A carrier I worked with ran an inventory of its enterprise software and found a familiar pattern. A meaningful share of the licenses it paid for every year were barely used. Some weren't running at all. One platform had been bought with executive fanfare 18 months earlier, announced in a press release, integrated about halfway, then quietly abandoned when the sponsor moved on. The seats were still provisioned. The vendor still listed the airline as a customer. The invoice still cleared every quarter.
Nobody had set out to waste the money. Each purchase had a reason at the time—a relationship, a conference, a competitor who'd just signed the same deal, a slide that needed a modern logo on it. What none of them had was an owner accountable for an outcome. Software got bought the way trophies get bought, and trophies don't get audited. Once a system is installed, asking whether it actually produced anything feels like criticizing the people who championed it. So the renewal sails through, the vendor adds your logo to its next pitch, and the question never gets asked.
And the license fee is the cheapest part of the bill. Every unused system still carries something: an integration to maintain, a security surface to watch, a slot in the architecture the next tool now has to work around. The most expensive cost never shows up on the invoice at all. It's the capability you didn't build because a box was already checked. "We already have a tool for that" is how good ideas die in enterprise IT—killed not by a no, but by a logo on a slide nobody ever switched on. Shelfware isn't idle. It's in the way.
The instinct when a platform underdelivers is to assume you bought the wrong one and go shopping again. A different vendor, a newer category, a bigger contract. But the second purchase fails the same way the first did, because the missing piece was never the technology. It was the discipline of tying every system to a result someone has to defend.
What changed things was boring. Before any renewal, one question: what outcome is this producing, and who owns it? Systems that couldn't answer got cut, not replaced—and the surprise was how few could answer at all. The savings funded the handful of capabilities that could actually point to a number. The software estate got smaller and the results got bigger, which is the opposite of how every modernization budget gets sold.
This is the trap underneath a lot of airline technology spend. A signed contract feels like progress. A logo on a vendor's wall feels like modernization. Neither is a result. The test isn't whether you bought the thing or whether it demos well—it's whether anyone can name what it changed.
Technology you can't tie to an outcome isn't an asset. It's a subscription for looking modern.