TV Dashboards: Put Your Metrics on the Office Wall
TL;DR
Most office TV dashboards are a spare laptop, permanently logged into somebody's account, with the screensaver disabled and a sticky note saying "don't touch." That setup is a standing security problem and it breaks whenever the session expires. Pairing solves it properly: the television opens one URL, you approve it from your phone, and the screen holds a narrow display token that can read the dashboards you assigned it and nothing else.
The usual way, and why it's bad
You want revenue on the wall. So you find a machine, plug it into the TV, open a browser, log in, full-screen the tab, and disable sleep.
What you've actually built:
A permanent session on a shared screen in a room people walk through.
Tied to one person's account. When they leave, the wall goes dark — or worse, it doesn't, and an ex-employee's session is still rendering live numbers.
Access far wider than the wall needs. That session can reach every dashboard, every setting, every datasource. The TV only ever needed to show one board.
Fragile. The session expires, the OS updates overnight, the screensaver wins. Someone has to go find a keyboard.
The failure isn't the laptop. It's that a display was given a user identity when what it needed was a display identity.
What pairing actually does
Open dashboardbase.com/tv on the television's own browser. You get an edge-to-edge pairing screen — no app chrome — showing a large short code and a QR, and it keeps the screen awake while it waits.
Then, from your phone: scan the QR with your camera (it deep-links straight into the app) or use the in-app scanner, or just type the code. The QR route is the smoothest, and it lands in the native app — iOS, Android — so it's worth having that installed before you walk over to the television. Name the screen something you'll recognise later — "Reception TV" — pick the dashboard it should play and how often it refreshes, and confirm.
The moment you confirm, the television switches into a permanent fullscreen kiosk view. No keyboard was involved, and the TV never signed in to anything.
Any workspace member can pair a screen. The approval happens from an account that's already logged in, which is what makes a stray code harmless — a code on its own can't claim a dashboard.
The security property that matters
A paired screen holds its own display token. Not your login, not a session, not an API key.
That token is narrow and revocable: it can read the dashboards assigned to that screen and nothing else. Unpair the screen from the Devices list and it drops straight back to the pairing view — no password changes, no session invalidation, no wondering what else that machine could still reach.
This is the same instinct behind keeping your data behind your own API instead of handing over database credentials: give each thing the narrowest capability that does its job.
One screen, several dashboards
A screen isn't limited to one board. Give it a rotation and it cycles on its own — revenue, then product, then uptime — holding each one for as long as you choose before moving on.
The display settings page is where you shape it: rename the screen, set how long each dashboard stays up (anywhere from ten seconds to an hour), switch between normal and larger TV-friendly text, and build the rotation by adding dashboards from a picker, reordering them, or dropping one. The order you see is the order it plays, and the page tells you how long a full rotation takes — so "does someone walking past at 9:04 see revenue?" is a question you can actually answer.
Two constraints worth knowing up front: only published dashboards can go on a screen, so drafts never surface in the picker, and a dashboard already in the rotation can't be added twice.
Changes land on the wall immediately rather than waiting for the next refresh cycle. Save on your phone, look up, it's already changed.
What it does when things go wrong
Wall displays fail in boring, specific ways, and most of them are network-shaped.
Network blips: the last-good data stays on screen. A wall showing slightly stale numbers is useful; a wall showing a spinner or an error page is worse than a blank one, because people stop trusting it.
Reconnection: it recovers quietly on its own. Live events stream back in.
A dropped connection during setup: the screen doesn't depend on catching one live "you've been paired" signal. It keeps checking, so it goes live even if its connection wobbled while you were still on your phone.
A dashboard in the rotation gets deleted: the TV carries on with the rest. If every dashboard on the screen is gone, it shows a calm "nothing to show yet" prompt and waits to be reconfigured, rather than throwing an error at a room full of people.
None of that is glamorous. All of it is the difference between a screen people glance at and a screen people ignore.
Making a dashboard readable across the room
A wall display is not a laptop dashboard shown bigger. The reader is eight metres away, walking, and looking for perhaps two numbers.
What works:
KPI widgets, few and large. One number per tile. A KPI can carry a goal progress bar underneath with a percentage readout, which reads well at distance.
