The Recovery Scorecard: Measuring How Fast You Actually Come Back

Ask the IT provider how long a full restore takes and you will get a number. Ask when that number was last measured and the room goes quiet. Most recovery times are estimates that have survived by never being tested, repeated often enough to feel like facts.

Resilience is not a state; it is a trend. A company that restores its core system in nine hours this quarter and five hours next quarter is becoming resilient in a way a certificate never shows. A company whose number has never moved is one untested assumption away from a very bad week.

The scorecard that captures this fits on an index card. What makes it powerful is that it gets updated, and that someone shows it to the leadership team.

TLDR

• An untested recovery time is a guess with confidence attached.

• Measure real restore times during tests and real incidents; record both.

• Track the trend quarterly: faster, slower, or unmeasured is the whole story.

• Every system on the scorecard has a target, a measured time, and a date.

• A measured miss is progress; an unmeasured success is not.

The number nobody has measured

The gap between the promised recovery time and the tested one is one of the most consistent findings in resilience work. The reasons are mundane: the backup was fine but the rebuild documentation was stale, the person who knew the order of operations was on vacation, the replacement hardware took two days to source.

None of those are exotic failures. They are the ordinary friction that only shows up when the clock is running, which is why the measurement matters more than the estimate.

The distinction between the two numbers is the entire point of comparing recovery time promised versus tested, and it deserves a permanent place in how the company talks about resilience.

The scorecard that fits on an index card

For each system the business cannot run without, record five things: the target restore time, the last measured restore time, when it was measured, how it was measured (test or real event), and the gap. That is the whole instrument.

This fictional example shows a three-system scorecard after a year of measurement. It is an illustration, not a prescription.

System	Target	Last measured	When	Gap
File shares	8 hrs	6.5 hrs (test)	Sep 2025	Beat by 1.5
ERP	24 hrs	31 hrs (test)	Jun 2025	Miss by 7
Email	4 hrs	26 hrs (real, May storm)	May 2025	Miss by 22

Read bottom-up, the card tells the year's story: email was a mess in May, the ERP test found a gap in June, and by September the file shares were genuinely fast. That is what improving looks like on paper.

Where the improvements come from

Restore times fall for boring reasons, and the scorecard exists to fund them. The runbook gets updated because the test proved it wrong. The spare hardware argument gets won because the card shows a seven-hour gap. The second person learns the restore procedure because the card has one name too few on it.

The order of operations matters as much as the speed; the reasoning in setting the restoration order is what makes a fast restore a useful one.

None of these fixes is a purchase order for new technology. They are practice, documentation, and sequencing, which is why the trend line on the scorecard is mostly a measure of organizational attention.

Show it upward, once a quarter

The scorecard earns its keep in the quarterly leadership review, in one minute: here are the three systems, here are the real numbers, here is the direction they are moving. No tooling tour, no acronyms, just time.

Directors and owners understand hours and trends instinctively. A company that can say 'our worst number is eleven hours better than it was a year ago' has made resilience concrete in a way no framework attestation can match.

That is also the answer that matters after a near miss: not that the plan exists, but that the recovery is measurably practiced.

The related discussion of recovery time promised versus tested is worth your time alongside this one. The related discussion of setting the restoration order is worth your time alongside this one.

Frequently asked questions

How many systems belong on the scorecard?

Three to seven: the ones whose absence stops revenue or shipments. More than that dilutes attention and testing capacity. A short list actually measured beats a long list never tested.

What if our measured time is embarrassingly long?

That is the scorecard working. The first honest number is supposed to be uncomfortable; it is the baseline the next two quarters improve from. The only bad outcome is not measuring.

Who should own the scorecard?

One named person on the business side, working with whoever runs IT. The owner schedules the tests, records the numbers, and puts the card in front of leadership every quarter.

A Defensible Next Step

This week, ask for the measured restore time of your most important system, and when it was last measured. If the answer is an estimate, the scorecard's first line just wrote itself. Book a Decision-Clarity Call if you want the first measurement designed with you.