"Johnny Yang, it's not that I don't want to pay them. It's that I have to recalculate everything every month, and then explain it separately to three different people."
The person saying this was the director of a medical aesthetics clinic, on a video call a little past 11 p.m. He shared his screen: one Excel file, seven worksheets. The consultants' closed-deal sheet, the doctors' session sheet, the front desk's add-on sheet, and a few lookup tables only he could read. He scrolled as he talked. This client got a refund last month, so it has to come back out of the consultant's bonus. That client is paying in installments and has only paid the first one, and he still hasn't decided whether the revenue counts in the month of the deal or the month the money arrives. And there's a batch of treatment packages sold six months ago that the clients never came back to use. Revenue on the books. Money not actually earned yet.
Each of these, on its own, is a ten-odd-minute judgment call. Added up, they cost him two or three evenings every month, plus the half hour the next morning when a consultant pushes open his door and asks, "Did I get shortchanged this month?" The amounts are usually small. What gets damaged is trust. I've seen it at more than one clinic: the strongest consultant quits, and the trigger isn't base pay. It's several months in a row of feeling her numbers are calculated differently from everyone else's, and never getting a straight answer.
Conclusion first: whether commissions can be automated has little to do with which system you buy. It depends on whether your scheme has been written as a rule a machine can actually decide. For a rule to be automated, it has to state three things at once: whose credit it is, which event triggers it, and in what order exceptions are handled. Miss any one, and the system just stops and waits for someone to tell it what to do. Your "automation" becomes someone going in every month and manually editing dozens of records. Worse, when a clinic with an unclear scheme buys a system, the disputes don't go away. They move. Before, the argument was "You calculated it wrong." After, it's "The system calculated it, and I don't know why either." Now there's a machine vouching for the number. Staff find it harder to challenge, you find it harder to explain, and the internal friction gets deeper. So the order is always the same: pin the scheme down until it's decidable, then talk about automation. I have not seen a single clinic do it the other way around and have it go smoothly.
I'm Johnny Yang, from Brightstream Technology in Taiwan. I've led teams doing implementation for seven years: 252+ projects across 32+ industries.
Why commissions go wrong so easily
Let me take your side for a moment. Your instinct that "this should be something a system can handle" points in the right direction. Calculating commissions is, at its core, adding up data according to rules. Machines do that faster and more consistently than people. The only mistake is thinking what you're missing is a tool.
Commissions differ from everything else in a clinic because they are three things at once. They're money, so every dollar that's wrong gets remembered. They're an evaluation: staff read what you value from how bonuses are structured. And they're a promise. That line from the interview, "Do well and I'll take care of you," gets either honored or discounted every month. Vagueness in other processes is a cost. In commissions it's a fuse. Get inventory wrong, add a correcting entry, and it's done. Get commissions unclear, and the employee won't blow up on the spot. They'll remember it, until another clinic tries to poach them, and then it all comes out at once.
And medical aesthetics clinics happen to have the most complicated structure of all. One purchase can involve the consultant, the doctor, the nursing side, the front desk and whoever in marketing brought the client in. Each share is counted on a different basis: some on the closed deal, some on treatment performed, some on payment received. That's the nature of the industry. The problem isn't the complexity. It's whether you've written the complex parts down.
Automation requires decidability: every rule has to pass three tests
When I judge whether a commission rule can be automated, I look at three things only. Try them on your current rules yourself.
The first test is whose credit it is. Whoever the bonus goes to must have a corresponding field in the data. Consultant A brought the client in, A was off that day, and B closed the deal. Whose is it? The client asked for a specific doctor, but a nurse performed the treatment. Who gets credit on the treatment side? You have answers to all of these in your head, but those are calls you make in the moment, not written rules. A system has no "depends on the situation" function.
The second test is which event triggers it. At what moment does the revenue count: when the contract is signed, when the payment lands, or when the treatment is actually performed? Clinics use all three, but you have to pick one, and you have to be able to explain it. Pick the closed deal, and selling a lot doesn't mean collecting it. Pick payment, and staff feel held hostage by how the client chooses to pay. Pick treatment performed, and the front-line people complain that they did the hard work of bringing the client in and now have to wait six months. What you're choosing is which cost you're willing to carry.
The third test is the order of exceptions. One purchase often triggers several rules at once: a returning client, who also used a new-client promotion, and was referred by another consultant. When all three apply, which wins? Can they stack? Is there a cap? This test gets skipped the most, because owners think "that rarely happens." My experience is the opposite: it eats most of the hours you spend on manual adjustments every month.
Rules that pass these three tests can be built in most systems. Rules that don't, you can switch systems ten times and get the same result.
Put every role's basis on the same sheet of paper
On every clinic project, I ask the director personally to fill in the table below once. On paper. Then read it out to the managers. The point isn't my suggestions. It's that you have to make a decision on every row.
| Role | Common basis | The one sentence you must settle before going on a system |
|---|---|---|
| Consultant | Closed-deal amount or number of closed deals | How to split when one person brought the client in and another closed; whether referrals count |
| Doctor | Sessions performed or procedures performed | For multi-session treatments, count every session or count the package once |
| Nurses and therapists | Number of assisted procedures or independently performed procedures | Which count as independent and which only as assisting |
| Front desk and admin | Add-ons, follow-up appointments booked | Whether add-ons the client chose on their own count for the front desk |
| Marketing and online customer service | Lead source, show-up rate | When a client came in through two sources, which one gets it |
| Director or manager | Overall target achievement or team bonus | Whether the deals you closed yourself are counted twice |
The right-hand column cannot be left blank. Every blank row is something you'll have to rule on personally, every month. Once you fill it in, you'll probably find that the special arrangements you made for individual employees over the past few years were never withdrawn and never written into the scheme. They have to be dealt with before you go on a system: either withdraw them, or formally adopt them as general rules. Leave them alone and they become the gap where "the system can't do this, so we calculate it separately in Excel." One gap is enough. The automation you paid for might as well not exist.
