The running-cost comparison always favours building; development, upgrades and timing usually reverse it, until the fund count rises.
Put both options on the same five-year cash flow, including the build's development cost upfront and the vendor's licence escalation, and discount them. For an illustrative manager with six funds, buying costs 1,470,636 in present value against 1,517,471 to build, although the build is 36,152 cheaper in undiscounted cash. The answer turns on the number of funds: the licence is priced per fund and the build is not, so building wins from about 6.27 funds upwards on these inputs.
Worked in full in The Private Markets COO by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
The build case is almost always presented the same way: the vendor wants 240,000 a year in licences plus an administrator to run it, an in-house engineer and some cloud capacity cost 220,000, so building saves 80,000 a year, 400,000 over five years. That comparison leaves out the three things that decide it: the development cost before the first report runs, the upgrade the platform will need when the data model or the reporting templates change, and the fact that the money is spent at different times. A COO who signs off a build on the running-cost comparison has approved a project that breaks even in cash only in year 5 and never does in present value at six funds.
| Input | Buy | Build |
|---|---|---|
| Funds on the platform | 6 | 6 |
| Year 0: implementation or development | 180,000 | 520,000 |
| Licence per fund a year, rising 5.0% a year | 40,000 | - |
| Internal administrator, a year | 60,000 | - |
| Engineer and analyst time, a year | - | 190,000 |
| Cloud and tooling, a year | - | 30,000 |
| Upgrade in year 3 | - | 150,000 |
| Discount rate | 8.0% | 8.0% |
All figures are illustrative. The build's year-3 upgrade stands for what every in-house platform eventually needs: a new data source, a reporting standard that changes the template, a security review that forces a rewrite. A vendor carries that cost inside the licence.
Cost in year t, buy = licence per fund × funds × (1 + escalation)t−1 + administrator; year 0 = implementation
Cost in year t, build = run + cloud + any upgrade; year 0 = development
Present value = Σ costt ÷ (1 + discount rate)t
Excel, costs for years 0 to 5 in B2:B7: =B2+NPV(8%,B3:B7)
| Year | Buy | Build | Buy less build | Cumulative |
|---|---|---|---|---|
| 0 | 180,000 | 520,000 | -340,000 | -340,000 |
| 1 | 300,000 | 220,000 | 80,000 | -260,000 |
| 2 | 312,000 | 220,000 | 92,000 | -168,000 |
| 3 | 324,600 | 370,000 | -45,400 | -213,400 |
| 4 | 337,830 | 220,000 | 117,830 | -95,570 |
| 5 | 351,722 | 220,000 | 131,722 | 36,152 |
| Total | 1,806,152 | 1,770,000 | 36,152 |
The cumulative column is the payback: the build's 340,000 head start in cost is earned back only in year 5, by 36,152. Discounted at 8.0 per cent, the year-0 cash weighs more than the later savings, and the order reverses: buy 1,470,636, build 1,517,471, a difference of 46,835 in favour of buying.
The licence scales with the number of funds and the build, to a first approximation, does not. Each additional fund adds 175,179 to the present value of the buy option over five years.
| Funds | Buy | Build | Build less buy |
|---|---|---|---|
| 4 | 1,120,278 | 1,517,471 | 397,193 |
| 6 | 1,470,636 | 1,517,471 | 46,835 |
| 8 | 1,820,994 | 1,517,471 | -303,523 |
| 10 | 2,171,352 | 1,517,471 | -653,881 |
| 12 | 2,521,710 | 1,517,471 | -1,004,239 |
The break-even is (1,517,471 − 419,563) ÷ 175,179 = 6.27 funds, where 419,563 is the present value of the buy option's costs that do not depend on fund count. A manager raising its seventh and eighth funds within the horizon should run the comparison on the fund count it will have, not the one it has. Two other inputs move the answer as much. Over a seven-year horizon, with a second upgrade in year 6, the build comes out 45,144 cheaper at six funds. A 30 per cent overrun on development, which is not unusual, makes it 202,835 dearer.
Comparing year-one running costs: 300,000 against 220,000, "a saving of 80,000 a year". It omits the 340,000 of extra spend in year 0 and the year-3 upgrade, and it treats a cost paid now as equal to a saving received in year 5. The second mistake is the opposite one: costing the vendor at today's fund count and its year-one price, when licences escalate and per-fund pricing grows with the platform. Neither option is cheap or dear in itself; the fund count and the horizon decide.
Cost is the first test, not the only one. A build depends on one or two people who understand the data model, and the operational risk of losing them belongs in the decision. A vendor contract needs data export rights and a termination path, or the switching cost lands in a later budget.
Data, service providers and the operating model of the management company are the book's subject, and the free workbooks for this book include the vendor register and capacity model a decision like this draws on. For the same arithmetic applied to the administrator, see in-house vs outsourced fund administration.
Only at scale or over a long horizon. In the illustration a six-fund manager pays 1,470,636 in present value to buy and 1,517,471 to build over five years. Because licences are priced per fund, each extra fund adds 175,179 to the buy option, and building becomes cheaper above about 6.27 funds.
The development cost before go-live, periodic upgrades when data sources or reporting standards change, and the timing of cash. The illustrative build needs 520,000 upfront and 150,000 in year 3; comparing running costs alone, 300,000 against 220,000 a year, suggests a saving of 400,000 over five years that does not exist in present value.
Very, because development is paid first and is not discounted. In the illustration a 30 per cent overrun on the 520,000 build turns a 46,835 disadvantage into 202,835. A longer horizon works the other way: over seven years, with a second upgrade, building comes out 45,144 cheaper at six funds.
This article is one calculation from The Private Markets COO. 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.