Nobody Gave Her The Headcount. She’s Doing It Anyway.

Sometime in 2022 I sat down to do a job I’d done a hundred times before and finished it in about four minutes.

It should have taken an afternoon. It had always taken an afternoon.

Little did my team know, I had developed a secret relationship with ChatGPT and I was firmly in the honeymoon period.

I sat there afterwards with the specific expression of a man who has watched a card trick up close and still cannot work out where the coin went, and then I did what any reasonable software founder does in that situation. I did it again. Then I did a different one. Then I did a spreadsheet job I’d been putting off for a fortnight, and that took six minutes, and at that point I stopped enjoying myself.

Because here is the thing about that feeling. It does not stay yours.

Everyone I sell to was about to sit down and have the same four minutes. And the second they did, every expectation they held about every piece of software they owned was going to shuffle six feet to the left, quietly, with nobody announcing it. Including their expectations of us.

I have spent a lot of my life being smug about being early to things. This was not that. This was watching a wave form a long way out and doing the arithmetic on whether you are far enough up the beach.

The bit where I tell you what we’d actually shipped

Our reporting, at that point, worked like this. You asked the platform for your data. The platform got you your data. It put it in a CSV. You downloaded the CSV.

That was it. That was the feature.

And I want to be careful, because I am about to be unkind about something that was, for its era, completely fine. In 2013 that was a respectable answer. In 2018 nobody complained. It did the job it was asked to do and it did it reliably, which is more than a great deal of enterprise software manages on a good day.

But a CSV is not an answer to a question. A CSV is a polite way of saying “here is everything, you sort it out.” And the people we sell to were about to stop accepting that, because they had just watched a machine sort something out for them in four minutes and nobody goes back from that.

Sigh.

Meet the person this is actually about

There is a Head of Events at a global corporate bank who I have known for years. Sharp, funny, extremely good at her job. And over the time I have known her I have watched that job quietly turn into a different job while everybody involved carried on describing it as the same one.

In 2018 the questions she got were about sparkle. How do we make this look extraordinary. What is the new technology, can we use the screens, can we do the thing with the lights. Confetti and sparkles, and I say that with real affection, because it was a good job, she was brilliant at it, and the events were genuinely wonderful.

In 2026 the questions are: can you justify this cost. Who attended, from which client, and what came of it. What will Compliance ask. What will Risk ask. What will the two people sitting above both of them ask.

Nobody changed her title. Nobody gave her one extra person. The questions simply started arriving and never stopped, the way water arrives in a basement.

I want to be precise about the actual request, because this is the paragraph where people who do this for a living either nod or close the tab. It is end-to-end reporting on individuals. Who was invited, who turned up, what it cost, what happened next. Reasonable. Obvious, even. You would assume it was a button.

It is not a button. Her firm’s CRM holds the client data. Her event platform holds the invitations and the attendance. And the glue is her effort. She pulls reports out of both. She builds a document that stitches them together. She sends it up the line. Then she throws it away, because it was never a report, it was a one-off, and next quarter she will build it again from nothing.

That is compliance work. It has a different name on the org chart and no headcount attached to it, but that is what it is, and she is doing it in the gaps around her actual job.

She is also not remotely unusual, and there are two solid reasons why.

The first is volume. The Bank Policy Institute asked twenty of its member banks, about half its membership, to count the hours their people spend complying with financial regulation and examiner mandates. Between 2016 and 2023 those hours rose 61 percent. Total employee hours at the same banks rose 20 percent. Compliance work grew three times faster than work did. It is US data, it is self-reported, and BPI is an organisation with firm views on whether regulation costs too much, so read it as directional rather than gospel. But nobody has produced a number pointing the other way, and seven years is a long time to be wrong in a single direction.

The second is that somebody has since written it into law. DORA has applied to EU financial entities since January 2025, and among other things it requires firms to keep a register of their technology suppliers and to perform documented due diligence on each one. Britain is not in it, but every London bank with a Dublin or Frankfurt entity is, and the UK runs its own version through the FCA and the Bank of England’s critical third parties regime. What that means on her desk is simple enough. Buying a piece of software has quietly become a regulated act, and somebody inside the building has to evidence it.

There is a third thing I will offer more carefully. Ncontracts, who sell compliance software and therefore have a dog in this fight, surveyed 183 financial institutions in December and found 64 percent expecting their compliance budgets to stay flat or shrink over the next 12 to 18 months. Smaller American institutions in the main, so hold it loosely. It is the only figure in this piece I would not stake a deal on. It simply happens to match what every compliance person I speak to says out loud.

