I've been on both sides of serious software delivery — buying it and selling it — and the thing that strikes me every time is how broken the incentives are.
Not broken in a subtle, academic way. Broken in a way that's obvious to everyone involved, and yet nobody changes it, because the brokenness is where the money is.
Getting through the door on serious work is brutal.
Procurement frameworks. Preferred supplier lists. Security questionnaires that take weeks. Due diligence processes designed by people who've never built software but have strong opinions about how it should be governed.
This isn't all bad. Some of it is necessary. If you're handling sensitive data or building systems that matter, you should ask hard questions about who's building them.
But the effect is that only large consultancies can afford to play. The cost of getting through procurement — the lawyers, the compliance teams, the months of unpaid effort — is a tax that only big companies can absorb.
So the big players win. Not because they're better. Because they can afford to wait.
Here's where it gets ugly.
Once a large consultancy wins a contract, their incentive is to make it last. Not consciously, not maliciously — but structurally. They bill by the day. More days, more revenue. The entire business model rewards duration.
So things take longer than they should. Teams are staffed with a mix of senior people (who win the pitch) and junior people (who do the work). Requirements are gathered slowly. Architecture is debated endlessly. The first line of production code ships months after the contract is signed.
Nobody's being lazy. The individuals are often talented and hardworking. But the system they're operating in rewards the wrong thing. It rewards presence, not progress. It rewards complexity, because complexity justifies more people for more time.
The client pays for this. They always pay for this.
And on the other side, the people actually doing the work get squeezed.
The consultancy charges the client £1,200 a day. The engineer doing the work sees a fraction of that. The rest goes to account managers, sales teams, office space, and the procurement machine that got them through the door in the first place.
The engineer knows this. They know they're the product being sold, and that the margin between what they cost and what they earn is someone else's business model. They accept it because the alternative — going independent — means facing that procurement barrier alone.
So they stay. They do good work inside a system that doesn't reward good work. They watch projects drag on and know they could have shipped it in half the time. They keep quiet because the incentive is to keep billing, not to finish.
This is the deal. Everyone knows it. Nobody talks about it.
What offends me isn't the money. It's the waste.
I've watched talented engineers sit in meetings they don't need to be in, working on things that don't need to be this complicated, for clients who are paying for speed and getting ceremony instead.
I've watched clients accept eighteen-month timelines for things that should take three months, because they've been told that's how long "serious" software takes.
I've watched the best engineers — the ones who could actually ship fast — leave the industry entirely, because they got tired of being a line item in someone else's margin calculation.
It's boring. That's the word I keep coming back to. The whole arrangement is profoundly, offensively boring. It's a system optimised for billing, not building. And everyone inside it knows it.
What if the incentives were aligned instead?
What if you paid for outcomes, not hours? What if the people building the software were the same people who won the contract — senior engineers, not a bait-and-switch of juniors? What if finishing fast was rewarded instead of punished?
This isn't utopian. It's just uncommon. Because it requires something that procurement frameworks can't mandate:
Trust.
Trust that the team will do good work without being watched. Trust that fast delivery means efficiency, not corner-cutting. Trust that a small team of experienced people will outperform a large team of mixed ability, every time.
I'll tell you what I've seen work, because I've built a company around it.
Senior engineers. AI-augmented. Small teams. Fixed scope, clear specs, fast delivery. The people who pitch are the people who build. Nobody's billing for time they don't need. Nobody's incentivised to drag things out.
The engineers are well paid — properly well paid, not "well paid minus 60% margin for the consultancy." They have autonomy to do the work the way it should be done. They're trusted to make decisions, because they're the kind of people who make good ones.
And the clients get what they actually wanted: the thing, built well, delivered fast, without the theatre.
Here's the part that companies don't want to hear.
The best engineers — the ones with fifteen years of experience, the ones who can architect a system and ship it — they have options. More options than ever, now that AI has made a single experienced person as productive as a small team used to be.
They will not work somewhere they're not trusted. They will not sit in status meetings explaining what they did yesterday. They will not accept being a billing unit in someone else's spreadsheet.
Give them autonomy, pay them well, trust them to deliver, and they'll build you things that would take a traditional consultancy a year in a fraction of the time.
Micromanage them, underpay them, surround them with process designed for people who can't be trusted, and they'll leave. Quietly. Without drama. They'll just stop being available.
And you'll be left with the people who are willing to tolerate that environment. Which is not the same as the people who are best at the work.
Trust each other. Look after each other. Deliver fast, deliver well, pay fairly.
That's it. That's the whole model.
It's not complicated. It's just incompatible with the way most of the industry works. Because most of the industry is optimised for the wrong thing, and the people who benefit from that optimisation aren't going to change it voluntarily.
So you have to go around them.
Find the people who are good. Trust them. Let them work. Pay them what they're worth. Ship the thing.
Everything else is theatre.