Stopping a Failing Project Meant Writing Off Two Years of Our Work

In the early 2010s, a team embarked on creating a customer portal for contract warehousing clients. As is typical with software development, the expenditures were capitalized by the finance department, meaning they were allocated over the expected lifespan of the software rather than being expensed in the year of creation. By the second year, the portal was found to be underperforming as clients began integrating directly with the company’s systems, rather than using the portal.

This led to a discussion at a steering meeting, during which the team proposed shutting down the project. However, the finance director raised a critical query – what would happen to the £1.4 million of capitalized costs? If the project were terminated, these costs would need to be written off, all at once, impacting profits in the same year that the CEO would present the results to a new owner.

Consequently, the project was not halted but was instead restructured, renamed, and continued for an additional eleven months, at an extra cost of £600,000. Ultimately, the portal was written off at a total cost of £2 million, despite the fact that no one was dishonest or intentionally misleading. The accounting methodology used was correct.

What remained unseen was the impact of capitalization on the decision to cease a project. Every month the project remained active, the financial cost of abandoning it increased, even if continuing the project was more expensive in cash terms. The team subsequently revised their approach, setting stop criteria for each project in collaboration with finance, such as the required usage levels in pilot tests.

Costs are now capitalized only after a project has passed a specific gate, proving its utility. Finance now reports the capitalized balance of all active projects to the board each quarter, alongside the project’s progress, making a write-off a matter of expected knowledge rather than a surprising revelation. Since implementing these changes, two projects were terminated early, both costing less than £200,000.

It became evident that a project that is inexpensive to initiate can turn out to be costly to terminate. The team had not previously considered how their own financial records could drive such outcomes.