The Report That Takes Too Long To Open

Nobody ever decides to stop checking their numbers.

What happens is smaller than that. A report that used to open takes a few seconds. Then longer. Then the floor manager clicks it at 7pm, waits, gets pulled away by the door, and comes back to a screen that has finished loading and no longer matters.

⚠️ After a few weeks of that, they stop clicking it. They are not being lazy. They have learned, correctly, that opening it costs more than the answer is worth.

The short version

• A slow screen does not get complained about. It gets abandoned.
• The cost is not the seconds. The cost is the decisions that stop being made.
• The screens that matter most are the ones opened during service, not at month end.
• Averages hide this completely. The tail is what people experience.
• Ask your staff which screens they have stopped opening. They will know instantly.

Why nobody reports it

What you expect	What actually happens
Staff complain the report is slow	They quietly stop opening it
You hear about it in a meeting	It never comes up — it is not a "problem", it is just how it is
The number is wrong	The number is fine. Nobody is looking at it.

The third row is the dangerous one. Nothing breaks. No error appears. The data is correct and complete and entirely unread.

One: the screens that matter are the ones used mid-shift

There are two kinds of report in a venue.

Month-end reports  - opened once, in an office, sitting down
Mid-shift screens  - opened on a phone, standing up, with a guest waiting

A month-end report can take thirty seconds and nobody minds. Somebody is sitting down with coffee.

⚠️ A mid-shift screen has about three seconds before it gets abandoned. The floor manager is between the door and the VIP room. If the answer is not there, they guess.

What gets guessed

"Is she free at 11?"          - guessed from memory
"Has this guest been in before?" - guessed from a face
"Are we covered for the 1am rotation?" - guessed, then fixed at 1am

Every one of those has a correct answer sitting in your system. The guess happens because the correct answer arrived too late.

Two: the average will tell you everything is fine

This is the part that fools most owners, and it fooled us.

Average load time: 290ms    - looks excellent
Slowest load:    4,804ms    - what the floor manager actually hit

Both numbers are true. The average is dominated by the small, fast queries — opening a single guest record, checking one shift. The slow ones are the wide screens: a full week, all rooms, every booking.

⚠️ The wide screens are the ones people use to make decisions. So the average is measuring exactly the calls that do not matter, and averaging away the ones that do.

What to look at instead

Not: the average
But: the slowest one in ten, on the screens people open during service

If you have somebody technical, ask them for the slowest ten percent, not the mean. If you do not, use the next method.

Three: ask, do not measure

You do not need instrumentation to find these. Ask three people three questions.

One   Which screen do you open most during a shift?
Two   Which screen have you stopped opening?
Three When you need a number fast, what do you do instead?

The third question is the one that pays. The answer is usually "I ask [somebody]" or "I check the paper sheet" or "I just remember."

⇒ Every one of those is a workaround somebody built because a screen was too slow. The workaround is now the process, and nobody wrote it down.

What the answers look like

"I keep a notebook for the regulars"     - the guest record is too slow
"I text the driver instead of checking"  - the dispatch screen is too slow
"I just wait for the end-of-night sheet" - the live numbers are too slow

⇒ Related: the weekly number your staff never see covers what happens when the number exists but never reaches the floor.

Four: what a slow screen actually costs

It is tempting to price this in seconds. Four seconds, fifty times a night, is three minutes. That sounds like nothing.

⚠️ The seconds are not the cost. The cost is the decision that got made on a guess.

Guessed availability  - a booking that could have been taken, was not
Guessed guest history - a regular treated like a walk-in
Guessed coverage      - a room left uncovered at peak

Each of those is worth more than an hour of anybody's time. And none of them show up as a complaint.

The compounding part

New staff are worse at guessing than experienced staff. So a slow screen falls hardest on the people who most need it.

⇒ That shows up later as "the new floor manager is not working out," when what actually happened is that the tool they were handed was unusable during service.

Five: the fix is usually narrow

The instinct is to rebuild. Usually you do not have to.

In our own system, the slowest screen was a booking view that took nearly five seconds on a busy venue. The cause was not the amount of data and not the design. It was a single missing index on the table — a lookup shortcut the database needed and did not have.

Before  4,804ms at the slow end
After     429ms at the slow end
Change  one index, added in 27 seconds, no downtime

⚠️ We had been looking at the screen layout and the number of things it loaded. Neither was the problem. The database was reading 126,000 rows to return 172.

