What you'll get: a calculator that sizes your counters from your own peak hour, the cheaper fixes to try before you buy a second till, the four things that break the moment you have one, what each failure actually costs, how the day close changes, and the questions to ask a vendor before you sign.
Search for this and you will find a glossary entry telling you a billing counter is where customers pay, and three American payment processors selling card readers. Nobody answers the actual question, which is not what is a billing counter but how many do I need, and what goes wrong when I have two?
The second half matters more than the first. Adding a till is easy — it is a tablet and a drawer. What nobody warns you about is that a restaurant's whole accounting model quietly assumes there is only one place money is taken, and the day you break that assumption, four things stop working at once.
How many counters does your peak hour need?
Size counters from the busiest sixty minutes of your week, not from your floor area or your seat count. Put your own numbers in.
Count them off last Saturday's report.
Ring up to change handed back. Time it.
Nobody can run a till at 100%.
Where that number comes from
No queueing theory, just division you can check:
- One counter clears 3600 ÷ seconds-per-bill bills an hour flat out. At 45 seconds that is 80.
- You cannot plan to run it flat out. Guests do not arrive evenly — they arrive in clumps when a film ends or an office empties. A till planned at 100% of capacity grows a queue during every clump and never gets the idle time back.
- So divide your peak by capacity × the most you are willing to load a till, and round up. 120 bills ÷ (80 × 0.8) = 1.9, so two counters — each then running at about 75%.
The number that moves the answer most is seconds per bill, and it is the one nobody measures. Go and time ten real bills at your own counter this Saturday, from the first item rung to the change back in the guest's hand. If it comes out at 90 seconds rather than 45, you do not have a counter problem, you have a speed problem — and a second till is the expensive way to fix that.
Four cheaper things to try before you buy a second till
A counter is a device, a drawer, a printer, a person and a piece of floor. Try these first; each of them removes load from the counter you already have.
- Pre-bill at the table while they are still eating. The slowest part of a dine-in bill is not ringing it, it is the guest reading it. Hand it over before they ask, and the payment itself takes twenty seconds.
- Take payment where the guest is. If a waiter can settle at the table on a phone, the queue at the counter is only takeaway.
- Give takeaway its own small station. Parcel and pickup orders are fast and repetitive; mixing them into the dine-in queue makes both slower.
- Let guests order for themselves. A QR order that arrives already itemised removes the whole ringing-up step from the counter.
If you have done all four and the queue is still there at 8 pm on a Saturday, you have a genuine capacity problem and a second counter will pay for itself. Now read the rest of this, because that is where it gets interesting.
The four things that break
Every one of these is invisible on the first day and expensive by the first month.
1. Bill numbers collide
Your invoice numbers have to be unique and unbroken — that is not a preference, it is what a tax auditor checks first. With one till it is trivial: the last bill was 41, the next is 42.
With two tills, both of them believe the last bill was 41. Both issue 42. Now you have two different invoices carrying the same number, given to two different customers, and no way to tell which is which after the fact.
There are only two honest fixes:
- Allocate every number from a server. Correct, but it means a network round trip on every bill. When the internet drops — and it will — either billing stops or the tills start guessing, which is the same collision arriving later.
- Give each counter its own series. Counter 1 issues its own run, counter 2 issues a separately marked run. They can never collide, even with both offline all evening, because they are not drawing from the same pool.
Ask this before you buy
“Two of my tills are offline at the same time and both take a bill. What number does each one print?” If the answer is vague, the collision is real and you will meet it during your busiest service.
2. One pooled drawer, and nobody is accountable
This is the one that costs actual money. Two cashiers working one drawer figure cannot be reconciled to a person. When it comes up ₹600 short, both can say honestly that it was not them, and you have no way to test either claim — so nothing happens, and it happens again next month.
One drawer, one cashier, one float, one blind count, one variance. That is the whole control, and it only works if the counters are genuinely separated rather than being two tablets pointed at the same books. We wrote the full method in the 20-minute cash reconciliation; all of it applies per counter.
3. One merged Z-report
If the shift report totals both counters, neither cashier can be cashed up on her own. She cannot leave until the other one finishes, the variance is a single blended number, and the moment there is an argument you are arguing about an average.
Each counter needs its own shift, its own open and close, and its own Z-report carrying its own expected, counted and variance. The day close then adds them up. Reconcile per counter, roll up second — never the other way round.
4. Nothing is attributable
Without per-counter reporting, sales, discounts, voids and variance are all averaged across the tills. That sounds harmless. It is not: it means the counter with a problem is being subsidised by the counter without one, and you cannot see either.
What you want to be able to read, per counter, over thirty days:
| Read per counter | What a difference means |
|---|---|
| Sales and bill count | Which till carries the load; whether the second one earns its keep |
| Average bill | A much lower one usually means it is the takeaway till, not a worse cashier |
| Discount value | One counter giving away twice as much is a conversation, not an accusation |
| Voids and cancellations | A higher rate on one till, repeatedly, is the single best theft signal you have |
| Cash variance, with direction | Errors scatter both ways; one-directional variance is systematic |
None of these are damning on their own. Three of them landing on the same counter, for a month, is a pattern — and the counter-fraud guide covers what to do next.