Status widgets in heartbeat mode, which render a live ECG-style pulse instead of a flat indicator — motion is legible from much further away than text.
Progress List for a handful of labelled bars, where the bars slide and the percentages count to each new value.
A Clock, and a Countdown if something is shipping — "days since last incident" is the classic, and it reads instantly.
Contributions Grid for activity over time, which scans as a shape rather than as digits.
The larger TV-friendly text setting, which exists precisely for this.
What doesn't: dense tables, six-series line charts, anything with a legend you'd need to squint at. If a widget requires reading a label to interpret, it belongs on a laptop.
On colour, one thing catches people out: colours are opt-in. A response that sends no color field renders in a neutral grayscale, so a dashboard stays monochrome until you deliberately reach for an accent. On a wall that's a feature — reserve colour for the one thing that means "look here," and it'll actually pull attention. Status colours (alert levels, trend badges) always keep their meaning and are never overwritten.
Where this isn't the right fit
Being straight about it:
You need someone to interact with the screen. Kiosk mode is a display, not a touch app. If people should filter, drill in, or click through, put a laptop there instead.
The room needs a dashboard nobody may see. A paired screen shows what it's assigned. It's a wall display, with the discretion that implies — think before you rotate payroll onto it.
You want one screen per person. This is shared-context infrastructure. Individual dashboards belong on phones, where push notifications reach you when a number moves.
You don't have a dashboard worth watching yet. A TV amplifies whatever you point it at. If the underlying board is four charts nobody agreed on, the wall makes that more visible, not less.
Getting one up
Publish the dashboard. Drafts can't go on a screen.
Open
dashboardbase.com/tvon the television's browser. Note the code.Scan the QR from your phone, or type the code in the app.
Name the screen and pick the dashboard. Set the refresh interval.
Confirm. The TV goes fullscreen on its own.
Add more dashboards to its rotation later from the display's settings page, if one board isn't enough.
The Share dialog on any dashboard also has a TV tab with the three-step version of this, a one-tap jump into pairing with that dashboard already selected, and a copyable dashboardbase.com/tv address — handy when the person standing at the television isn't the person holding the phone.
Every paired screen then shows up under Devices on the workspace page, with its name, what it's currently rotating through, and when it was last seen. That "last seen" column is the one you'll actually use — it's how you find out a screen has been quietly unplugged for a week.
The honest bottom line
The hard part of an office dashboard was never rendering it. It was that "put this on the wall" had no answer other than "log a computer in and leave it there," and every team quietly accepted a permanent session on a shared screen because the alternative was more work.
Pairing makes the display its own thing, with its own narrow, revocable identity — which is the shape the problem always had. If you already have a dashboard you like, it's about ten seconds of work to find out whether it's a good wall.
TV Dashboards: Put Your Metrics on the Office Wall
TL;DR
Most office TV dashboards are a spare laptop, permanently logged into somebody's account, with the screensaver disabled and a sticky note saying "don't touch." That setup is a standing security problem and it breaks whenever the session expires. Pairing solves it properly: the television opens one URL, you approve it from your phone, and the screen holds a narrow display token that can read the dashboards you assigned it and nothing else.
The usual way, and why it's bad
You want revenue on the wall. So you find a machine, plug it into the TV, open a browser, log in, full-screen the tab, and disable sleep.
What you've actually built:
A permanent session on a shared screen in a room people walk through.
Tied to one person's account. When they leave, the wall goes dark — or worse, it doesn't, and an ex-employee's session is still rendering live numbers.
Access far wider than the wall needs. That session can reach every dashboard, every setting, every datasource. The TV only ever needed to show one board.
Fragile. The session expires, the OS updates overnight, the screensaver wins. Someone has to go find a keyboard.
The failure isn't the laptop. It's that a display was given a user identity when what it needed was a display identity.
What pairing actually does
Open dashboardbase.com/tv on the television's own browser. You get an edge-to-edge pairing screen — no app chrome — showing a large short code and a QR, and it keeps the screen awake while it waits.
