Expected annual cost of three recovery tiers, the quarter-end outage that sets the real requirement, and the payment workaround that is cheaper than a faster tier.
Set the recovery time objective from the gap between hard payment deadlines, and the recovery point objective from what lost work costs to rebuild. On illustrative figures, backup and restore with a 30-hour RTO costs a fund manager 77,520 a year in expectation, the cheapest of three tiers. But an outage in a quarter-end week, when deadlines fall every 8.3 business hours, costs 562,000. A 15,000 manual payment workaround cuts that to 173,200, which is cheaper than buying a faster tier.
Worked in full in Operational Due Diligence in Private Equity by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
The recovery time objective (RTO) is how long systems may stay down after a failure. The recovery point objective (RPO) is how much data may be lost, measured back from the failure: a 10-hour RPO means a day's entries may need to be rebuilt. Operational due diligence questionnaires collect both, and managers answer with whatever their IT provider offers. The useful question is whether those numbers fit the firm's own calendar, because a private markets manager does not lose money by the minute like a trading desk. It loses money when it misses a wire that had to go that day.
| Input | Value |
|---|---|
| Major outages needing recovery, per year | 0.20 |
| Hard payment deadlines a year (calls, closings, distributions, facility interest) | 40 |
| Of which in the 4 quarter-end and call weeks (20 days) | 24 |
| Business days and hours | 250 days, 10 hours, 2,500 hours |
| Average cost of a missed deadline | 120,000 |
| Cost of staff unable to work, per business hour | 3,000 |
| Cost to rebuild lost work, per hour of RPO | 4,000 |
| Tier | RTO (business hours) | RPO (hours) | Annual cost |
|---|---|---|---|
| Backup and restore | 30 | 10.0 | 40,000 |
| Warm standby | 8 | 1.0 | 120,000 |
| Hot standby | 1 | 0.1 | 320,000 |
Deadlines missed = deadline density per business hour × RTO
Cost per outage = RTO × idle cost + deadlines missed × penalty + RPO × rebuild cost
Expected annual cost = tier cost + outages a year × cost per outage
Average density = 40 ÷ 2,500 = 0.0160 an hour. Backup and restore: 0.0160 × 30 = 0.48 deadlines missed; 90,000 + 57,600 + 40,000 = 187,600 per outage; 40,000 + 0.2 × 187,600 = 77,520 a year
| Tier | Deadlines missed | Idle | Penalties | Rebuild | Per outage | Annual total |
|---|---|---|---|---|---|---|
| Backup and restore | 0.48 | 90,000 | 57,600 | 40,000 | 187,600 | 77,520 |
| Warm standby | 0.13 | 24,000 | 15,360 | 4,000 | 43,360 | 128,672 |
| Hot standby | 0.02 | 3,000 | 1,920 | 400 | 5,320 | 321,064 |
Warm standby saves 144,240 per outage for 80,000 more a year, so it pays only if major outages happen 0.55 times a year or more, about once every 1.8 years. On averages, the cheapest tier wins.
Averages hide the calendar. In the quarter-end and call weeks, 24 deadlines fall in 200 business hours, a density of 0.120 an hour, one every 8.3 hours. In the rest of the year the other 16 are spread over 230 days, one every 144 hours.
| Tier | Deadlines missed | Penalties | Total | With payment workaround |
|---|---|---|---|---|
| Backup and restore | 3.60 | 432,000 | 562,000 | 173,200 |
| Warm standby | 0.96 | 115,200 | 143,200 | 39,520 |
| Hot standby | 0.12 | 14,400 | 17,800 | 4,840 |
The 30-hour RTO, three business days, is longer than the gap between deadlines when it matters most. That is the real finding, and an average-cost model never shows it.
Most of the quarter-end loss is penalties, not idle staff, and penalties arise only because payments depend on the same systems as everything else. A manual payment route, bank portals reachable from a clean device, signatories holding their own tokens, a printed list of the week's wires and callback numbers, costs perhaps 15,000 a year to keep tested and is assumed here to avoid 90 per cent of missed-deadline cost. It turns a 562,000 quarter-end outage on backup and restore into 173,200. The callback numbers matter: a crisis is when fake payment instructions arrive, as what an untested callback control costs shows.
| Outages a year | Backup and restore | Warm standby | Hot standby | Backup with workaround |
|---|---|---|---|---|
| 0.1 | 58,760 | 124,336 | 320,532 | 68,576 |
| 0.2 | 77,520 | 128,672 | 321,064 | 82,152 |
| 0.5 | 133,800 | 141,680 | 322,660 | 122,880 |
| 1.0 | 227,600 | 163,360 | 325,320 | 190,760 |
On expectation the workaround costs 4,632 a year more at 0.2 outages; what it buys is the removal of the 562,000 year. Hot standby is the dearest tier at every outage frequency in the table.
The rating model in which a finding like this is scored, with veto rules for the issues an average would bury, is in the free workbook for this book.
The recovery time objective is how long systems may stay down; the recovery point objective is how much recent data may be lost, measured back from the failure. In this illustrative case backup and restore gives an RTO of 30 business hours and an RPO of 10 hours, warm standby 8 and 1, and hot standby 1 hour and 0.1.
One shorter than the gap between hard payment deadlines at the busiest time of year, or a payment route that does not depend on the main systems. Here quarter-end weeks carry 24 of 40 deadlines in 20 days, one every 8.3 business hours, so a 30-hour RTO misses 3.60 deadlines if the outage lands then, against 0.48 on average.
Rarely on cost. At 320,000 a year it costs 321,064 in total expectation, against 77,520 for backup and restore and 128,672 for warm standby, because a private markets firm has few deadlines measured in minutes. Even at one major outage a year, warm standby, at 163,360, remains cheaper than hot standby at 325,320.
This article is one calculation from Operational Due Diligence in Private Equity. The book takes the same case from first principles to the decision, chapter by chapter, and every figure it prints is a live formula in the free companion workbooks.
Get the book on Amazon →Free companion files
Also on Amazon UK · Amazon Germany · Amazon France · Amazon Canada
Reading guide: private equity and private markets → · All 1340 articles →
If this book helped, or didn’t, a few lines on Amazon are worth more than they look: they are what the next reader goes on. Write a review. The workbook stays free either way.