Outfox Software's Sunk-Cost Fallacy: How to Spot It and Escape

Think back to the day you picked that one software solution—the one you feel saddled with today.

You weren't careless about it. You ran the demos, read the reviews, and weighed your options. That vendor made the shortlist, in part because everyone recognized the name. Nobody would blame you for picking it. It was the safe choice.

Now that safe choice is causing trouble. The invoice came in higher than you expected, the implementation ran for months, and every support ticket starts over with someone who doesn't know your account (and that’s when you hear back from their support team).

If you're wondering how the safe option became the most stressful one, you wouldn’t be the first. Until you start using a tool in a given software category, it’s impossible to know exactly what makes one solution better or worse than the alternatives. During evaluations, all we can do is act on the information available to us at the time.

But after? Why does it feel so hard to leave a solution you know is causing problems?

Psychologists call this phenomenon the sunk-cost fallacy. And fortunately for software buyers, psychologists have done more than name it: they’ve studied how to escape it.

Why Leaving Feels Harder Than It Should

The sunk-cost fallacy is the habit of continuing to invest in something because of what you've already put into it, not because of what it will return in the future. In one classic 1985 study, people who paid more for a theater subscription went to more plays, even though they didn’t start out with a higher desire to attend more plays than the group that paid less for the same subscription.

Several factors make it stronger in the workplace.

One is self-justification. In a well-known business simulation, people who were personally responsible for an investment that went badly put the most additional money into it. Researchers call this escalation of commitment. For the person who championed a software solution, admitting it isn't working can feel like it casts their judgement into question. Ironically, admitting a bad investment isn’t working out is a demonstration of good judgement, even if it doesn’t feel that way.

Another factor is loss aversion. Walking away makes a loss feel final, even though the money is gone either way. The 1985 theater study also pointed out that people don't want to appear wasteful to others.

Inside a company, all of this is harder to see.

The investment isn't one payment. It's budget, training hours, engineering time, and the reputation of whoever championed the tool. Everyone carries a slice of the total cost. Many people may have weighed in on the evaluation too, so nobody has a holistic perspective on whether the software has been worth the investment.

Vendors Have Little Reason to Reduce Your Sunk Costs

In software, the way vendors earn money and buyers’ psychological shortcuts are often aligned. Most vendors aren’t cynically looking to exploit cognitive fallacies. They just have little reason to counter them.

Economists have studied the dynamics of switching vendors when the sunk-cost effect is in play. When switching is costly, companies tend to choose to harvest profit from customers who are already locked in instead of competing for more market share. The harder it is to leave, the less pressure a vendor feels to earn your renewal.

The costs that lock you in go beyond the invoice. Researchers who study software adoption count the time and effort already spent learning a tool, and the skills built around it, as sunk costs. In MIS Quarterly's best paper of 2012, students deciding whether to move from email to Google Docs were less open to the new tool when they felt more invested in the old one and used it out of habit.

Now think about what a complicated implementation does to that tally. Every month of setup, every workflow rebuilt, and every person trained adds to the effort already sunk. By the MIS Quarterly researchers' logic, more sunk effort means more switching inertia.

Not every vendor is going to capitalize on the psychology of sunk costs, but you shouldn’t expect the ones who do to change. A vendor selling a complex software solution has very little reason to make it less complex.

How to Step Back: What an Airport Can Teach Us About Sunk Costs in Software

Fortunately, getting out of a bad commitment is a studied problem too. Researchers Ramiro Montealegre and Mark Keil examined an (in)famous case: the baggage system at Denver International Airport.

In the early 1990s, Denver was building an airport twice the size of Manhattan, and its planners decided to move every bag on one automated system. The system wasn't ready for the October 1993 opening, and was in fact delayed four times. The airport finally opened in February 1995, 16 months late and close to $2 billion over budget.

Not ideal, right?

Life rarely is, and business is no exception. The Denver airport’s baggage system was a failure, but the airport’s switch from a failing commitment to a system that worked was a success.