Then, from your phone: scan the QR with your camera (it deep-links straight into the app) or use the in-app scanner, or just type the code. The QR route is the smoothest, and it lands in the native app — iOS, Android — so it's worth having that installed before you walk over to the television. Name the screen something you'll recognise later — "Reception TV" — pick the dashboard it should play and how often it refreshes, and confirm.
The moment you confirm, the television switches into a permanent fullscreen kiosk view. No keyboard was involved, and the TV never signed in to anything.
Any workspace member can pair a screen. The approval happens from an account that's already logged in, which is what makes a stray code harmless — a code on its own can't claim a dashboard.
The security property that matters
A paired screen holds its own display token. Not your login, not a session, not an API key.
That token is narrow and revocable: it can read the dashboards assigned to that screen and nothing else. Unpair the screen from the Devices list and it drops straight back to the pairing view — no password changes, no session invalidation, no wondering what else that machine could still reach.
This is the same instinct behind keeping your data behind your own API instead of handing over database credentials: give each thing the narrowest capability that does its job.
One screen, several dashboards
A screen isn't limited to one board. Give it a rotation and it cycles on its own — revenue, then product, then uptime — holding each one for as long as you choose before moving on.
The display settings page is where you shape it: rename the screen, set how long each dashboard stays up (anywhere from ten seconds to an hour), switch between normal and larger TV-friendly text, and build the rotation by adding dashboards from a picker, reordering them, or dropping one. The order you see is the order it plays, and the page tells you how long a full rotation takes — so "does someone walking past at 9:04 see revenue?" is a question you can actually answer.
Two constraints worth knowing up front: only published dashboards can go on a screen, so drafts never surface in the picker, and a dashboard already in the rotation can't be added twice.
Changes land on the wall immediately rather than waiting for the next refresh cycle. Save on your phone, look up, it's already changed.
What it does when things go wrong
Wall displays fail in boring, specific ways, and most of them are network-shaped.
Network blips: the last-good data stays on screen. A wall showing slightly stale numbers is useful; a wall showing a spinner or an error page is worse than a blank one, because people stop trusting it.
Reconnection: it recovers quietly on its own. Live events stream back in.
A dropped connection during setup: the screen doesn't depend on catching one live "you've been paired" signal. It keeps checking, so it goes live even if its connection wobbled while you were still on your phone.
A dashboard in the rotation gets deleted: the TV carries on with the rest. If every dashboard on the screen is gone, it shows a calm "nothing to show yet" prompt and waits to be reconfigured, rather than throwing an error at a room full of people.
None of that is glamorous. All of it is the difference between a screen people glance at and a screen people ignore.
Making a dashboard readable across the room
A wall display is not a laptop dashboard shown bigger. The reader is eight metres away, walking, and looking for perhaps two numbers.
What works:
KPI widgets, few and large. One number per tile. A KPI can carry a goal progress bar underneath with a percentage readout, which reads well at distance.
Status widgets in heartbeat mode, which render a live ECG-style pulse instead of a flat indicator — motion is legible from much further away than text.
Progress List for a handful of labelled bars, where the bars slide and the percentages count to each new value.
A Clock, and a Countdown if something is shipping — "days since last incident" is the classic, and it reads instantly.
Contributions Grid for activity over time, which scans as a shape rather than as digits.
The larger TV-friendly text setting, which exists precisely for this.
What doesn't: dense tables, six-series line charts, anything with a legend you'd need to squint at. If a widget requires reading a label to interpret, it belongs on a laptop.
On colour, one thing catches people out: colours are opt-in. A response that sends no color field renders in a neutral grayscale, so a dashboard stays monochrome until you deliberately reach for an accent. On a wall that's a feature — reserve colour for the one thing that means "look here," and it'll actually pull attention. Status colours (alert levels, trend badges) always keep their meaning and are never overwritten.
Where this isn't the right fit
Being straight about it:
You need someone to interact with the screen. Kiosk mode is a display, not a touch app. If people should filter, drill in, or click through, put a laptop there instead.
The room needs a dashboard nobody may see. A paired screen shows what it's assigned. It's a wall display, with the discretion that implies — think before you rotate payroll onto it.
You want one screen per person. This is shared-context infrastructure. Individual dashboards belong on phones, where push notifications reach you when a number moves.
