In a famous experiment, owners demanded twice what buyers would pay for the same mug. That's the endowment effect, and it quietly distorts the cloud versus self-hosting decision in both directions. The fix is to decide as if you owned neither option.
When people own something, they value it more than they would if they were buying it new. In a well-known experiment, people handed a coffee mug demanded about twice as much to sell it as a second group was willing to pay to buy the same mug. The gap is the endowment effect, documented by Kahneman, Knetsch, and Thaler, and it survived every attempt the researchers made to teach it away.
It shows up a long way from the lab. The choice to run your own infrastructure or host in the cloud gets argued in spreadsheets: capital versus operating cost, control versus elasticity, total cost of ownership. It should be. But the spreadsheet is filled in by a person, and that person is rarely neutral about what they already run.
If you've built and maintained your own servers, you've put in more than money. There's expertise and identity in a system that works, and loss aversion, the asymmetry where giving something up hurts more than gaining it pleases, makes moving off it feel like a loss rather than a trade. So the system gets valued above what a neutral party would pay for it.
The same force works in reverse. Samuelson and Zeckhauser's work on status-quo bias shows that people stick with the current arrangement largely because changing it requires a decision and keeping it doesn't. If you're already in the cloud, the bill arrives on its own and someone else runs the hardware. Staying is effortless. Leaving is a project. The familiar option wins by inertia, not by merit.
Neither instinct is wrong on its own. The Flexera 2025 State of the Cloud Report found more than half of enterprise and SMB workloads now run in public clouds, and only about a fifth of what moves there ever comes back. That's a real trend, but adoption is not the same as deliberation. A lot of that footprint accreted the way most infrastructure does, one reasonable decision at a time, until where things run was settled by momentum rather than actually chosen. Cloud is the right home for plenty of workloads. Owned or managed infrastructure is the right home for others, especially where data sovereignty, compliance, or steady, predictable load come into play, which is the case I've made in the case for self-hosted infrastructure. The problem is never which option you pick. It's picking it for the wrong reason, or never quite picking it at all.
You can't switch the endowment effect off. Kahneman's subjects couldn't, and they knew they were being tested. What you can do is structure the decision so the bias has less room to work:
Nobody is neutral about something they own. That isn't a gap in skill, it's how valuation works, which is why a second set of eyes is worth more on this decision than on most. An independent read separates "this is the right architecture" from "this is the architecture I'm attached to."
It also helps because the modern version of this question isn't only about cost. Cloud and owned environments both have to be architected, secured, and governed, which puts identity and access management, data protection, and security posture squarely inside the comparison. Those are specialist disciplines, and an honest evaluation treats them as part of the decision rather than something the familiar option handles for free.
This is the kind of call Kief Studio is built to make with a client, precisely because we don't win by selling one answer. We win when the infrastructure fits the business, whether that turns out to be cloud, owned, or some of each.
When the honest answer is the cloud, the real work is architecture, identity, and data governance, not just picking a provider. For that we partner with JDR Security Solutions, a cloud security firm specializing in cloud architecture, identity and access management, and data protection across AWS, Azure, and GCP. Their cloud readiness and health check gives an objective, CISO-level read on whether a cloud environment is built, secured, and costed the way it should be.
When the honest answer is owned or managed infrastructure, we run that end to end on hardware we control, for the workloads and compliance postures where it genuinely fits. Either way the recommendation comes from the workload, not from whatever is easiest for us to sell. You can see how we approach it at kief.studio.
The decision deserves a real answer, and the honest one usually starts the same way. Notice which option you're holding onto, set that feeling down, and choose as if you owned neither. The mug is only worth twice as much because it's yours.
The endowment effect is the tendency to value something more highly simply because you own it. In the Kahneman, Knetsch, and Thaler experiments, owners of a mug demanded roughly twice the price that prospective buyers would pay for the identical item, a gap driven by loss aversion.
Someone who has built self-hosted infrastructure tends to overvalue it because of the labor and identity invested, so leaving it feels like a loss. The same psychology, through status-quo bias, keeps cloud-based teams from re-examining the cloud even when it no longer fits. The bias works in both directions.
For most workloads today the cloud is the sensible default. Industry data shows more than half of enterprise and SMB workloads run in public clouds, with only about a fifth ever repatriated. Owned or managed infrastructure tends to win for workloads with strong data-sovereignty, compliance, or predictable-load characteristics. The right answer depends on the workload, not on ownership instinct.
Ask what you would choose starting from scratch with no existing investment, leave out costs already spent, compare on total cost of ownership rather than sticker price, and account for architecture, security, and data governance. Then match the choice to where the business is headed.
Because nobody is neutral about something they own, which is a feature of human valuation rather than a gap in expertise. An independent assessment, such as a cloud readiness and health check, is the cleanest way to tell whether an architecture is genuinely right or simply familiar.
Survivorship bias overstates the typical mutual fund's reported returns by roughly 1.6 percentage points a year, according to the University of Chicago's Center for Research in Security Prices: surviving U.S. stock funds averaged 8.8 percent over the decade ending 2003, while counting...
In a set of experiments published in 2011, people who assembled a plain IKEA storage box valued it 63 percent higher than an identical prebuilt one. The IKEA effect, named by the researchers who ran that study, is the bias that makes you overvalue what you built yourself. It quietly...
Imagining that a project has already failed raises a team's ability to correctly identify the reasons for a future outcome by about 30 percent, according to research by Deborah Mitchell, Jay Russo, and Nancy Pennington published in the Journal of Behavioral Decision Making. That single...
Work With Us
Kief Studio builds, protects, automates, and supports full-stack systems for businesses up to $50M ARR.
Newsletter
Strategy, psychology, AI adoption, and the patterns that actually compound. No spam, easy to leave.
Subscribe
