If someone in your business writes down the cold room temperature in a notebook, weighs the goods arriving at the warehouse and logs them on a sheet, or tracks fuel consumption per route in a notebook, you are already paying the cost of measuring. What you are not collecting is the benefit. The number gets written down and dies there: no one adds it up, no one compares it with last week’s, and no one finds out in time when something goes out of range.
The useful question is not whether measuring helps. It is what changes in your operation when that data stops being an isolated note and becomes something you can review, compare, and use to decide.
What you already measure without calling it measurement
Almost no business starts measuring for fun. It starts because something is costing it money.
- Temperature: cold rooms, display cases, ovens, sterilizers, boilers, pool water.
- Weight: what the supplier delivers to you, the portion you serve, what the truck carries, the day’s shrinkage.
- Consumption: electricity, water, gas, fuel, supplies, machine hours.
Each of those numbers exists because there is a loss behind it: goods that spoil, portions that do not add up, bills that rise without a clear explanation, deliveries that yield less than expected.
The problem is not that you do not have data. It is that you have it in a format you cannot get anything useful from in time.
What you gain, in concrete terms
You stop finding out too late
A handwritten log is almost always written once a day, at the end of the shift and from memory. If the cold room went out of range at 2 a.m. and came back at 5, the notebook will show a 4 and nothing will happen. The thermometer worked. The problem is that the data arrived when it was no longer useful for preventing the loss.
When the log has a time and is reviewed regularly, you see the pattern before the damage: the temperature that starts rising every night, the weight that drops two kilos per sack, the consumption that spikes on Mondays.
You end internal arguments with a table
“The supplier always sends less,” “that route always uses more,” “the portion is fine.” Arguments like these persist because no one has the number at hand. With comparable records, you show the table and the conversation changes subject.
You know what production really costs you
If you divide consumption by units produced or services provided, you get your cost per dish, per delivery, per appointment, per occupied room. That is what lets you set prices without guessing and identify which of your products leaves you a margin and which does not.
You separate whether the problem is the machine or the process
A compressor working overtime, an air conditioner badly calibrated, an oven taking twice as long. Temperature and consumption tell you whether the equipment is failing or whether the problem is how it is used. Replacing equipment is expensive. Changing a schedule is free.
You can test changes and see if they worked
You change suppliers, adjust a route, move the start time. Without prior history, you do not know if it improved or if you just had a good week. With four weeks of comparable records, you do know.
What the notebook really costs you
- It cannot be added up quickly. A month of notes takes an entire afternoon to transfer to a spreadsheet.
- It gets lost. It gets wet, misplaced, or someone takes it to another location.
- It is written from memory. At the end of the shift, in a hurry, without the correct unit.
- No one reviews it until there is a big problem. That is the worst time to find out.
- When the person leaves, the judgment leaves too. The numbers are still there, but no one knows what was normal.
- There is no way to compare. Everyone writes things down however they want and in whatever order they want.
Writing it down is not the same as measuring
| With a notebook | With usable measurement |
|---|---|
| It gets written down when someone remembers | Fixed interval, with time |
| Free text (“normal,” “fine”) | Number and unit |
| Comparing takes hours | Comparing takes seconds |
| You find out when the product is already spoiled | You find out beforehand |
| The history is lost | The history remains and can be reviewed |
| Only one person understands the notebook | Anyone can read the series |
The difference is not paper versus screen. It is that the data is comparable, complete, and that someone looks at it.
What to do this week, without buying anything
- Choose a single variable. The one costing you the most today. Do not start with three.
- Define unit, frequency, and owner. For example: temperature in degrees, every four hours, the shift manager.
- Move the notebook to a shared spreadsheet. Fixed columns: date and time, value, unit, who logged it, note. Nothing else.
- Set a review rule. Ten minutes every Friday looking at the full series.
- Define the threshold. The number at which it starts to matter to you. If you do not define it beforehand, everything will seem normal.
- Wait three or four weeks before deciding whether it is worth automating. With that history, you already have what you need to justify it or rule it out.
When the spreadsheet is no longer enough
The spreadsheet works well until one of these cases appears:
- You need more than one reading per hour.
- You want to know what happens when you are not at the location.
- The team forgets to log it on the busiest days.
- You have several locations and want to see them together.
- You need someone to receive an alert right away, not on Monday.
That is where sensors, connected scales, or consumption meters come in, along with automation that logs and alerts. Each case has its own scope and cost, and no two are alike, so it is worth reviewing before committing to dates. If you want an idea of how that looks in delivery operations, see what to automate in logistics for faster deliveries. If your case is cold chain or inventory, the starting point is the same: first the threshold, then the sensor.
Recurring mistakes
- Measuring everything in the first month. You end up with data no one reviews and the feeling that it did not help.
- Buying the sensor before knowing what threshold matters to you. The device does not decide for you.
- Not defining who reviews it. Data without an owner is dead data.
- Saving numbers without a unit or time. Later they cannot be compared.
- Measuring without a reference value. If you do not know your normal range, you cannot detect what is abnormal.
- Turning measurement into a punishment. If the data is used only to find someone to blame, the team learns to write down what you want to see.
How to know if it is paying off
You do not need a study. You need two numbers of your own.
First, how much the last loss you could have prevented cost you: damaged goods, a remade order, the bill that rose without explanation. Second, how many hours per month your team spends logging, transferring, and reviewing that data.
If measurement helps you avoid even one of those losses per month, it has already paid for itself. And if it avoids none in three months, you traded a notebook for a spreadsheet and did not lose much.
What does not hold up is continuing to write things down by hand and never using what you wrote. There you are indeed paying the full cost and getting none of the benefit.