You don't have a dashboard worth watching yet. A TV amplifies whatever you point it at. If the underlying board is four charts nobody agreed on, the wall makes that more visible, not less.
Getting one up
Publish the dashboard. Drafts can't go on a screen.
Open
dashboardbase.com/tvon the television's browser. Note the code.Scan the QR from your phone, or type the code in the app.
Name the screen and pick the dashboard. Set the refresh interval.
Confirm. The TV goes fullscreen on its own.
Add more dashboards to its rotation later from the display's settings page, if one board isn't enough.
The Share dialog on any dashboard also has a TV tab with the three-step version of this, a one-tap jump into pairing with that dashboard already selected, and a copyable dashboardbase.com/tv address — handy when the person standing at the television isn't the person holding the phone.
Every paired screen then shows up under Devices on the workspace page, with its name, what it's currently rotating through, and when it was last seen. That "last seen" column is the one you'll actually use — it's how you find out a screen has been quietly unplugged for a week.
The honest bottom line
The hard part of an office dashboard was never rendering it. It was that "put this on the wall" had no answer other than "log a computer in and leave it there," and every team quietly accepted a permanent session on a shared screen because the alternative was more work.
Pairing makes the display its own thing, with its own narrow, revocable identity — which is the shape the problem always had. If you already have a dashboard you like, it's about ten seconds of work to find out whether it's a good wall.
TV Dashboards: Put Your Metrics on the Office Wall
TL;DR
Most office TV dashboards are a spare laptop, permanently logged into somebody's account, with the screensaver disabled and a sticky note saying "don't touch." That setup is a standing security problem and it breaks whenever the session expires. Pairing solves it properly: the television opens one URL, you approve it from your phone, and the screen holds a narrow display token that can read the dashboards you assigned it and nothing else.
The usual way, and why it's bad
You want revenue on the wall. So you find a machine, plug it into the TV, open a browser, log in, full-screen the tab, and disable sleep.
What you've actually built:
A permanent session on a shared screen in a room people walk through.
Tied to one person's account. When they leave, the wall goes dark — or worse, it doesn't, and an ex-employee's session is still rendering live numbers.
Access far wider than the wall needs. That session can reach every dashboard, every setting, every datasource. The TV only ever needed to show one board.
Fragile. The session expires, the OS updates overnight, the screensaver wins. Someone has to go find a keyboard.
The failure isn't the laptop. It's that a display was given a user identity when what it needed was a display identity.
What pairing actually does
Open dashboardbase.com/tv on the television's own browser. You get an edge-to-edge pairing screen — no app chrome — showing a large short code and a QR, and it keeps the screen awake while it waits.
Then, from your phone: scan the QR with your camera (it deep-links straight into the app) or use the in-app scanner, or just type the code. The QR route is the smoothest, and it lands in the native app — iOS, Android — so it's worth having that installed before you walk over to the television. Name the screen something you'll recognise later — "Reception TV" — pick the dashboard it should play and how often it refreshes, and confirm.
The moment you confirm, the television switches into a permanent fullscreen kiosk view. No keyboard was involved, and the TV never signed in to anything.
Any workspace member can pair a screen. The approval happens from an account that's already logged in, which is what makes a stray code harmless — a code on its own can't claim a dashboard.
The security property that matters
A paired screen holds its own display token. Not your login, not a session, not an API key.
That token is narrow and revocable: it can read the dashboards assigned to that screen and nothing else. Unpair the screen from the Devices list and it drops straight back to the pairing view — no password changes, no session invalidation, no wondering what else that machine could still reach.
This is the same instinct behind keeping your data behind your own API instead of handing over database credentials: give each thing the narrowest capability that does its job.
One screen, several dashboards
A screen isn't limited to one board. Give it a rotation and it cycles on its own — revenue, then product, then uptime — holding each one for as long as you choose before moving on.
The display settings page is where you shape it: rename the screen, set how long each dashboard stays up (anywhere from ten seconds to an hour), switch between normal and larger TV-friendly text, and build the rotation by adding dashboards from a picker, reordering them, or dropping one. The order you see is the order it plays, and the page tells you how long a full rotation takes — so "does someone walking past at 9:04 see revenue?" is a question you can actually answer.
