The Friday That Was Not Actually Good

The room was full. The bar was three deep. Everyone worked hard, and the number at the end of the night was the best in weeks.

And the month still came in soft.

This happens often enough that it deserves a proper answer. A big night is not automatically a good night, and the difference is visible in about four numbers.

Revenue alone cannot tell you

A total is the product of two things you should never look at together.

Revenue = heads × spend per head

Those two move independently, and they respond to completely different actions.

Heads up,  spend down  → you got busier and poorer per guest
Heads down, spend up   → a smaller, better night. Often more profitable
Both up                → genuinely good
Both down              → obvious, and rarely the confusing case

The confusing Friday is almost always the first line. Heads went up, spend per head fell, and the total looked fine because volume covered it.

That is not a good night. It is a busy one, and the two cost different amounts to produce.

Why a full room can be worse

A packed room has costs that a three-quarters-full room does not.

Service degrades at the top of capacity. The wait at the bar goes from two minutes to eight. The guest who would have bought a third drink buys two. That is the spend-per-head drop, and it is caused by the fullness itself.

Staff cost is not linear. Past a point you add people to hold the same service level, and the marginal person costs more than the marginal revenue they enable.

The good guests leave first. The table that would have stayed until close leaves at one because it stopped being pleasant. You lose the tail of the night, which is the high-margin part.

And the balk. People who look at the queue and leave are invisible in every report you have. You will never see the revenue you did not take.

Four numbers, per night

Heads                    … door count, not covers
Revenue per head         … the one that carries the most information
Capacity utilisation     … peak occupancy ÷ comfortable capacity
Mix                      … share of revenue by category

Comfortable capacity is not licensed capacity. Licensed is a legal ceiling. Comfortable is where service still works — usually 70–85% of licensed, and every room's number is different. Find yours by looking at the nights where revenue per head held up.

Mix is where the story usually is

Two nights with identical revenue can have completely different mixes, and the mix predicts next month.

Night A … 60% bottles / 25% bar / 15% entry
Night B … 20% bottles / 60% bar / 20% entry

Same total. Different businesses.

Night A is a smaller number of high-spend groups. Night B is volume. A is more profitable and more fragile — losing three tables changes the night. B is more robust and more expensive to run — it needs more staff, more stock movement, more door.

Neither is wrong. But if your mix shifted from A to B without anyone deciding it should, the room drifted and nobody noticed.

The comparison that means something

Not "this Friday vs last Friday." Too noisy — weather, events, paydays.

This Friday vs the trailing 8-Friday median, on all four numbers

Median, not mean. One outlier Friday distorts a mean badly at n=8.

And compare the four separately. A night that is up on heads and down on revenue per head is a specific, diagnosable event. A night that is "up 12%" is not.

What to do with a big night

The instinct is to repeat it. Same promoter, same booking, same door policy.

Check the mix first. If the big night was volume with depressed spend per head, repeating it repeats the cost structure without the margin.

Three questions:

Did revenue per head hold?           If no, you were over capacity
Did the mix hold?                    If no, the guest type changed
Did the tail hold?                   Compare the last two hours to the median

The third is the most overlooked. A night that peaked early and emptied by one made its money in a compressed window with full staffing. The revenue looks the same; the margin does not.

The number that stops the argument

Revenue per head, against the trailing 8-night median for that weekday

One number, one comparison, weekday-matched. It settles most "was that a good night" conversations in about ten seconds, and it moves the discussion from feel to fact.

If it held or rose, the night was good. If it fell while heads rose, you were past capacity — and the fix is a door policy or a service change, not more marketing.

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 by pushing harder on sales.

This guide is one of the arguments that growth settled. Peak nights felt like success and sometimes were not. Revenue per head against the weekday median is what ended the guessing.

What it does here: revenue and guest counts per night, with weekday comparison. Settlement broken out by line — course, extension, nomination, options, transport, discounts — so mix is a field rather than an estimate. Guest records that persist across visits, so you can see whether a big night produced returning guests or one-time volume. Staff hours per shift, so the cost side of a peak is a number.

What it does not do: it does not hold a door count separate from bookings, and it has no queue or balk measurement. Comfortable capacity is a judgment you set.

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.

It bears on this directly. A channel that fills the room with low-spend heads on a night that was already near capacity has negative marginal value — it pushed you past the point where revenue per head falls. The ceiling is computed from lifetime value, which catches this. Head-count-based channel reporting does not.

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 — how last Friday compared to its median, what the mix was, which sources brought the heads.

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 with eight Fridays

1. Pull heads and revenue per head for the last eight Fridays. Take the median of each.
2. Compare your best recent night to those medians, separately. Not the total.
3. Find your comfortable capacity — the occupancy above which revenue per head starts falling.
4. Split revenue by mix for the same eight nights. Watch for drift nobody chose.
5. Check the tail. Last two hours against the median. A compressed night costs more than it looks.

Read next

• The Slow Tuesday Problem
• The Door: Where Your Night Is Decided
• Bottle Service Math: The Table Numbers Most Nightclubs Never Run

---

On benchmarks. No target revenue per head, capacity ratios, or mix splits appear in this guide. We do not have a dataset broad enough to publish them, and they vary enormously by format, city, and price point. Eight of your own Fridays will give you a median that means something. An industry average will not.