Software's Inevitable Crossroads: Traditional Maintenance Mode Solutions Fail to Deliver, Survey Reveals

Software systems inevitably reach a ‘maintenance mode’ when demand for new features declines, and maintenance tasks like defect fixes, vulnerability updates, and platform upgrades outpace value-adding work. This phase, often termed ‘keeping the lights on’ or BAU, signifies a shift in funding away from new product features rather than an absence of user needs or defects. Despite its inevitability and maturity, traditional approaches to maintenance mode are proving fundamentally flawed, failing to meet desired outcomes such as increased capacity, reduced costs, sustained service reliability, protected future feature delivery, and job satisfaction.

Two primary traditional solutions persist, yet both exhibit significant drawbacks. The first involves fully transitioning services to an operations team post-delivery. While this can increase delivery team capacity and reduce costs, it often weakens live service reliability, as operations analysts may lack the domain knowledge or technical skills for numerous services, leading to slower, lower-standard task completion. Future feature delivery becomes arduous due to painful ‘reverse transitions’ and skill gaps, and job satisfaction plummets for both delivery teams (feeling like a ‘feature factory’) and operations (a ‘dumping ground’). A survey of 40 enterprise organizations by Equal Experts revealed that 55% experienced a drop in service reliability, 55% reported slower feature delivery, and a stark 88% observed a negative impact on job satisfaction with this model. The second common approach sees delivery teams ‘backgrounding’ maintenance, allocating minimal time (e.g., an hour a week) to older services while developing new ones. This can preserve reliability due to retained domain knowledge but fails on other fronts. It creates difficult prioritization conflicts for future feature delivery, leads to low job satisfaction due to context switching and repetitive work on old tech stacks, and, critically, does not reduce costs, as teams cannot be easily resized or retired. The Equal Experts survey confirmed this, with 100% of respondents reporting no cost reduction, 52% noting slower feature delivery, and 58% experiencing worsened job satisfaction, often due to ‘drowning in background work’ or ‘slowing down work on the new service to fix up the old service’ impacting morale. These findings underscore the urgent need for more radical maintenance mode solutions that genuinely deliver on all desired outcomes.