Two constraints worth knowing up front: only published dashboards can go on a screen, so drafts never surface in the picker, and a dashboard already in the rotation can't be added twice.
Changes land on the wall immediately rather than waiting for the next refresh cycle. Save on your phone, look up, it's already changed.
What it does when things go wrong
Wall displays fail in boring, specific ways, and most of them are network-shaped.
Network blips: the last-good data stays on screen. A wall showing slightly stale numbers is useful; a wall showing a spinner or an error page is worse than a blank one, because people stop trusting it.
Reconnection: it recovers quietly on its own. Live events stream back in.
A dropped connection during setup: the screen doesn't depend on catching one live "you've been paired" signal. It keeps checking, so it goes live even if its connection wobbled while you were still on your phone.
A dashboard in the rotation gets deleted: the TV carries on with the rest. If every dashboard on the screen is gone, it shows a calm "nothing to show yet" prompt and waits to be reconfigured, rather than throwing an error at a room full of people.
None of that is glamorous. All of it is the difference between a screen people glance at and a screen people ignore.
Making a dashboard readable across the room
A wall display is not a laptop dashboard shown bigger. The reader is eight metres away, walking, and looking for perhaps two numbers.
What works:
KPI widgets, few and large. One number per tile. A KPI can carry a goal progress bar underneath with a percentage readout, which reads well at distance.
Status widgets in heartbeat mode, which render a live ECG-style pulse instead of a flat indicator — motion is legible from much further away than text.
Progress List for a handful of labelled bars, where the bars slide and the percentages count to each new value.
A Clock, and a Countdown if something is shipping — "days since last incident" is the classic, and it reads instantly.
Contributions Grid for activity over time, which scans as a shape rather than as digits.
The larger TV-friendly text setting, which exists precisely for this.
What doesn't: dense tables, six-series line charts, anything with a legend you'd need to squint at. If a widget requires reading a label to interpret, it belongs on a laptop.
On colour, one thing catches people out: colours are opt-in. A response that sends no color field renders in a neutral grayscale, so a dashboard stays monochrome until you deliberately reach for an accent. On a wall that's a feature — reserve colour for the one thing that means "look here," and it'll actually pull attention. Status colours (alert levels, trend badges) always keep their meaning and are never overwritten.
Where this isn't the right fit
Being straight about it:
You need someone to interact with the screen. Kiosk mode is a display, not a touch app. If people should filter, drill in, or click through, put a laptop there instead.
The room needs a dashboard nobody may see. A paired screen shows what it's assigned. It's a wall display, with the discretion that implies — think before you rotate payroll onto it.
You want one screen per person. This is shared-context infrastructure. Individual dashboards belong on phones, where push notifications reach you when a number moves.
You don't have a dashboard worth watching yet. A TV amplifies whatever you point it at. If the underlying board is four charts nobody agreed on, the wall makes that more visible, not less.
Getting one up
Publish the dashboard. Drafts can't go on a screen.
Open
dashboardbase.com/tvon the television's browser. Note the code.Scan the QR from your phone, or type the code in the app.
Name the screen and pick the dashboard. Set the refresh interval.
Confirm. The TV goes fullscreen on its own.
Add more dashboards to its rotation later from the display's settings page, if one board isn't enough.
The Share dialog on any dashboard also has a TV tab with the three-step version of this, a one-tap jump into pairing with that dashboard already selected, and a copyable dashboardbase.com/tv address — handy when the person standing at the television isn't the person holding the phone.
Every paired screen then shows up under Devices on the workspace page, with its name, what it's currently rotating through, and when it was last seen. That "last seen" column is the one you'll actually use — it's how you find out a screen has been quietly unplugged for a week.
The honest bottom line
The hard part of an office dashboard was never rendering it. It was that "put this on the wall" had no answer other than "log a computer in and leave it there," and every team quietly accepted a permanent session on a shared screen because the alternative was more work.
Pairing makes the display its own thing, with its own narrow, revocable identity — which is the shape the problem always had. If you already have a dashboard you like, it's about ten seconds of work to find out whether it's a good wall.