The Door Count That Tells You Nothing

Door count is the easiest number in the building to collect and the hardest to act on.

Six hundred through the door on Friday, six hundred again the following Friday, and the second night did substantially less. Nobody can explain it, so it goes down as "quieter crowd" and the count keeps getting recorded because it always has been.

A number you collect every night and never act on is not a metric. It is a habit.

The fix is not to stop counting. It is to record two or three things alongside it that turn the count into something that predicts.

Why the count alone cannot work

Door count treats every guest as identical. Your revenue does not.

On a typical weekend night, a large share of the spend comes from a small share of the guests — the tables, the bottle buyers, the groups that arrive early and stay. The rest of the room is atmosphere, and atmosphere is necessary, but it is not where the night is decided.

Two nights with the same count can differ enormously in how many of those higher-spend guests were in it. The count cannot see that, which is why it cannot explain the gap.

The first thing to add: arrival curve

Not how many came in, but when.

Record the count in hourly buckets rather than as a single total. It takes no additional effort at the door and it changes the number completely.

Two nights with six hundred each look identical. Split by hour and one shows a steady build from opening with a peak at midnight, the other shows almost nothing until 23:30 and then a wall of people.

Those are different nights with different economics. The early-build night has guests who ate, drank through multiple rounds, and settled at leisure. The late-wall night has guests who came for the last ninety minutes, ordered once, and left.

The arrival curve is the single most useful thing you can add to a door count, and it costs one extra column.

Once you have it for a few weeks you can also see what moves it — which promotions pull people earlier, whether an entry-price change shifts arrivals, whether the room down the street closing at a different hour is feeding you.

The second: revenue per guest, by night

Divide the night's revenue by the door count. Track it over weeks.

This is a blunt instrument and it is still more informative than either number alone. A falling revenue-per-guest with steady count means the crowd is changing, and it usually changes for a reason you can find: a promotion that pulled a cheaper crowd, an entry price that filtered out the guests who used to arrive early, a promoter widening their list.

A rising count with falling revenue per guest is one of the clearest warning signs a venue produces, and it is invisible if you only watch the count.

Look at it by night of week. Your Tuesday number and your Saturday number should be different, and each should be stable against itself.

The third: turned-away and capacity time

Most venues do not record the hours they were at capacity, and almost none record turn-aways.

Both matter. A night at capacity from 23:00 is a night where your count is a ceiling rather than a result — it measures the room, not the demand. Comparing that count against a night where you never filled is comparing two different things.

The practical version: record the clock time you hit capacity, and a rough turn-away count in fifteen-minute blocks at the door.

Rough is fine. The purpose is not precision; it is knowing whether the constraint was demand or the room. Those two situations call for opposite responses, and the count on its own cannot distinguish them.

What the combination shows

With arrival curve, revenue per guest, and capacity time recorded for six weeks, questions that were previously arguments become readable.

Should we open earlier? Look at the arrival curve on your strongest nights and whether the early hours produce spend or just staffing cost.

Is the entry price right? Watch what a change does to arrival timing and to revenue per guest, not to the count. Count is the least sensitive of the three.

Are we actually full, or does it just feel full? Capacity time answers this, and the answer frequently surprises people who work the floor, because a busy room feels full an hour before it is.

Is this promotion working? Compare revenue per guest and arrival curve on promotion nights against matched nights, rather than comparing counts.

The mistake worth avoiding

The temptation once you start measuring is to optimize for revenue per guest, which pushes toward a smaller, higher-spending crowd.

That is a real strategy and it is not automatically the right one. A fuller room does work you cannot see in a monthly report — it sets the perception that determines next month's arrivals, and it is what makes the venue feel like somewhere to go.

Revenue per guest is a diagnostic, not a target. Watch it for direction and for surprises. Do not manage the door to it, or you will optimize your way into an empty room with good margins and a declining future.

What a workable record looks like

At the door, per night:

• Count in hourly buckets
• Time capacity was reached, if reached
• Approximate turn-aways in fifteen-minute blocks during any capacity period
• Anything unusual in one line — weather, a nearby event, a competitor closed

That last line is the one people skip and the one that saves the analysis. Six weeks later, nobody remembers that it rained hard from ten until midnight, and that night will sit in the data as an unexplained trough.

The join that makes it useful

Door data on its own is still just door data. It becomes operationally useful when it sits next to the sales for the same hours and the staffing for the same shift.

Arrival curve plus sales by hour tells you whether your peak is a revenue peak or just a crowd peak. Arrival curve plus staffing tells you whether you are rostered for when people actually arrive or for when you assume they do — and those two answers differ in most venues by roughly an hour.

tasteck keeps door records, hourly sales, and shift assignments against the same night, so those comparisons are one view rather than three spreadsheets and a reconstruction.

The count is not the problem. Counting it alone is. Add the hour it happened and what it spent, and the number you have been collecting for years starts answering questions.