In January 2018, Manchester United signed Alexis Sanchez from Arsenal. He was 29, one of the most talked-about attackers in the league, and his base wage was £350,000 a week — rising past £500,000 once bonuses and image rights were counted. Eighteen months later, with five goals across all competitions and no sign of the player who'd torn the Premier League up at Arsenal, he was loaned to Inter Milan and eventually left on a free transfer.

The wage was paid in full regardless.

Pick any season in the last decade and there's a version of this story. Mesut Ozil's final years at Arsenal. Gareth Bale on Real Madrid's bench. Eden Hazard at the Bernabeu. The shape is the same: the club commits to the full cost up front, and whatever happens next — injury, loss of form, tactical mismatch, a falling-out with the manager — the money leaves the account every Friday anyway.

Zoom out from the individual case and the pattern is structural, not anecdotal. UEFA's annual finance report has, for years, treated a wage-to-revenue ratio of 70% as the informal warning line, and Premier League clubs in aggregate have spent the last decade bumping up against it, with several specific clubs above it in lean years. That's before transfer amortisation, before agent fees, before the rest of the commitments stacked on top. The wage bill is the single biggest line item in football finance, and most of it is contractually fixed whether the team turns up or not.

This is the same problem enterprise software had fifteen years ago.

Before cloud, software was sold on licenses. You paid a fixed fee for a seat or a server or a site, whether your team used the thing heavily or barely at all. The pricing had nothing to do with the value you actually got out of it. Then AWS showed up and made the model look absurd. Rent a virtual machine by the hour, shut it down overnight, and you don't pay for overnight. Idle resource, no bill. You pay for what you consume. The rest of the software industry has spent the last decade slowly, unevenly, moving in that direction.

(For readers who don't work in software: the shorthand for this is "usage-based billing" — you pay per API call, per gigabyte, per transaction, per whatever unit of value the vendor actually delivered. The older model is closer to paying a flat annual fee to have the software in the building, regardless of how much of it you touch.)

Football sells the old model. A contract is a commitment to pay for time, not performance. The club rents the seat; what happens in the seat is a separate matter.

It's worth naming where the analogy stops being useful before going further. Virtual machines are fungible — if one is broken you spin up another and nothing upstream notices. Footballers aren't. You can't horizontally scale your centre-forwards, and when the £500k-a-week striker is out, the bench option isn't a cheaper version of the same thing; it's a different player having a different effect on a team that was built around somebody else. That matters, and it's part of why the problem in football is arguably worse than it ever was in software.

A few reasons it's worse.

Commitments are longer. A three-year software contract is long. A Premier League deal of 3-5 years, with options, is standard. Very few industries lock in labour cost on that horizon.

The secondary market is thin and visible. Every transfer fee and every wage is reported, and each one becomes the benchmark the next agent walks in with. When the wage slip for a rival's centre-back leaks, that number is now the floor for the next negotiation.

And the risk concentrates rather than pools. A software vendor can spread failures across thousands of customers. A club can't spread across eleven players. When the one expensive striker tears an ACL in September, the replacement isn't a reserved instance; it's a seventeen-year-old from the academy or a panic signing in January.

None of this is news to anyone who follows football finance. The wage-to-revenue ratio gets cited in every piece about Financial Fair Play. The "wasted wages" narrative is a staple of every summer transfer window. What hasn't really been asked — at least, I haven't seen it asked seriously — is the question enterprise software answered fifteen years ago: is there a way to tie the cost more directly to what the club actually gets out of it?

That's what the next few posts in this series are going to chew on. Not because football should literally meter goals and charge by the API call — a handful of the more excitable takes on this idea have already gone that way, and they read as cute more than useful. But the mechanisms that solved the equivalent problem in software — metering actual consumption, credits for downtime, availability guarantees, outcome-linked pricing — are worth taking seriously, one by one, and asking which ones survive contact with the actual shape of football.