The researchers found that Denver didn't back out in one decision. It moved through four phases that can be generalized to breaking out of other failing commitments (like a software that isn’t delivering what it should).

1. Recognizing the Problem

What the Denver case study showed: The Denver airport contractor had said from the start that it would need at least another year, and another airport's technical advisers gave the city the same message, but no one wanted to hear it. Nothing changed until a public test failed in April 1994.

How it applies to you: The authors advise staying close to the project and questioning how you read feedback, not relying on the view from the executive suite. Ask the people who use your software every day how it's going, and if it isn’t delivering, state so plainly to business stakeholders.

2. Re-examining the Prior Course of Action

What the Denver case study showed: As costs from the delayed airport opening began to rise, Denver’s mayor changed the question from how to finish the automated system to how to open the airport, and he formed a task force of people who'd had no part in the original decision.

How it applies to you: To translate this to a software commitment, add up what staying costs you, then ask what you need the software to do. Hand the review to someone with some distance to the original evaluation.

3. Searching for an Alternative Course of Action

What the Denver case study showed: Denver city officials went looking for a low-tech option that would get the airport open, and the city settled on a conventional tug-and-cart system that could be built in under 5 months.

How it applies to you: The authors suggest starting from the business problem the project was meant to solve. Do the same: start from what your software was supposed to do, and look at what else does it. Researching options says nothing about your earlier choice, just that you’re more informed now than you were then.

4. Implementing an Exit Strategy

What the Denver case study showed: United Airlines, the airport's largest tenant, objected to the new system. It took 3 days of closed-door talks with the city and the contractor to restructure the contract.

How it applies to you: The authors advise asking who will resist switching before you act. List who has a stake in keeping your software, such as its champion and the teams whose workflows depend on it, and bring them in early. Then write down what leaving would take: the timeline and the migration work.

Sound Familiar?

If the software you feel you’ve sunk too much into to leave is OneTrust, you're in a very common spot. It's a good fit for large enterprises with big, dedicated privacy teams, which is who it was built for–but it’s not who OneTrust is sold to. It's a harder fit for teams of one to five, and a lot of people reading this are in that camp.

You picked it because it was supposed to protect you from privacy violations. But regulators have penalized big companies, including Honda and Ford, for running default OneTrust configurations that violated privacy law. Even though these companies relied on their privacy vendor to keep them compliant, OneTrust’s product philosophy is to provide tooling; not compliance support.

It's not just these two companies either. InfoTrust's 2025 audit of 450 US websites found that 79% still load at least one targeting or advertising tag after a user opts out. InfoTrust and Osano audited one such site running OneTrust on a recent webinar, showing why a solution that’s complex to configure and manage, like OneTrust, can hurt compliance outcomes.

The same thing that keeps teams on OneTrust–the sunk cost of a complex implementation, learning to use it, configuring it, and managing it on a daily basis–also directly contributes to higher compliance risk.

You spent months setting up a platform because it was the safe choice, and a banner that looks right can still sit on top of tags that keep firing after someone says no. The more you've put into it, the harder it is to ask whether it's working.

Thinking About Switching?

You wouldn’t be the only one. We’ve talked with hundreds of teams evaluating OneTrust alternatives. In fact, 40% of our customers came to Osano from OneTrust.

Above, we laid out four steps to extricate yourself from the sunk-cost fallacy:

1. Problem recognition
2. Re-examination of the prior course of action
3. Searching for alternatives
4. Implementing an exit strategy

If you’re past steps 1 and 2 already, we can help with steps 3 and 4.

Take a look at our comparison detailing the differences between Osano and OneTrust here. We break down how the two platforms stack up, report on how current customers rate both platforms, and identify the essential criteria users should look for when making their evaluation.

If you’re mainly concerned about a lengthy and complicated migration, we can help quantify the process and sketch out a migration roadmap with our migration calculator. Just input the capabilities you want to port over and provide a few details like numbers of integration, domains, and other relevant variables–we’ll provide an estimate of what your migration path could look like.