Keep the Lights On (KTLO)
What is Keep the Lights On (KTLO)?
Keep the lights on (KTLO) is the work and spending needed to run and maintain existing systems and services, such as support, maintenance, patching and licences, as opposed to building new capabilities.
Updated November 2026 · 2 min read
What counts as KTLO
KTLO covers anything needed to keep today's products and systems working at their current level. It does not make anything new; it stops things from getting worse. The term is most common in IT and engineering, where leaders track how much of the budget or team capacity it consumes.
| Category | Examples |
| Operations and support | Incident response, on-call, help desk, monitoring |
| Maintenance | Bug fixes, security patches, upgrades to stay on supported versions |
| Licences and contracts | Software renewals, hosting, support agreements |
| Compliance | Audits, certificate renewals, required regulatory changes |
| Technical upkeep | Dependency updates, small refactors to keep code workable |
KTLO vs growth and transformation
Many organisations split technology spending into three buckets, often called run, grow and transform. KTLO is the "run" bucket. Growth covers improvements to existing products; transformation covers new capabilities and business models. The split matters because KTLO tends to grow on its own. Every new system adds to the run cost once it goes live, so without active management, the share left for new work shrinks year by year.
| Bucket | Purpose | Typical question |
| Run (KTLO) | Keep current services working | What is the minimum we need to spend? |
| Grow | Improve and extend what exists | Which improvements give the best return? |
| Transform | Build new capabilities | Which bets move the strategy forward? |
An example
A software company finds that 68% of engineering capacity goes to KTLO: support tickets, an aging billing system and four databases on unsupported versions. That leaves less than a third of capacity for the roadmap. Leadership sets a target to reduce KTLO to 50% over 18 months. It retires two little-used products, migrates the databases, and puts the billing rebuild into the portfolio as a funded initiative. A year later, KTLO is at 55% and roadmap capacity has grown by almost half, without adding headcount.
Managing KTLO
- Measure it: tag work and spend as run, grow or transform so the share is visible
- Set a target share in capacity planning, and protect it
- Retire or consolidate systems; every system removed reduces KTLO permanently
- Invest in automation and in paying down technical debt that drives support load
- Include future run costs in every business case, not just build costs
Common pitfalls
Treating KTLO as unavoidable overhead is the most common mistake. Much of it can be reduced, but only with deliberate investment. The opposite mistake is starving KTLO to fund new projects, which leads to outages, security risk and slower delivery later. Many teams also hide KTLO inside project work, so leaders never see its true size. Clear reporting, often through technology business management practices or resource allocation reviews, makes trade-offs visible. See also CapEx vs OpEx.
Frequently asked questions
What percentage of IT budget goes to keeping the lights on?
Estimates vary by industry and company age, but surveys commonly find that well over half of IT budgets go to running and maintaining existing systems, and figures of 60 to 80% are often quoted. Older organisations with complex legacy estates tend to sit at the higher end, leaving less room for new investment.
Is KTLO the same as technical debt?
No, but they are related. KTLO is the ongoing work to keep systems running. Technical debt is the accumulated cost of past shortcuts, which often increases KTLO by making systems harder to maintain and more likely to fail. Paying down technical debt is one of the main ways to reduce KTLO over time.
How do you reduce KTLO?
Retire or consolidate systems that are little used, automate repetitive operations work, pay down technical debt that drives incidents, move to managed services where they cost less to run, and include run costs in investment decisions. Reductions usually need upfront investment, so they should be planned and funded like any other initiative.
Related terms
Capacity Planning
Matching planned work to the people and time actually available.
Resource Allocation
Assigning people and budget to the work that matters most.
Kill Criteria
Conditions agreed up front that tell you when to stop.
Demand Management
Capturing and shaping new work requests so only the best ones are committed.