How to record a KRI reading on Mac
You can do this with RiskOS, a risk register for macOS. A value, a date, a note — and an indicator that starts telling you something.
A key risk indicator earns its place only once it has numbers in it. The thresholds, the direction, the link to a risk say nothing until a reading arrives and there is something to compare. Recording one takes under a minute, and doing it on the same day each period is the difference between a list of metrics and an early warning.
An indicator with no readings reads as no data, which is deliberately not the same as being in tolerance. Until a value arrives, RiskOS says it does not know rather than implying everything is fine.
Where readings live
Choose Indicators under Signals in the sidebar. The table opens worst first, so anything breached or in warning is already at the top: Ref, the indicator with the direction it is read in, the latest value, its status, the trend, the thresholds, the risks it watches and when the next reading is due.
Click a row and the panel on the right fills with that indicator — its thresholds, its reading history chart and the risks it is linked to. Readings are added from here, and the chart redraws the moment one lands. The list is searchable, so you can type part of a name instead of scrolling.
Record a reading, step by step
-
Open the indicators list
Choose Indicators under Signals in the sidebar. The list is ordered worst first and shows the latest value and status of every indicator, so you can see which ones are already asking for attention before you record anything.
-
Select the indicator you are updating
Click its row, or search for part of its name to narrow the list first. The panel on the right opens on that indicator, showing its thresholds, its target if it has one, the risks it is linked to and the readings recorded so far.
-
Check which direction is bad before you type
The direction sits beside the indicator's name and decides how every threshold is read. For Critical patches outstanding beyond SLA a higher number is worse; for Backup restore test success rate a lower number is worse. A value that breaches under one setting is in tolerance under the other, so confirm it before the first reading rather than after twenty.
-
Enter the value
Record a reading and type the number in the unit the indicator is measured in — 14 for a count of outstanding patches, 72 for a percentage, 3 for days to detect. Record the figure you measured rather than a rounded version of it; the chart is only worth reading if the points on it are real.
-
Date the reading to the period it measures
Set the date to the period the number describes, not the day you are typing. A restore rehearsal run on the last Friday of the month is a reading for that Friday, even if it reaches the register the following Tuesday. Readings are plotted in date order, so a misdated point bends the trend line for everyone who reads it later.
-
Add a note that explains the number
Write where the figure came from and anything that makes it unusual: a frozen change window, a supplier outage, a backlog cleared in an afternoon. A spike with a note is evidence. A spike without one gets argued about six months later by people who were not there.
-
Save the reading and read the status it produces
Once the reading is in, the latest value updates in the list and the status is worked out against the warning and breach thresholds in the indicator's own direction. Thresholds are inclusive: a value landing exactly on the warning threshold is a warning, and one exactly on the breach threshold is a breach.
-
Check the chart and the risks it feeds
The reading history chart draws the warning and breach lines across the series, so a number drifting towards one is visible several periods before it crosses. Then open a risk the indicator is linked to: it appears there with its live status, which is the reason for recording the reading at all.
What the status is telling you
The status is one word derived from the most recent reading, the two thresholds and the direction. It follows the indicator everywhere it appears, and it is what a colleague sees on opening the list without reading a single number.
| Status | What produced it | What it calls for |
|---|---|---|
| No data | No reading has been recorded yet. | Record one, or pause the indicator until it is genuinely being measured. |
| In tolerance | The latest reading sits on the safe side of both thresholds. | Nothing. Keep the cadence and watch the trend rather than the single value. |
| Warning | The latest reading has reached the warning threshold. | Look at the linked risk before the next reading, not after it. |
| Breached | The latest reading has reached the breach threshold. | Treat it as a finding: check the controls behind the risk and raise an action. |
| Overdue | The cadence has passed without a reading. | Record the measurement, or change the cadence to one you can keep. |
Thresholds are inclusive and direction-aware
Inclusive means the boundary belongs to the worse side: set a breach at 80 per cent where lower is worse, and 80 is a breach, not the last acceptable value. Direction-aware means the same pair of thresholds is read upwards for a count you want low and downwards for a rate you want high, which is why the direction sits beside the name in the list.
No data is not the same as in tolerance
An indicator nobody has measured is a gap in the register, not a clean result. Keeping the two apart means an empty indicator is never mistaken for a healthy one in a report, and it makes the first reading worth recording the same week you create the indicator.
Dates, cadence and pausing
An indicator is a number measured on a rhythm, and the rhythm matters as much as the thresholds. A value from March tells you little in September, and a series with holes cannot show a trend.
What overdue actually means
Overdue is about the cadence, not the value. An indicator can be in tolerance and still be marked overdue, because the period it should have been measured in passed without a reading. Two separate questions: is the number bad, and do we still know what the number is.
Set a cadence you can keep
Lists fill with overdue indicators when the cadence reflects how often a team wished it measured something. Weekly indicators need a weekly owner. If a figure is produced once a quarter, set the cadence to match, and the overdue mark stays meaningful when it appears.
Pause an indicator instead of deleting it
When a measurement stops for a while — a system is being replaced, a supplier relationship is on hold — pause the indicator from the list. It stops asking for readings and stops counting as overdue, while everything already recorded stays where it is. Resume it from the same place and the chart carries on.
Reading the history chart
A single reading answers "where are we". The series answers "where is this going", which is the only version of the question a risk conversation can act on early. The warning and breach lines are drawn across the chart, so the distance between your latest point and the line you care about is visible rather than calculated.
