# Three slides, not thirty

> How many slides should a digital sign have? Work it out from the loop, not from taste. Three slides at eight seconds each comes round in 24 seconds. Thirty takes four minutes, and most of it is never seen.

Source: https://channelos.tv/academy/three-slides-not-thirty
Published: 2026-10-29
Track: Make the board worth looking at
Lesson 2 of 4
Plan: Free plan
Author: The ChannelOS team

**Three to five slides is the right size for most screens, held about eight seconds each, which is a loop of 24 to 40 seconds. The constraint is dwell time, not design. As a rule of thumb worth testing on your own floor, treat someone walking past as four seconds and someone in a queue as forty: at thirty slides the loop runs four minutes, so most of the board is never seen by either of them.**

---

## The short version, in 5 steps

1. Count your slides, multiply by the seconds each one holds, and write that total down as your loop.
2. If the loop runs past about 45 seconds, decide which slides come out rather than which ones stay.
3. Say what each remaining slide is for in four words, and split or cut any slide that needs more than four.
4. Put the price on the same slide as the thing it prices, and keep the action you want on the last slide only.
5. Open the show on your phone, stand where your customer stands, and watch one full loop from there.

---

The show started with two slides. Then the opening hours went on, then the wifi password, then the loyalty card, then three photographs of the room, and it is eighteen slides long now. Nobody has mentioned the board since the week it went up. That is not a taste problem and no amount of better typography fixes it. It is arithmetic, and the number nobody wrote down is how long the whole thing takes to come round again.

**Three to five slides, held about eight seconds each, so the loop comes round in under 45 seconds.** Everything else here follows from that one figure.

## How many slides should a digital sign have?

There are two people looking at your screen, and only two. One is walking past, for about as long as it takes to cross a shop window. The other is standing still in a queue or a waiting room, because there is nothing else to look at.

**Call it four seconds for the first and forty for the second.** Both are a rule of thumb rather than a measurement, so check them against your own floor: time a real queue at your counter, and every number below moves with whatever you get.

Write the loop out for both of them. Eight seconds a slide is the ChannelOS default, so:

| Slides | Loop at 8 seconds each | What the person in the queue sees |
|---|---|---|
| 3 | 24 seconds | the whole board, most of it twice |
| 5 | 40 seconds | the whole board, once |
| 10 | 1 minute 20 | about half of it |
| 30 | 4 minutes | about a sixth, and never the same sixth |

The second reader is the one who settles it. At thirty slides, any one slide is on screen for one thirtieth of the time, so the odds that the slide you spent an afternoon on is up when somebody glances over are about one in thirty. It is not that slide twenty nine is a bad slide. It is that you paid to design something with a one in thirty chance of being the board.

And there is a harder consequence for the four second reader: **there is no first slide.** Whichever slide happens to be up is the entire board, mid-loop, with no run-up. A slide that only makes sense after the one before it does not make sense.

## How long should each slide stay up?

Budget about two seconds for someone to notice the board changed at all, then a little over a second for each line they have to read. Another rule of thumb rather than a measurement: it gets you a starting number to adjust once you have watched a loop from the floor.

A single headline is done in five or six seconds. A panel of five to seven menu items needs ten to twelve, which is worked out properly in [the arithmetic behind type size and items per panel](/blog/digital-menu-board-design). The editor's presets are 3, 5, 8, 10, 15 and 30 seconds, and the setting is **per slide**, not per show, so a dense board can hold while a title card goes past quickly.

If a slide genuinely needs thirty seconds, that is the signal, not the solution. Thirty seconds of reading is two slides or a leaflet.

The other half of the budget is the ceiling. Keep the whole loop under about 45 seconds and the person at the counter sees all of it. Past about ninety seconds you are authoring slides for nobody in particular.

## What should I put on each slide?

Four rules, and they all come out of the four second reader.

- **One job per slide.** Try saying what the slide is for in four words. "Today's special, with price." "Where to collect." If it takes a sentence, it is two slides, or none.
- **The price goes on the same slide as the thing it prices.** A slide of photographs followed by a slide of prices is a board nobody can use, because the two halves are never on screen together.
- **The action goes on one slide, last.** A code to scan, a counter to walk to, a website to type. On every slide it is wallpaper; on one slide it is an instruction. What makes a code worth putting up at all is in [QR codes people actually scan](/academy/qr-codes-people-actually-scan).
- **Anything that needs a paragraph is not a slide.** Terms, a history of the business, the full allergen list. Nobody reads a paragraph standing up. Print it and put it by the till.

