Comps and House Accounts: What They Actually Cost

Every night venue gives things away. A bottle for a regular's birthday. A round for the promoter's table. A tab that the floor manager signs off and nobody chases.

None of it appears on an invoice. So most rooms never total it, and the number stays invisible until the month is tight and nobody can explain why.

Comps are not the problem. Untracked comps are.

First, price a comp correctly

Most operators either ignore comps entirely or value them at menu price. Both are wrong, and they are wrong in opposite directions.

Method	What it says	Why it misleads
Ignore it	"It cost nothing"	The bottle still left the store
Menu price	"That bottle cost us £300"	You would not have sold it to that guest anyway
Cost of goods	"That bottle cost us £40"	Closest to the truth for a comp given to someone who was not buying
Menu price	Correct only when the comp replaced a sale you would have made

The test is simple: would that guest have paid?

Guest was going to buy anyway  → the comp cost you the full menu price
Guest was never going to buy   → the comp cost you the goods

A comped welcome drink to a first-time guest costs the pour. A comped bottle to a regular who buys one every week costs the bottle.

Most rooms use one number for both, and it is always the wrong one for half the cases.

Four numbers, monthly

Comp value at cost      … goods that left without payment
Comp value at menu      … what the same items would have sold for
Comps as % of revenue   … at cost. Sanity number
Comps by issuer         … which staff member signed off

The last one is the one nobody keeps, and it is the one that changes behaviour.

Not because people are dishonest. Because an unmeasured discretion drifts. A manager who comps three rounds a night is not stealing — they genuinely believe each one earned goodwill. They have simply never seen the monthly total of their own decisions.

Showing someone their own number is usually enough. No policy required.

House accounts are a different problem

A comp is a decision. A house account is a loan.

Comp          … you gave it away. Decided, closed
House account … you extended credit. Open until settled or written off

House accounts are how venues get quietly wounded. They start with a good regular who forgot their card. They end with a balance nobody wants to raise because the guest is a friend of the owner.

Three rules keep them survivable:

One — a house account needs a ceiling and a date. Not "we'll settle it eventually." An amount and a day.

Two — someone owns each balance by name. Not the venue. A person, who is the one who will make the call.

Three — aged balances get reported weekly, not monthly. By the time a monthly report shows a 60-day balance, it is a 60-day balance. A weekly one shows it at 7 days, when a text message still fixes it.

A house account that nobody has looked at in three weeks is not a receivable. It is a comp that has not admitted it yet.

Who should be allowed to comp

The instinct is to restrict it to management. That usually makes things worse, because the floor staff who see the moment worth rewarding then have to find a manager, and the moment passes.

A better structure separates size from frequency:

Floor staff   … small value, unlimited frequency, logged
Manager       … medium value, logged with a reason
Owner         … anything

Everything logged, nothing blocked. The point is not to stop comps — comps work. The point is that at the end of the month you can see who gave what to whom, and decide whether it paid.

Did it pay?

This is the question comps exist to answer, and almost nobody asks it.

Comp payback = revenue from that guest after the comp
             ÷ cost of the comp

Track it per guest, not in aggregate. The aggregate always looks fine because a handful of regulars carry it.

What usually shows up:

• Comps to existing regulars: good payback. They were coming anyway; the comp buys loyalty cheaply
• Comps to first-time guests: highly variable. Some convert, most do not
• Comps to industry and staff friends: near zero payback, and often the largest bucket by value

That last line is where the money is. It is also the hardest conversation, which is why the number needs to exist before the conversation, not during it.

The number that decides the policy

Comps at cost ÷ revenue

There is no correct value. A room building a regular base should be higher. A mature room should be lower. What matters is that the number is known and moving in the direction you chose.

If you do not know it, it is drifting upward. That is the default for anything unmeasured that feels generous in the moment.

What to record

Value at cost   … what left the store
Value at menu   … what it would have sold for
Who issued it   … by name
Which guest     … so payback is computable
Reason          … one word is enough

Five fields at the point of issue. If comping requires a form, it stops happening in the moment it should. If it requires nothing, it stops being visible.

The system we built

tasteck is a booking and analytics system for night venues, built by people who ran them for sixteen years and grew from ¥200 million to ¥1.2 billion a year — six-fold, by attacking the business with systems rather than pushing harder on sales.

Why that matters here: comps were one of the lines we cut. Not by banning them — by making the monthly total per issuer visible. The behaviour changed before any policy did.

What it does: guest records that persist across visits, so comp payback is computable per guest rather than in aggregate. Settlement with a line-item breakdown — course, extension, nomination, options, transport, discounts and events — so a discount is a field rather than a note. Staff records tied to each transaction. Bookings, dispatch, and a reservation screen wired to inbound calls.

What it does not do: there is no dedicated comp module with approval tiers. Discounts and events are a settlement line. The ceiling-and-date discipline for house accounts is yours to run. What the product supplies is the guest-level history that makes payback answerable at all.

The output no one else produces

tasteck outputs the maximum you can spend on each marketing channel next month, as an amount in your currency.

A comp is acquisition spend that skipped the budget. If a comped first-visit converts a guest who returns for two years, that comp was cheaper than the channel that delivered them. If it does not, it was a channel with no invoice and no ceiling.

The ceiling is computed from the lifetime value of guests each source actually delivered, with your target margin applied. Nothing else in the nightlife category produces that figure.

Ask it from ChatGPT

tasteck connects to ChatGPT over MCP: ask your numbers as a question and the answer comes back in the chat — who spent what last month, which guests are overdue, how each source performed.

Measured against every vendor listed on Japan's principal nightlife-industry directory, this is the first implementation of it in the category, and the same interface is callable from anywhere rather than being tied to one assistant.

Multi-language is built in, the operating surface itself, with your language set put in place during onboarding.

From $34 a month for up to two venues. Thirty days free on every plan, cancel any time. → Pricing

Start this month

1. Total last month's comps at cost. Most rooms have never seen this number.
2. Split it by issuer. Show each person their own line. Say nothing else.
3. Put a ceiling and a date on every open house account. Aged balances go weekly, not monthly.
4. Compute payback for one month of comps, per guest. Not aggregate.
5. Then decide the policy. Not before — you will guess, and the guess protects whoever is loudest.

Read next

• The Cash Room: Reconciliation That Actually Reconciles
• What to Pay a Promoter
• Four Questions Your POS Cannot Answer

---

On benchmarks. No target comp ratios, house-account limits, or payback multiples appear in this guide. We do not have a dataset broad enough to publish them, and they vary enormously by format, city, and how much of the room runs on regulars. One month of your own numbers, split by issuer, will tell you more than any industry average.