Here is a restore-rehearsal indicator read downwards — lower is worse — with a warning at 90 per cent and a breach at 80 per cent.
| Reading date | Value | Status | What the series was saying |
|---|---|---|---|
| 31 March | 94 % | In tolerance | Healthy, and the first point of a series. |
| 30 April | 91 % | In tolerance | Still fine, but the first move in the wrong direction. |
| 31 May | 88 % | Warning | Past the warning line. Worth a look at the linked risk now. |
| 30 June | 83 % | Warning | Three consecutive falls. The trend, not the value, is the finding. |
| 31 July | 72 % | Breached | Past the breach line, and predictable four months earlier. |
Nothing there was a surprise by July. The fifth row is what people react to; the third row is where the work would have been cheap. Read the direction of travel across three or four points before you read the last one.
How a reading reaches the rest of the register
A reading is intelligence, and its job is to reach the risks it was set up to watch. Open a risk in Risks and its Intelligence section lists the linked indicators with their live status, beside any events recorded against it. Anyone reading that risk sees the measurement has gone against it.
A reading does not re-score a risk by itself
This is deliberate. Residual scores come from the effectiveness of the controls you link, not from a metric, so a breached indicator never quietly rewrites a rating. What it gives you is a reason to re-score, with a date and a number attached. Open the risk, ask whether the controls are still doing what you assumed, and change the rating yourself if they are not.
Carry indicators into a report
Reports in RiskOS include a risk indicators section you can switch on, so the readings arrive in the board pack alongside the register rather than as an attachment nobody opens. Every output is built from the same snapshot, so the status in the document matches the status on screen when you exported it.
Troubleshooting
I recorded a reading and the status did not change
Check the direction first. An indicator where lower is worse reads its thresholds downwards, so a number you expected to trigger a warning may be on the safe side of the line. Confirm the direction beside its name, then the two threshold values.
The indicator still says no data
The reading has almost certainly landed on a different indicator with a similar name. Search the list and read the Latest column across the results; whichever row shows your value is the one that received it.
An indicator is marked overdue although I measure it every month
Two things usually cause it. The cadence may be shorter than the rhythm you measure on, in which case set it to match. Or readings are dated the day they were typed rather than the period they describe, which leaves gaps.
I typed the wrong value
Press ⌘Z to undo, and ⇧⌘Z if you undo a step too many. If it comes to light later, record the corrected figure for the same period and use the note to say what was wrong.
A breached indicator is not showing on the risk I expected
The indicator is not linked to that risk yet. Link it to the risk it watches, and the live status appears in that risk's Intelligence section from that moment. One indicator can watch several risks, and a risk can carry several indicators, so a measurement you rely on in two places only needs recording once.
A routine that keeps indicators honest
Indicators decay faster than anything else in a register, because they need something from you on a schedule. A short routine holds them together.
- Record on a fixed day. The first working day of the month, or every Monday. A fixed day survives a busy week better than an intention to do it when the number is ready.
- Do the whole list in one sitting. It is ordered worst first, so working top to bottom updates the indicators that matter even if you run out of time.
- Write the note while you remember. It is the only part of a reading that cannot be reconstructed afterwards.
- Read the series, not the point. Three readings moving the wrong way inside tolerance deserve more attention than one bad reading in a flat line.
- Chase warnings, not breaches. By the time an indicator breaches, the cheap options have gone. The warning threshold exists to be acted on.
- Pause what you are not measuring. An honest paused indicator is better than an overdue one everybody has learned to ignore.
- Take breaches back to the risk. Check whether the linked risk's controls still deserve the effectiveness you gave them, and re-score if they do not.
- Back up after the monthly pass. Choose File ▸ Back Up RiskOS… and RiskOS writes readings, risks, controls and settings to a single file you keep, wherever you choose to save it.
Frequently asked questions
How do I record a KRI reading?
Choose Indicators under Signals in the sidebar, select the indicator, then add a reading with a value, the date of the period it measures and a short note. The latest value, the status and the history chart update as soon as it is saved, and every risk the indicator is linked to shows the new status.
How often should I record a reading?
On the cadence set for that indicator, and no less. Monthly suits most operational measures; weekly suits a fast-moving count such as outstanding critical patches. RiskOS marks an indicator overdue when its cadence passes without a reading, so choose an interval you can genuinely keep rather than the one that sounds diligent.
What does no data mean on a risk indicator?
It means no reading has been recorded, and it is kept separate from being in tolerance on purpose. An unmeasured indicator is a gap in your monitoring rather than a good result, so RiskOS says it does not know instead of showing a reassuring status that nothing behind it supports.
Does recording a reading change a risk score?
No. Residual scores come from the effectiveness of the controls linked to a risk, never from a metric, so a breached indicator cannot rewrite a rating on its own. It appears on the risk with its live status, which is your prompt to look again at the controls and re-score if the evidence has moved.
Can I record a reading for a date in the past?
Yes, and you often should. Set the date to the period the number describes rather than the day you are typing it in. A test run at the end of last month is a reading for last month, and dating it correctly keeps the chart, the trend and the cadence all telling the same story.
What happens if a reading lands exactly on a threshold?
It counts as the worse state. Thresholds in RiskOS are inclusive, so a value sitting exactly on the warning threshold is a warning and one exactly on the breach threshold is a breach. They are also direction-aware: the same numbers are read upwards for a count you want low and downwards for a rate you want high.
How do I stop an indicator being marked overdue while nobody is measuring it?
Pause it from the indicators list. A paused indicator stops asking for readings and stops counting as overdue, and everything already recorded stays intact. Resume it from the same place when the measurement restarts, and the chart carries on from the series you built rather than starting again.
Do my indicator readings leave my Mac?
No. There is no account, no sign-in and no sync, nothing is uploaded, and there is no tracking or analytics. Readings stay in your register on your Mac and leave it only when you export or back it up yourself. The only network use is Apple's App Store, for purchases, and it never sees your data.