What actually gets stuck isn't commission rates. It's refunds, installments and unused treatments
When most directors talk about automating commissions, they talk about commission rates. But on the ground, what forces the calculation back into manual work is almost never the rate itself. It's these three things that happen afterward.
Refunds: go back, or deduct going forward
Going back means recalculating that month's bonus as if the sale never happened. Deducting going forward means offsetting it against the bonus in the month of the refund. Going back touches salary that's already been paid, which is messy for both the books and the employee. Deducting going forward is easy to execute, but if that employee's numbers were already weak this month, the deduction can take them to zero or below. Does a negative carry over to next month? What if the person has already left? Should a refund because the client changed their mind be handled differently from one over a dispute about results? The system won't think through a single one of these for you. If you don't decide, it stops there every month and waits for a person.
Installments: the gap between revenue and cash flow
The revenue counts the moment the contract is signed, but the money comes in over several months. If bonuses follow the closed deal, you're paying out money you haven't received. The safer approach is two stages: pay part of the bonus when the deal closes, and the rest once payment is complete. But for that to work, the system has to see "this installment belongs to this contract and this client," not just a lump-sum bank deposit. In many systems, payment records and sales records are two separate ledgers reconciled by hand. Then staged payouts can't be automated. Ask about this directly when choosing a system.
Unused treatments: revenue on the books, a liability in reality
A treatment package is sold, and the client hasn't come in yet. Financially, that money is a prepayment: a service you owe the client. But on the sales sheet, it has usually already been booked as a closed deal, and the bonus has been paid. The better you sell, the bigger the unused balance, the prettier the books, and the more the service load and refund risk pile up at the back end.
So you first have to decide what unused treatments mean in your scheme. No effect on bonuses? Or, past a certain period, a provision, with part of the bonus held back until the treatment is used? This depends on whether your package deduction logic is clean. If you have to phone someone to find out how many sessions a client has left and which ones were complimentary, commission automation isn't even on the table. I broke this down in "How to deduct treatment packages without starting a fight". Treat it as groundwork.
For clinics with an unclear scheme, buying a system just moves the disputes into the system
When commissions are calculated by hand, disputes at least have a way out. An employee thinks something's wrong, comes to ask, you explain once, and if it was wrong you fix it. That takes time, but it's alive: the vague parts get forced into the open.
Once you're on a system, if the scheme is still vague, the system runs on whatever interpretation someone casually configured, and the numbers it produces carry the authority of "the computer calculated it." When staff come to ask, the manager's standard answer becomes "that's just how the system calculates it." Once that sentence is said, the discussion is over. The problem isn't. It just goes from the table to underground. Staff start comparing bonuses with each other in private. Managers stop explaining, because they aren't sure why the system calculated it that way either.
Among the failed implementations I've seen, one kind is especially painful. The system runs fine, the data is correct, and nobody trusts it. The few rules that were never settled got handled with "we'll set up a common approach for you for now," and the director didn't check them one by one at acceptance. Six months after launch, the disputes blow up, and it turns out the rules in the system were never the same as the ones in the director's head. Fixing them means touching six months of bonuses already paid.
So for acceptance testing on commissions, I always require this: run last month's real data through the system and compare it line by line with the manual calculation. For every line that doesn't match, someone has to be able to explain which one is right. Skip this, and you're testing a new system with your employees' pay. I covered how to choose and accept a clinic system overall in "How to choose a medical aesthetics clinic system". Commissions are just the one part where you can least afford to settle.
When you shouldn't rush into a system
I don't like seeing people spend money on something that doesn't really hurt yet. If your clinic is still small, few people share in commissions, and calculating them isn't yet crowding out other work, what you need isn't a system. Write the scheme as a one-page document, hand it out, let everyone argue it through once, and after it's final, calculate by hand against it for three months. Those three months will surface every exception you didn't think of, and finding exceptions is faster and cheaper by hand than by system. The other situation where you should hold off is when you're planning to change the bonus structure. If the scheme is going to change anyway, moving the old one into a system first means paying to lock in something you aren't happy with yourself.
The signals that it's time to act are actually clear. Someone regularly spends more than a day every month reconciling commission figures. The same kind of dispute has come up two or more times in six months, and the ruling wasn't quite the same each time. Or you've already lost an employee you didn't want to lose because the numbers couldn't be explained. By then, what you're paying already exceeds the cost of a system.
What you can do tomorrow
Copy the table above and fill in the right-hand column, row by row. Don't skip the ones you can't fill in. Circle them. Those rows are the real source of friction in your clinic right now, and exactly where you should press hardest when you talk to vendors.
Next, take last month's commissions and pick three records you adjusted by hand. Write down why you adjusted each one. Those three will tell you how far your scheme is from decidable. If every reason can be written as a rule without "depends on the situation," you can start talking about automation. If not, write the scheme first.
Then set aside an afternoon and settle one rule each for refunds, installments and unused treatments, all on the same page. These three don't need perfect answers. They need answers everyone knows. Once they're settled, you'll find that even before any system, monthly disputes drop by more than half.
None of these three things cost money, and nobody can do them for you. The scheme is your promise, and only you can decide what you're promising. Do them first, then go look at systems. The questions you ask will be completely different.
If calculating commissions already keeps you up late every month, or you've just had an argument over how they're calculated, I offer a free 30-minute diagnostic of your process. Walk me through how you do it now, and I'll tell you whether to fix the scheme first or change the system first.




