The Developer They Would Not Hire
Twenty actuaries on six-figure salaries spend their charity day rebuilding a small charity's databases, because actuaries are good with numbers and everyone knows they code. By Friday the charity has a system nobody on its staff understands. Nobody asks why it did not simply hire a developer.
The brief
The idea arrives in a team meeting. The local hospice shop, food bank or youth charity has run on spreadsheets for twelve years. Donor lists, Gift Aid declarations, volunteer rotas, a grant tracker. Most of it was built by a retired treasurer who has since moved to Spain, and it is held together with macros and hope.
Somebody says the obvious thing. We do this every day. We are actuaries. We will rebuild the lot in a day and give them something proper.
It sounds generous. It sounds like skills-based volunteering, the respectable cousin of the painted fence. It is not. It is unskilled labour wearing a skilled badge.
Actuaries are not developers
Actuaries write a great deal of code. Almost none of it is software.
It is VBA behind a reporting workbook, an R script that runs the mortality study, Python that reshapes model output, Prophet code that only its author can follow. It is written to be run by its author, on his own machine, against a deadline. It is checked by reconciling the answer, not by testing the code.
Software engineering is a different trade. Requirements gathered from the people who will use the system. A data model designed before any code is written. Access control, backups, version control, error handling. An interface a volunteer can use on a Saturday morning without phoning anyone. Documentation. Someone to fix it in three years.
Actuaries do almost none of this as a matter of course. Every insurer now runs end-user computing controls precisely because its own spreadsheets cannot be trusted unsupervised. Being good with spreadsheets is not the qualification for this job. It is how the charity got into the mess.
A system is not a day's work
A charity's core records cannot be rebuilt in eight hours. Nobody scopes the job properly, because scoping would reveal that. So the day ends in one of three ways.
The first is a half-finished migration: some data moved, some not, two systems running in parallel and neither trusted. The second is a new tool that works on the day and breaks at the first Office update, or when the one actuary who understood it changes jobs. The third is quieter: weeks of unpaid evenings from two keen volunteers, then silence when their own year-end arrives.
In every case the charity is left maintaining code it did not ask for, cannot read, and cannot get support for. The old spreadsheet was ugly, but somebody on site knew how it worked. The new one is understood by nobody in the building.
The data is not theirs to touch
A badly painted fence harms nobody's privacy. A charity's databases are different. They hold donors' names, addresses and bank details, Gift Aid declarations, volunteers' DBS records, and often details of the people the charity serves, including health information that UK GDPR treats as special category data.
The charity is the data controller. On charity day, twenty outsiders get access to all of it, usually by copying files onto firm laptops or personal drives. There is no data processing agreement, no access log, and no plan for deleting the copies afterwards. If one laptop goes missing, the charity has a reportable breach and its trustees answer to the Information Commissioner.
Gift Aid adds a sharper risk. HMRC expects a charity to hold an auditable record of every declaration it claims on. A migration that drops or garbles those records can leave the charity unable to evidence past claims, and HMRC can recover Gift Aid that cannot be supported. The free rebuild can cost the charity money it has already spent.
The developer nobody counted
There is a trade that does this work. Small software houses, freelance developers and IT providers that specialise in the charity sector. For them a small charity's database rebuild is a real job, the kind that keeps a two-person firm going for a month.
Off-the-shelf charity CRMs also exist, sold at charity pricing, with support, backups and Gift Aid handling built in. A competent provider would probably tell the charity to buy one rather than build anything bespoke. That advice alone is worth more than the day.
When twenty actuaries arrive for free, that customer disappears. The need never surfaces as demand. The transfer runs from small technology businesses to a large firm's reputation, and the firm records it as generosity.
Nobody has read the policy
Who is on cover when the rebuilt donor database corrupts the Gift Aid file?
The actuaries' firm carries professional indemnity insurance for its professional business: actuarial advice, priced by underwriters who understand it. Building software for a third party, unpaid, is not that business, and the wording may well exclude it. The individuals carry no cover of their own.
The charity may hold cyber insurance, but those policies come with conditions on security and on who touches the systems. Twenty unvetted outsiders with copies of the data is not the risk the underwriter priced. The claim lands, once again, in the gap between two policies.
A developer or CRM provider carries the cover as a condition of doing the work. That is part of what the charity gives up when it accepts the free version.
Only one party wants the system
The charity would rather have the money, or a supported product. The developer would rather have the job. One participant is left who prefers the rebuild: the firm.
And the firm gets real value. A hackathon is good fun. Junior staff practise Python on live data with no reviewer. Partners get a skills-based volunteering story for the annual report. That is training and team-building, and the charity is supplying the venue and the data.
Stated honestly, it would read: we used a charity's records as a training exercise and left it to maintain the result. Nobody would put that in a press release, which is why it is never phrased that way.
The test is the same as for the fence. Offer the charity a straight choice: twenty actuaries for a day, or the cash to pay a developer. Then abide by the answer.
The honest version
None of this argues for corporate indifference. It argues for three separate things, each named as what it is.
Buy the team day commercially. If the aim is a hackathon, book one, or run it on the firm's own data where the reviewers and the controls already exist.
Donate cash for the real job. Fund a proper developer, or two years of a supported charity CRM. The charity knows what it needs. It is rarely bespoke code.
Give the actual skill. An actuary's comparative advantage is not building the system. It is being the informed client. Write the requirements. Check the provider's quote. Reconcile the migrated data against the old spreadsheets, record by record, which is exactly what actuaries are good at. Then take a trustee seat, model the reserves, and look hard at the defined benefit scheme many older charities still carry. That work is scarce, and it undercuts nobody.
How it stops
Everyone in the chain is behaving rationally. HR is measured on engagement, marketing on content, the fundraiser on corporate relationships. Nobody's objectives include whether the charity's systems still work next year.
It stops when someone asks four questions. What would the charity have chosen? Who maintains it now? Where did the data go? Who was on cover? None of them hostile. All of them currently unasked.
Until then, actuaries will keep rebuilding charities' spreadsheets, badly, and the developer who should have been hired will keep wondering why the sector stopped calling.