Innovation, Now Sold Separately

Or, how your bug fix became a technical advance at the very frontier of human knowledge.

“Image generated using OpenAI’s DALL·E.”

There is a special time of year, dear reader, one that arrives with all the warmth and spontaneity of a compliance calendar reminder, mandatory “health and safety” training, or just the usual dollop of HR mandated fun, when some poor soul (very often the one who has touched the actual codebase least in living memory) is handed a spreadsheet, a consultant’s template thick with italicised “guidance notes”, and the unenviable task of proving, to the satisfaction of a tax authority, that the company spent the entire year “Innovating”.

And that’s “Innovating” with a capital “I”.

This isn’t your usual fare of your common or garden software type either. Not the fixing of bugs at two in the morning (out of OCD induced anxiety rather than overtime, note), not the forking of someone else’s open source project and quietly filing the serial numbers off (at the instruction of management, over the phone, and never by traceable email, note), and not the wrapping of the same library everyone else also wrapped and calling it a “platform” (see last parenthesis).

No, Innovating, at the very frontier of human knowledge, under conditions of grave and heroic “technical uncertainty”, Innovating that ticks the boxes of the tax authorities rather than the technicals themselves.

You see, the R&D tax credit does not reward research, and it certainly does not reward development. It rewards the narration of both. Somewhere a government decided, quite reasonably (which, granted, is highly unusual for a government), to hand money back to companies genuinely pushing the boundaries of science and technology, and somewhere shortly afterwards an entire cottage industry sprang up to explain to the rest of us how our perfectly ordinary year of graft could be dressed, powdered, and paraded as precisely that.

As SMEs have come to learn, to the delight of the PowerPoint brigade and denizens of middle-management but definitely not to the people actually doing the hard work — the deliverable was never the code, oh no, the deliverable is the paperwork.

Yes, indeed, SMEs have become the EU Commission itself, in miniature.

And what a load of paperwork it really is. The template speaks a peculiar dialect, one you will never hear at a standup and never see in a commit message, in which every developer becomes a lone pioneer and every afternoon of debugging becomes a “systematic investigation”.

You are required to identify the “technical advance you were seeking”, the “technical uncertainty” you so bravely faced, and the “readily deducible solution” that, thank the Elder Gods, did not previously exist¹.

The guidance notes, helpfully italicised throughout, all but tell you which uncertainty you were supposed to have felt at the time. You did not feel it², and that’s not through lack of soft skills — that’s actually through your genuine technical acumen. You were just trying to get the sprint closed before the retro and didn’t notice the middle-managers rubbing their hands reading the latest government pamphlet, well, infographic, on how much they could claim this year (rather than doing actual work, which would require the employment of actual technical people and customers that prefer solutions rather than going on “a journey” during every Teams meeting).

The real fun begins with the archaeology, of course, because nobody, and I do mean nobody, wrote any of this down while it was happening. Well, ok, I’m a little rash. I document furiously and with purpose, but so few people do these days — especially when their SlopBot session hangs.

Their entire heroic narrative has to be reconstructed, months later, from git blame, from stale Jira tickets, from the fossil record of a dead Slack channel, and from the flickering recollections of engineers who genuinely cannot recall what they did last Tuesday, let alone which frontier of human knowledge they were nudging back in March.

The ones who can remember, naturally, embellish. A caching tweak becomes “a novel approach to distributed state”. A retry loop becomes “resilience research”. Everyone is suddenly Alan Turing, retroactively, for tax purposes.

Here is the part that tickles me most: the person compiling the claim is almost always the person who did the least of the actual building.

It has to be, because narration is a wholly different skill from engineering, and the greybeards who did the real work are far too busy doing more of it to sit around mythologising the last lot. So the definitive account of the year’s collective genius ends up authored by whoever is furthest from the keyboard, and, in a delicious little inversion, that document becomes the single source of truth. The most authoritative artefact in the entire building is the one describing work its author never once touched.

Then comes what is termed, somewhat laughably, “consolidation”. Twelve small, honest, entirely unremarkable jobs are quietly rammed together into four grand “Programmes”, because four Programmes read as “Research” whereas twelve tickets read as Tuesday.

And then, naturally, the whole assembled edifice goes up to the CTO for a “final polish”, which as you can imagine is yet another wonderfully genteel euphemism. Because a chief “technology” officer, presented with a sheaf of documents describing work they mostly did not do and largely did not follow, has precisely one reflex… to make it all sound considerably more impressive than it was.

The modest is recast as ambitious, the routine as visionary, “we upgraded a dependency” becomes “we undertook a strategic re-platforming”, and the odd project they dimly recall championing from the doorway of a standup acquires, in the retelling, a heroism it never once earned.

They may even bolt on one of their own, hastily, at the eleventh hour. Or two. Or three. After all, like Bilbo famously said, “Why Shouldn’t I?”

It is the CTO doing what they do best, you see… not building the thing, but describing, magnificently, the thing they would have built if they only had the time, the resources, oh and the technical skill (but they usually breeze over that bit when challenged, or just don’t reply to mails about it).

Grandeur, you understand, is the unit of account. And the whole glorious confection is then graded for “merit”. This is the merit of work that, may I remind you, nobody documented at the time and half the participants cannot even remember, by people whose entire remit is to find it overwhelmingly meritorious!

Which brings us, inevitably, to the “Specialists”.

The consultancy that supplies the template, coaches the dialect, and blesses the finished fiction is, more often than not, paid a percentage of the claim. And, yes, of course there are, because that is how things are now — layers, upon layers, upon layers, of smoke, mirrors, slop, and sludge.

Every. Single. Day.

I shall just let that settle for a moment. Their incentive is not accuracy, their incentive is a big number for paper shuffling, which is (call me a cynic) most likely these days done by a copy paste into a language model and a walk to a local coffee shop (and back).

“Claim everything,” they murmur soothingly, “and we’ll trim it later³,” which is a marvellous way to align every soul in the room behind the most flattering possible version of the truth, complete with a laminated methodology to lend the whole business a reassuring heft.

And so the annual festival concludes with a triumphant submission, a tidy rebate, and a shared, unspoken understanding that the actual innovation, if there ever was any, happened in the small, unbilled, undocumented moments nobody thought were worth recording.

The very moments, again, of course, the credit is structurally incapable of ever seeing at all.

The hours on a timesheet were always fiction, as we established once before. But now it turns out that the genius is too.

Innovation, now sold separately. Batteries not included. Terms and conditions apply. Someone, somewhere, is quietly trousering a percentage of the whole beautiful lie.

[1]: “Readily deducible”, incidentally, is the phrase doing all the heavy lifting, and it means roughly “any competent professional could have worked it out”, which describes, if we are being honest with each other, very nearly everything we do. [2]: I did try to feel the uncertainty, dear reader but, to be perfectly honest, I mostly just felt the deadline. [3]: A percentage. Of the fiction they helped you write. Nice work if you can get it, and they certainly got it.