## How do I cut a show that is already too long?

1. Count the slides and multiply by the seconds each one holds. Write the total down. That is your loop.
2. If it runs past about 45 seconds, decide what comes **out**. Deciding what stays is how you end up with eighteen slides, because every slide survives that question.
3. Run the four word test on each survivor.
4. Check that each one still works cold, with nothing before it.
5. Open the show on your phone, stand where your customer stands, and watch one full loop from there. This is the step people skip, and it is the only one that tells you the truth.

![Starting a new show in ChannelOS](/shots/new-show.webp)

## How often should I change it?

More often than the slide count would suggest, and this is the cheap part. What a screen shows is derived from the show in real time, so an edit made on a phone reaches every screen tuned to that show by itself. There is no publish step to remember and no device to go and sync, which means the cost of changing a board is the edit and nothing else.

We are not going to tell you what changing it does for your takings. We have no evidence for that, so any figure printed here would be invented. What we can tell you is the mechanism and the price of using it, which is one edit.

A workable plan: prices and anything dated get changed the day they change, and one slide of the three is the one you rewrite every week. If you cannot remember writing a slide, it has been up long enough to have become part of the wall.

## Why does the free plan stop at three slides?

The free plan holds two shows at three slides each, and on this particular question the cap and the advice agree. Three slides is a 24 second loop, which is the shape most single screens want anyway.

Where it does bite is a fourth slide you can justify, and that is a real reason to move up rather than something to work around. Worth knowing before you count on the free screen: a free screen carries a ChannelOS banner along the bottom, and about every two minutes that banner opens into a full ChannelOS slide for about 30 seconds and then folds back. That is what pays for it, and any paid plan removes both.

## When is thirty slides the right answer?

Plenty of the time, and the test is always the reader rather than the board.

**When the reader came specifically to read it.** A departures board, a class timetable, a staff rota, a spec sheet in a showroom read at arm's length. These people will wait, because they are looking for one row and they know it is coming. Dwell time is minutes, so the budget on this page does not apply.

**When each board is its own screen.** A menu wall is not one screen of fifteen slides. It is three screens of five, and nothing waits for anything.

**When you have a long wait rather than a queue.** A waiting room where people sit for twenty minutes is a genuinely different brief from a shop window, and a longer, slower loop is right there.

And there is an answer that is neither three slides nor thirty: more **time** instead of more slides. Split the day so breakfast runs its three and dinner runs three different ones, which is six slides authored and three in the loop at any moment. That needs scheduling, which starts on Solo, and it is covered in [breakfast, lunch and dinner on one screen](/academy/dayparting-a-day).

Go and count your own loop. Slides times seconds, written down, before you add anything else to it. If the number surprises you, the next thing to check is whether the text is big enough to read from where people actually stand, which is [how big the text has to be](/academy/how-big-the-text-has-to-be).

## Frequently asked questions

### How many slides should a digital sign have?

Three to five for a screen people walk past or queue in front of. The reason is the loop: at the default eight seconds a slide, three slides come round every 24 seconds and five every 40, so somebody waiting at a counter sees the whole board. Thirty slides is a four minute loop, which means a given slide is on screen for about one thirtieth of the time anybody is in the room.

### How long should each slide be?

Eight seconds is a sensible default and it is what ChannelOS uses for a new slide. As a rule of thumb to test rather than a measurement, budget roughly two seconds for a person to notice the board changed, plus a little over a second for each line they have to read. A single headline is fine at five or six seconds; a seven item menu panel wants ten to twelve. The presets in the editor are 3, 5, 8, 10, 15 and 30 seconds, set per slide rather than per show.

### What should I put on a digital sign?

One job per slide. A slide that sells one thing, with its price on the same slide, beats a slide carrying four things and a paragraph. Anything that needs explaining in sentences belongs on a leaflet by the till, because nobody reads a paragraph standing up. Keep whatever you want the person to do, a code to scan or a counter to walk to, on one slide at the end rather than on all of them.

### How often should I change digital signage content?

Change prices and anything dated the day it changes, and plan on one slide you rewrite weekly. The practical argument is cost: what a screen shows is derived from the show in real time, so an edit on your phone reaches the screen by itself with no publish step and no device to sync. If you cannot remember writing a slide, it has been up long enough to have become part of the wall.

### Is three slides enough for a menu?

For one counter, usually yes: a headline board, the five to seven items you actually want to sell, and a closing slide. A full menu is not three slides, and it is not thirty either. It is three or four separate boards, each on its own screen or each airing in its own part of the day, so nobody has to wait for the drinks list to come round again.


## Next lesson

[Set your colours once so the fifth show still matches](https://channelos.tv/academy/brand-kit-every-show-matches)


## Further reading

- https://channelos.tv/blog/digital-menu-board-design
- https://channelos.tv/blog/digital-signage-size-guide
- https://channelos.tv/blog/how-channelos-works


The whole curriculum: https://channelos.tv/academy and https://channelos.tv/academy.md