Skip to main content
Broadpin

The Numbers Trap: Why Teams Underestimate What They Built

This blog was written by Stephan La Rocca (Director of Business Development).

Why do modernization projects so often blow past their time and budget targets, even when the team knows the legacy system inside out? This three-part series explores the cognitive biases behind that pattern. After Part 1 covered how teams start wrong, Part 2 turns to estimation: why a system that already works makes the job of rebuilding it look deceptively small. For all biases and the methodological response to each, download our whitepaper Software Modernization and Its Cognitive Biases.

Download the Whitepaper

Part 1: Why Modernization Projects Fail Before the First Line of Code

Scale Insensitivity: "Ten minutes per item" times four thousand

In a recent project, the person responsible for UI data binding told me: "I only need ten minutes to rewrite this event. I would rather do it manually."

A fair point, in isolation. Unfortunately, the legacy application contained more than 4,000 places that needed the same change. At ten minutes each, that person would have spent over half a year doing nothing but copying existing code. No errors, no tests, no other tasks.

This bias is unfamiliar to teams that have mainly worked on new development. In agile delivery, work arrives sprint by sprint, and even a large backlog feels manageable because refinement breaks it into digestible pieces. Modernization delivers its work differently. If the legacy application has not been properly inventoried, then suddenly millions of lines of code, hundreds of thousands of UI elements, and thousands of table APIs land on the board at once. Every individual decision multiplies with the full complexity of what already exists.

The Planning Fallacy: "The system already works, we just have to rebuild it"

In estimates for modernizing legacy applications, stumbling blocks, risks, and necessary iterations are consistently underestimated more than in a complete new build. The reason is counterintuitive: precisely because the existing system already works, it creates the impression that the new solution merely has to be reconstructed.

The original wrong decisions, rework cycles, and late-night debugging sessions that went into making it work are invisible. What remains is a running system, and running systems look easy to replicate.

Overconfidence: "We will manage somehow"

This bias often has organizational rather than personal causes. Especially with in-house applications, the business department's expectations of IT grow while the budget does not allow for external expertise. In that situation, IT tends to accept the modernization assignment and respond with "we will manage," even though risks, dependencies, and missing specialist knowledge have not been sufficiently understood.

When overconfidence meets scale insensitivity meets the planning fallacy, the result is a project plan that looks achievable on paper but has no realistic foundation underneath it.

The structural answer

The common thread across all three is the same: decisions are based on what the team believes the system contains, not on what it actually contains. The fix is not more optimism or more caution. It is data.

An automated inventory of the legacy application, its code volume, its UI elements, its dependencies, its implicit business logic, replaces gut feeling with a factual baseline. Estimation stops being a guessing exercise and starts being an engineering task.

In the final post, I will look at the biases that distort how teams remember the past and learn from others, and at how Broadpin’s modernization approach is built to counter all of these patterns.

Replace gut feeling with facts.

Our whitepaper Software Modernization and Its Cognitive Biases explains why teams consistently underestimate legacy systems and how an automated inventory turns estimation into an engineering task instead of a guessing exercise.

Download the Whitepaper