Closing a day with more than one counter
- Stop billing on that counter and close its shift. Not the whole restaurant — just that till.
- The cashier counts her drawer blind and locks the number before the system shows her what it should have been.
- Read that counter's variance and record it, however small.
- Repeat for each counter. A counter that finished at 9 pm should not be waiting for one that closes at midnight.
- Run the day close last. It adds the counters into one figure for the business — sales, tenders, expected, counted and total variance.
The order is the point. A single combined total tells you money is missing. Per-counter totals tell you where from.
Seven questions to ask a POS vendor
Ask these before you sign, and ask for the answers in writing. Every one of them is something restaurants discover afterwards.
- If two tills are offline at once and both take a bill, what number does each print?
- Can each counter have its own cash drawer, float and blind count?
- Can each counter close its own shift and print its own Z-report, without waiting for the others?
- Can I read sales, discounts, voids and variance per counter, for a date range?
- What does the second, third and fifth till cost me — per device, per month, in writing?
- If a counter's tablet dies mid-service, what happens to its open bills and its shift?
- Can I turn all of this off? A shop that goes back to one counter should get its old behaviour back exactly.
That last one matters more than it sounds. Multi-counter is a mode, not a destination — plenty of restaurants add a counter for a festival season and take it away again.
Two to nine counters, each with its own everything
In Dinelax every billing counter gets its own bill number series, its own drawer and float, its own blind count and Z-report, and its own row in a Billing Counter report — while End of Day still totals every counter into one close. It is off until you switch it on, and a shop that never does is unaffected.
rocket_launch Start free — no card14 days free. Included in every plan — not an add-on.
Frequently asked questions
Size it from your peak hour, not your floor area. One counter clears roughly 3600 divided by your seconds-per-bill: at 45 seconds a bill that is about 80 bills an hour flat out. Plan to run each counter at about 70 percent of that, because arrivals bunch and a till at 100 percent utilisation grows a queue that never clears. So a peak of 120 bills an hour at 45 seconds each needs two counters, and 200 an hour needs four. Most single-outlet restaurants under 150 covers never need more than one.
Four things, and none of them are obvious on day one. Bill numbers collide, because two tills each think the last number was 41 and both issue 42. The cash is one pooled drawer figure, so a shortfall cannot be attributed to a person. The Z-report merges both counters, so neither cashier can be cashed up alone. And sales cannot be read per counter, so you cannot tell which till is slow or which one carries the discounts.
They can, but only if the number is allocated by something both terminals agree on, and that means a round trip to a server on every single bill. That is fine until the internet drops, at which point either billing stops or the tills start guessing. The safer design is one series per counter with its own prefix or marker, so counter 2 issues its own run of numbers and can never collide with counter 1 even while both are offline.
Yes, and its own opening float and its own blind count. A shared drawer worked by two cashiers cannot be reconciled to a person: when it is 600 short, both of them can say honestly that it was not them, and you have no way to test either claim. One drawer, one cashier, one count, one variance. It is the single biggest reason to separate counters properly rather than just adding a second tablet.
That depends entirely on the vendor, and it is worth asking before you buy rather than after. Some price per terminal, so a second counter doubles the bill. Others allow several devices on one restaurant account at no extra cost. Ask for the year-one number in writing with the number of tills you actually plan to run, including the kitchen screen and the waiter phones.
Close each counter first, then the day. Each cashier counts her own drawer blind, locks her number, and gets her own Z-report with her own expected, counted and variance. The day close then adds the counters together into one figure for the business. Reconcile per counter first and roll up second — a single combined total tells you money is missing but not from where.
Usually not. A second counter costs a device, a drawer, a printer and a person, and it only pays for itself if the queue at your peak is genuinely costing you covers. Before adding one, try the cheaper fixes: take payment at the table, pre-bill while the guest is still eating, or move takeaway pickup to its own small station. A counter that is idle for 22 hours a day is an expensive way to solve a 40-minute problem.
You need sales, discounts, voids and variance broken out per counter, over 30 days. One counter with a higher void rate, a higher average discount, or a variance that only ever goes one direction is a pattern worth asking about. Without per-counter reporting all of that is averaged away, and the counter with the problem is subsidised by the one without it.
Start here, tonight
- Time ten bills at your counter on your busiest night. That single number decides everything above.
- Count your peak hour off last Saturday's report and put both into the calculator.
- If it says one counter, you are done — spend the money on the speed of the one you have.
- If it says two or more, ask the seven questions before you buy anything. The tablet is the cheap part; the bill numbers, the drawer and the Z-report are what you are actually buying.
A second counter is not a bigger version of the first one. It is a different accounting model, and the restaurants that get burned are the ones that found that out in month two.
Every counter its own numbers, its own drawer, its own close.
Start free — no card14 days free, no card. Android tablets or a Windows till.