# A kitchen display per station cuts errors by 60%. But not in the first month

Moving from a shouted tab to a kitchen display has a learning curve. This article explains why errors fall later rather than on day one, and how to shorten the adjustment.

- **Published:** 21 February 2026
- **Reading time:** 4 minutes
- **Category:** Kitchen
- **Author:** The Oasys team

The kitchen display system, the KDS, works. In the kitchens we follow, order errors fall by close to 60% once the screen replaces the shouted tab and the paper on a spike. But I need to be honest about something a system salesperson will not say: the first month is bad. The drop in errors comes after the curve, not on the day you switch the screen on.

## Why errors fall, further down the line

The shouted tab has three points of failure that every cook knows. The order nobody heard in the rush. The paper that fell behind the stove. The handwriting nobody could read.

A KDS split by station kills all three. The order appears legibly on the screen of the right station, what belongs to the grill goes to the grill, what belongs to the fryer goes to the fryer. Nothing disappears, nothing depends on a shout beating the noise. And the kitchen starts to see the time on each order, so whatever is running late shows up before it becomes a complaint at the table.

## Why the first month is hard

Because your team does not work with screens. It works with muscle memory. Years of shouting orders out and spiking paper do not turn into something else in a week.

At the start, the cook forgets to mark an order as ready. Looks at the spot where the paper used to be. Finds the new rhythm strange. And there is always the veteran who swears it was faster by shouting. That is normal. It is an adjustment curve, not a defect in the system. Whoever gives up in the first month throws away the gain that would have arrived in the second.

## How to speed up the adjustment

You can shorten the curve. Here is what works in the kitchens we follow.

Start on a slow night, not on Saturday. Tuesday or Wednesday, quiet trade, is where the team learns without the weight of the rush.

Keep the paper as a safety net for a week. Screen and paper together at the start takes away the fear of what happens if the screen fails. After a week, take the paper away, otherwise nobody lets go of the old habit.

Pick one owner for the screen on each shift. One person responsible for marking orders ready and teaching the others. With no owner, everybody assumes it is somebody else's job.

Train the gesture, not the concept. A cook does not want to understand the system. A cook wants to know where to tap to mark an order ready. Show the gesture, repeat it, done.

## What comes after the curve

Once the first month is past, what was resistance turns into the opposite complaint: nobody wants to go back to paper. Errors fall, the time to send a dish out becomes visible, and the Saturday rush stops turning into chaos. The gain is not the screen. The gain is a predictable kitchen.

## Why we are writing about this

At Oasys, the KDS routes each item to the right station and shows the time on each order in real time. But we do not sell first-day magic. We sell the system and the truth about the curve, because a kitchen that understands that month one is adjustment reaches month two with 60% fewer errors. A kitchen that expected a miracle on day one switches the screen off on the first difficult Friday.

The screen works. It just does not work on its own, and it does not work instantly. It works once the team flips the switch, and that switch is one we can help flip.

## See also

- [All insights](https://oasysai.com.br/insights)
- [For restaurants](https://oasysai.com.br/para-voce/restaurantes)
- [For burger joints](https://oasysai.com.br/para-voce/hamburguerias)
- [For Japanese restaurants](https://oasysai.com.br/para-voce/restaurantes-japoneses)

---

This is the English markdown representation of https://oasysai.com.br/insights/kds-por-estacao. The rendered page
is in Brazilian Portuguese and is the authoritative version.
