The Comps You Never Approved

Every other cost on your P&L is something you agreed to. Rent, payroll, liquor, the linen contract — you signed for all of it.

Comps are the one line where someone else spends your money, one table at a time, and tells you afterward.

That is not an argument against comping. A comped round is one of the cheapest retention tools a venue has. The problem is not that comps exist. The problem is that in most rooms nobody can answer three questions about them: how much, who, and why.

The number you are looking for

Comps and voids together, as a percentage of gross sales.

Below 2% usually means one of two things: either the floor is disciplined, or your staff have stopped comping entirely because the approval process is painful. The second one costs you more than the first saves.

Between 2% and 5% is where most rooms with an active floor sit. It is not a target — it is a range where the number is probably being used rather than leaking.

Above 5% and the number has stopped being a decision. Something structural is producing it: a promoter arrangement nobody wrote down, a POS button anyone can press, a manager who comps to avoid conversations, or a service problem being papered over drink by drink.

The percentage on its own is not the finding. The percentage over three months, split by who authorized it, is.

Comps and voids are not the same problem

They land in the same bucket and they come from opposite places.

A comp is a decision. Someone chose to give something away — a round for a regular, a bottle for a birthday, a drink to fix a mistake. It has an author.

A void is a correction. The item should never have been rung: wrong drink, wrong table, duplicate entry, guest changed their mind before it was made.

When you look at them together you see a total. When you split them you see two different stories, and they need two different fixes.

A rising comp number is a policy question. A rising void number is a training question — or a shrinkage question, because a void after the drink was poured and served is the oldest way to take money out of a bar.

The single most useful split is void-before-service versus void-after-service. Most POS systems record the timestamp of the ring and the timestamp of the void. If the gap is routinely longer than the time it takes to make the drink, that is worth a conversation.

The comp with no name on it

Ask your POS who authorized last month's comps. In a lot of venues the honest answer is "the manager card," and the manager card was in a drawer that four people had access to.

That is the whole problem in one sentence. A comp without an author is not a comp — it is a hole.

The fix is not a new rule. It is making the existing rule produce a record:

• Every comp carries the ID of the person who approved it, not the person who rang it
• Every comp carries a reason code, chosen from a short list, not free text
• The list is short enough that people actually pick the right one — five or six codes, not twenty

Six codes cover almost everything: service recovery, regular/VIP, staff/industry, promotion, birthday/celebration, management discretion. If a room needs a seventh, add it deliberately and watch it.

The moment reason codes exist, the number stops being one number and becomes a shape. Service recovery at 40% of your comp total is a kitchen or bar problem wearing a comp costume. Management discretion at 40% means the codes are not being used and you are back where you started.

What "service recovery" is actually telling you

This is the code worth watching most closely, because it is the only one that points at something else.

A remade drink is a comp. A remade drink is also a defect. If your service-recovery comps are concentrated on two cocktails, the problem is the spec or the training on those two cocktails, and no amount of comp policy fixes it.

Run the recovery comps by item for a month. In most bars they cluster hard — a handful of drinks account for most of the remakes, usually the ones with the most steps or the least-used ingredients.

That list is a training list, and it pays for itself twice: once in the comps you stop giving, once in the guests who did not have to send a drink back.

The regular who is not a regular

The second-largest comp category in most rooms is some version of "taking care of" a guest.

That is real value. A guest who feels recognized comes back, spends more, and brings people. The question is not whether to do it. The question is whether the guest you are comping is actually worth what you are spending.

Most venues cannot answer this because the comp lands on a table, not on a person. The round goes out, the tab closes, and the guest's name is nowhere in the record.

If your system connects guests to visits, the check is simple: pull the guests who received comps last quarter and look at their total spend across all visits. You will usually find three groups.

• Guests whose spend clearly justifies it — keep going, and consider doing more
• Guests who receive comps regularly and spend at or below average — this is a habit, not a strategy
• Guests who received one comp and never returned — the comp did not fail, but it did not do the job it was given

The middle group is the one that costs real money, and it is almost always invisible without guest-level records. Nobody decided to comp that person every week. It just became what happens when they walk in.

Industry drinks and the shift that never ends

Staff and industry comps deserve their own paragraph because they are the easiest to defend and the hardest to see.

Buying the bartender from the room down the street a drink is how the trade works. It is genuinely good for business. It is also the category most likely to drift, because there is no guest to notice and no tab to close.

The check that works: cap it per shift, per person, and let the cap be generous. A generous cap that is recorded beats a vague understanding that is not. Staff will tell you what a reasonable number is if you ask them, and the number they give you is usually lower than what is happening now.

Three checks that hold

One — comps by authorizer, monthly. Not to catch anyone. To see the distribution. If one person is 60% of the total, either they work the busiest shifts or something needs a conversation. Both are worth knowing.

Two — void-to-ring time gap. Anything voided more than a few minutes after it was rung, listed by shift. The list is short in a healthy room. If it is not short, you have found something.

Three — comp percentage by night of week. A Saturday comp rate double your Tuesday rate is not automatically wrong — bigger nights carry bigger tables and more promoter arrangements. But the ratio should be stable over months. When it moves, something changed in how the busiest nights are being run, and you want to know what before the quarter closes.

What this looks like when it works

You do not end up comping less. In most rooms that tighten this up, the total stays roughly where it was.

What changes is where it goes. Service-recovery comps fall because the underlying defects got fixed. Discretionary comps rise, because the floor now has a clear allowance and uses it on the guests who justify it. The category that disappears is the one nobody could explain.

The goal is not a smaller number. It is a number where every line has an author, a reason, and a guest attached to it.

At that point the comp line stops being a leak and starts being what it should have been from the start: a marketing budget spent one table at a time, by the people closest to the guest.

Where the record has to live

None of this works on paper, and it does not work in someone's memory of a busy Saturday.

It works when the comp, the authorizer, the reason, and the guest are recorded at the moment the drink goes out — because that is the only moment when all four are known. An hour later the table has turned over and the reason is gone.

tasteck records comps against the guest, not just the tab, so the quarterly question — what did we spend on this person, and what did they spend with us — has an answer that does not require anyone to remember the night.

That is the whole difference. Not stricter rules. Just a record that survives the shift.