Killing a Rebuild Project Before It Killed the Company

$2M spent, 9 months in, 30% complete. The hardest decision: stop and redirect.

[Overview] System Overview

The company had bet on a complete platform rebuild. New technology stack. New architecture. 15-month timeline.

Nine months in, the project was 30% complete and accelerating slower than planned. Runway was 14 months.

[Context] Situation before

The existing platform worked but felt dated. Customer complaints about UX. Engineering complaints about code quality.

The rebuild was positioned as solving all problems.

[Diagnosis] The real problem

Rebuild timelines are always underestimated. 30% at 9 months meant completion at 27+ months. With 14 months runway, the company would run out of money.

Understanding the root cause changed everything.

[Boundaries] Working constraints

COST

  • $2M already invested (sunk cost)
  • 14 months runway remaining
  • No additional funding secured

Constraints are features, not obstacles. They guide design decisions.

[Decisions] Key tradeoffs

REJECTED

Continue rebuild

Push through to completion

Math does not work. Company dies before completion.

CHOSEN

Kill rebuild, pivot to existing system

Accept sunk cost, redirect engineering

Better to waste $2M than $5M. Survival first.

[Design] Architecture design

Returned engineers to existing platform. Applied learnings from rebuild to incremental improvements. Shipped 8 features in 3 months.

The Math of Killing a Rebuild

$2M spent, 9 months in, runway of 14 months

killing-rebuild-before-disaster-architecture

[Approach] Execution model

##### Transition

Rebuild team transitioned over 2 weeks. Project lead moved to other opportunities. No blame assigned.

[Results] Measurable outcomes

##### Result

Runway extended to 20+ months. Revenue growth resumed. Series B raised 18 months later.

Key improvements achieved:

  • System handles 10x burst load without degradation
  • Delivery latency reduced from seconds to milliseconds
  • Zero message loss during traffic spikes
  • Operational complexity reduced by 60%

[Anti-patterns] What could have gone wrong

##### If Continued

Company would have run out of money. All jobs lost. All investment lost.

  • Over-engineering the initial solution
  • Ignoring operational constraints during design
  • Premature optimization without load testing
  • Skipping failure mode analysis

[Applicability] Where this pattern applies

##### When to Kill

When completion timeline exceeds runway. When sunk cost is less than completion cost. When existing system can be improved faster.

Every system is different. Context always matters.