⇒ The general lesson is not about indexes. It is that slow screens usually have one cause, not many, and you will not find it by redesigning.

How to get it looked at

If you have a vendor or a developer, the useful request is not "the system is slow." It is:

One   Which screen (name it)
Two   What you were doing (filters, date range, how many rooms)
Three Roughly how long (count out loud - "about four seconds")
Four  Whether it is always or only sometimes

The fourth one matters most. "Only on Fridays" and "only for the whole-week view" are both enormous clues. "It is slow" is not.

Six: the screens to check first

In rough order of what tends to hurt:

One   The schedule view for a full week
Two   The booking list with filters applied
Three The guest record opened from a phone
Four  Anything opened right after login
Five  Anything that loads "everything" by default

⚠️ Number four is a common trap. The screen right after login often loads five or six things at once because that is the moment the system sets itself up. It feels slow at exactly the moment somebody is starting their shift and is least patient.

The screen that is only slow for one venue

If you run more than one room, check each separately. A screen that is fine at your smaller venue and slow at your busiest one is not "sometimes slow" — it is slow in proportion to how much business you do there.

⚠️ That is the worst shape, because it punishes your best venue hardest and looks like a coincidence.

The default-everything problem

Screens that load the full history by default are slow forever and get slower every month you operate.

⇒ A screen that defaults to "this week" and lets you widen it is faster on day one and still fast in year three. This is a settings question more often than a rebuild question.

Why it gets worse without anybody changing anything

Nothing has to break for a screen to slow down. You simply keep trading.

Year one   the table holds a few thousand bookings - anything is fast
Year three the same table holds a million - the same screen crawls

⚠️ The code did not change. The volume did. This is why "it used to be fine" is true and useless as a diagnosis.

Numbers worth keeping

One   Which screens staff open during service (ask, do not guess)
Two   Which screens staff have stopped opening
Three The slowest one in ten, per screen - not the average
Four  How many decisions per night are made from memory

⚠️ Four is the one nobody counts. Ask a floor manager to make a mark every time they answer a guest from memory rather than from a screen. One busy Friday is enough to see the shape.

⇒ On choosing which number to look at at all: the number you check first.

The five lines to pin up

────────────────────────
  Slow screens

  1. Name the screen, not "the system"
  2. Say what you were doing when it was slow
  3. Count out loud - "about four seconds"
  4. Say whether it is always or sometimes
  5. Say what you did instead while waiting

  Point 5 is the one we most need.
────────────────────────

Point five is the one owners never ask for and the one that explains everything. What somebody does while waiting becomes what they do instead of waiting.

Common objections

"It is only a few seconds"

⚠️ True, and irrelevant. The question is not how long the wait is. It is whether the wait is longer than the patience available at that moment. Three seconds at a desk is nothing. Three seconds with a guest in front of you is a guess.

"Nobody has complained"

⇒ That is the symptom, not the counter-argument. People complain about things that are broken. They quietly route around things that are merely slow.

"We will look at it when we redesign"

⚠️ Redesigns are measured in months. The single-cause fixes are measured in minutes and you will not know which kind you have until somebody looks.

"We already replaced the system once"

⇒ Replacing the system changes which screens are slow. It does not remove the pattern, because the pattern comes from how much data you have accumulated, and you carried that with you.

"Our staff would tell us"

⇒ Ask them question three above — what they do instead. If they have an answer ready, they have been routing around it for a while and did not think it was worth mentioning.

What to do this week

One   Ask three staff the three questions. Write the answers down.
Two   Pick the single screen that came up most.
Three Time it yourself, during service, on a phone.
Four  If it is over three seconds, say so to whoever maintains it -
      with the screen name, the filters, and what you did instead.

⇒ If you are not sure what "during service" reveals that a quiet afternoon does not, the slow Tuesday problem is the same shape applied to trading patterns.

Summary

• Slow screens are not complained about, they are abandoned
• The cost is decisions made from memory, not seconds lost
• Averages hide it. Look at the slow tail, on the screens used mid-shift
• Ask what staff do instead of opening a screen - that is the whole diagnosis
• Slow screens usually have one cause, findable in an afternoon
• New staff are hurt most, and it looks like a hiring problem

Nobody decides to stop measuring their venue. The screen just gets slow, and the decision quietly moves back into somebody's head.