Why salon numbers drift

A shop sells a product and the stock reduces. A salon sells a stylist's time, sells products alongside it, uses products during treatments, and pays commission on some of that. Four different movements, one shelf, and usually two or three systems that each know about part of it.

The drift is not caused by dishonesty most of the time. It is caused by a walk-in charged without a stylist selected, a colour mixed from a tub nobody logged, a retail sale rung through as a treatment because it was faster. Each is trivial. By month end they add up to a stock count that cannot be explained and a commission figure somebody disputes.

Salon POS in Dubai: the front desk is the real problem

The counter is the busiest point in the business. It takes bookings, handles walk-ins, sells retail, processes payment and answers the phone, usually staffed by whoever is free rather than by someone trained on the system. Every second of friction there happens in front of a client.

That is why a slow front-desk computer is not an inconvenience but a cost. Most are slow for three reasons: a mechanical hard drive, too little memory, and years of software that starts at login. Replacing the drive with an SSD is usually the single biggest improvement available and costs a fraction of a new machine.

The second requirement is that the diary and the till are on one screen. If adding a walk-in means switching applications, staff will take the payment and add the appointment later, which is exactly how a sale ends up with no stylist against it.

Making stock honest

The only arrangement that works is one system where all three movements reduce the same stock figure: a retail sale, an internal usage entry for product consumed in a treatment, and an adjustment for anything else. If professional-use product is tracked separately from retail, or not tracked at all, the count will never reconcile.

Then make adjustments attributable. With named logins, every write-off has a person and a timestamp against it, which changes the conversation entirely: a monthly count becomes a check rather than an investigation, because the discrepancies have an audit trail.

Counting frequency matters more than count accuracy. A small count of fast-moving colour lines weekly catches drift while it is still explainable. A full count quarterly produces a large number nobody can account for, which is demoralising and useless.

Commission without arguments

The link between discounts and commission is worth dwelling on. Most commission disputes we see are not about the rate; they are about a discount applied at the desk that reduced the commissionable value, with no record of who approved it. Once discounts are attributable, the argument disappears because the data answers it.

• Make stylist selection mandatory before a charge can be completed.
• Split commission rules for services and retail, since they are rarely the same rate.
• Handle the shared-service case explicitly, where two stylists worked on one client.
• Produce the commission report from the system that took the money, not a spreadsheet.
• Give each stylist visibility of their own figures during the month, not just at the end.
• Log discounts and who authorised them, because discounted services change commission.

Walk-ins are where the data breaks

Booked appointments behave well because the booking already carries the client, the service and the stylist. A walk-in carries none of that. Someone arrives, a stylist who is free takes them, and the transaction is created at the till after the service rather than before it.

That is the moment most bad data is produced. The service is rung through as whatever is closest on the button layout, the stylist is whoever was standing there, and the client is not recorded at all because entering details takes longer than taking the cash. Multiply that across a busy Saturday and both the commission report and the client list are wrong.

The fix is to make the walk-in path as short as the booked one. A quick-create client entry needing only a name and a phone number, a mandatory stylist field, and a button layout that matches what the salon actually sells rather than the supplier's default. Salons that get this right are usually the ones that redesigned the till screen, which is configuration nobody thinks to ask for.

Client records are more sensitive than owners assume

A salon holds phone numbers, treatment and colour history, allergy and patch-test notes, and very often before-and-after photographs. Patch-test records are health-adjacent information and the photographs are personal data about an identifiable person.

The photographs are the weak point in almost every salon we assess. They sit in a staff member's personal phone gallery, which the salon does not own, cannot audit and cannot clear when that person leaves. They belong in the salon's own system, with access limited to the staff who need them, and the phone folder cleared once they are moved.

Anything reachable from outside the salon should require multi-factor authentication, which Microsoft describes in its guidance on how MFA works. And cloud-held records still need a separate backup, because a deletion inside a platform is not something the platform will reverse for you; Microsoft draws the same distinction in its Microsoft 365 Backup documentation.

Repeat business is the number that matters

Most salon reporting focuses on daily takings, which is the least useful figure available. The number that predicts the business is how many clients come back, and how long they take to do it. A salon with strong takings and weak rebooking is quietly living on new clients, which is the most expensive kind.

Any competent salon system can report this, and almost none are configured to. What you want is rebooking rate by stylist, average gap between visits, and a list of clients who have not been in for longer than their usual interval. That last list is the most directly useful thing software can give a salon, and it is generally sitting there unused.

It only works if client records are captured properly, which loops back to the walk-in problem. A salon that does not record who walked in cannot tell whether they came back, and so cannot see the one trend that matters most.

The network under all of it, and what fails when it is busy

Two things cause most salon technical faults. The card terminal shares a network with client Wi-Fi, so a client streaming video slows a payment. And a single access point at the front does not reach the back of the salon, so staff at the rear basins have no signal.

Both are straightforward. Client Wi-Fi goes on a separate network with a bandwidth limit, the terminal goes on its own segment, and coverage is measured on the actual floor rather than assumed. Our network and Wi-Fi service covers the design, and it is normally a small job rather than a rewire.

The third fault is the printer, which is nearly always a queue or driver problem rather than hardware, and is worth fixing properly once rather than restarting weekly.

The pattern worth noticing is that most of these are capacity problems rather than faults. Everything works on a quiet Tuesday and fails on the day it matters. That is why we assess salons during trading hours rather than in the morning, because a salon at ten and a salon at seven in the evening are genuinely different networks.

• The card terminal, when client Wi-Fi is saturated and both share a network.
• The front-desk machine, which is the oldest computer and has been running since eight.
• Wi-Fi at the back basins, where the single front access point does not reach.
• The receipt printer, generally a queue problem rather than a hardware one.
• Staff taking shortcuts at the till because the queue is long, producing the data problems above.
• The internet line, with no second path, taking bookings and payments down together.

Staff turnover is a data problem too

Salons have high turnover, and access tends to lag behind it. A former stylist whose login still works can see the client list, which for a salon is the most commercially valuable thing it owns.

Offboarding should be a written step someone owns: on the last day, the named login is disabled, any shared passwords that person knew are changed, and access is verified rather than assumed. Our employee offboarding checklist sets out the order.

The reason it needs to be written rather than remembered is that departures are rarely tidy. The ones that matter most are the abrupt ones, which is precisely when nobody is thinking about logins.

What to fix first

Start with mandatory stylist selection and named logins, because together they fix commission attribution and make everything else auditable, and neither costs anything beyond configuration time.

Then move client photographs off personal devices into the salon's system. That is the highest-risk item in most salons and it is resolved in an afternoon.

Then the front-desk machine and the network separation, which are small spends with immediate daily effect. Only after that is it worth considering new software, because most salons discover the system they already pay for does what they need once it is configured properly. If you want that assessed rather than guessed, our salon IT page explains what an assessment covers.

One last point on sequencing. Do not buy new software to solve a data problem until the current system has been configured properly, because in most salons we assess the software already does what is needed and nobody finished setting it up. Changing products resets the staff training, moves the client list, and leaves the same habits in place at the till, which is how a salon ends up with a second system and the same unexplained stock count.

Frequently asked questions

Written by Srinivasan Ramachandran, Founder, Listonics · Published Sep 24, 2026 · Last reviewed October 11, 2026 · 2,043 words