More work. Same money. The difference has to land on somebody, and it never lands on the person who wrote the regulation.

And now the bit I would rather not write

We did not escape this. We are in it. We are, in a small and unflattering way, part of it.

When a client asks for reporting the platform we built a decade ago was never designed to produce, we build it by hand. Our project managers assemble the report. To assemble it they go and ask back-end engineers to run deep data exports, specific to a client, an event, sometimes an individual attendee. It is slow, it is expensive, it burns people who should be building other things, and it is the reason the whole platform is being reimagined rather than extended. I would rather tell you that in my own words on my own website than have you discover it in a reference call.

There is a rule we made early that I still stand behind. Bespoke work never goes into the product. If a capability will be used across the whole client base, we build it properly. If it exists for one client, we absorb it as effort and the product stays clean. That rule is why our software has stayed coherent for thirteen years instead of becoming the usual haunted house of features nobody remembers commissioning.

The problem is that the rule quietly assumed these requests would stay rare.

They did not stay rare. The same request kept arriving, from every client, one ticket at a time, and we kept absorbing it, and eventually I had to sit down and admit I was not looking at a series of exceptions. I was looking at a product telling me, clearly and for several years, what it needed to become.

That is what convinced me. Not a competitor. Not a board, because there isn’t one, and every one of these calls has been mine to get right or get wrong. It was a stack of hand-built reports and the growing certainty that the gap between what she needed and what we handed her was widening, and that we were the ones allowing it to widen.

Tom’s story, which is worse than mine

Our Commercial Director has a better version of this than I do, and he tells it with less self-pity.

A law firm had all but decided. The events lead was bought in, the case was made, and the security review was already running, which in a legal sale is not a complication, it is Tuesday. Tom does those reviews for a living. The documentation goes over before anybody asks for it. Nobody was surprised by any of it.

What stopped the deal was the gate behind the gate. A due diligence step that had not existed the last time that firm bought software: a supplier register to update, diligence to evidence, questions their own process had never thought to ask. And no, before anybody says it, that was not a firm inventing work for itself. This is what buying software has quietly turned into, and it arrived in person, wearing a lanyard.

Nobody owned it. Not IT, who had already done their bit. Not procurement, whose process was written before the requirement existed. So it went where this work always goes. The events lead gathered what the step needed, learned more about data hosting than any reasonable person would want to, argued about access controls with people she had never met, and defended the lot internally on top of the job she was hired for. Which, I remind you, is running events.

She also, and this is the part Tom repeats, built the business case herself. She used the pain of running events without a proper platform as the evidence, because she was the only person in that building who could describe it first hand.

She was not asked to run an event. She was asked to staff a compliance step that did not exist a year earlier, win an argument with two departments, and then get back to her desk and sort dinner for ninety people.

Nobody is staffed for this. That is the entire point. There is no team called Proving Things.

So what do you actually do

If you recognised yourself somewhere up there, the most useful thing I can give you is not a demo. It is a reframe, and it is free.

The reporting you rebuild by hand every quarter is not an irritation at the edge of your job. It is a permanent part of your job that arrived without paperwork. So treat it like one. Name it. Put a number on the hours. Count them properly, across a quarter, including the evenings.

That number is always larger than everyone expects, including the person doing the work, and it is the most persuasive line available to you in any business case, because it is money already being spent. You are not asking for new budget. You are showing your finance director an invoice the firm has been paying invisibly for four years.

And when you assess any event platform, ours very much included, ask one question early and watch what happens to the room. How much of this weight do you take off me? Ask for the security documentation before anybody offers it. Ask whether the report you rebuild every quarter could simply be a page you open.

The good ones will have answers ready, in a form you can forward rather than rewrite. The rest will tell you about their roadmap.

We have not finished this work ourselves, and I would be wary of any vendor who tells you they have.

The end bit

Thirteen years in, and what our clients need from us has changed more in the last four than in the previous nine. The work of proving things, justifying things, defending decisions and answering for spend has quietly relocated into jobs that were never built to hold it. Nobody sent a memo. Nobody attached a budget line. It moved in and started making demands, like a lodger.

The tools have to carry that weight now, because the alternative is that she carries it, forever, at eleven at night, rebuilding the report she already built in March.

The event is still the job. Proving it is now also the job.

Nobody announced that. It happened anyway.