Back to Blog
22/11/2026
6 min readBy Angel Anson

How to set a KTLO budget and act at 20%, 40% and 60%

How to set a KTLO budget and act at 20%, 40% and 60%

Years ago, a leader I worked with introduced a concept that changed how I run teams: a budget for KTLO, keeping the lights on. KTLO is all the work that isn’t building features: incidents, maintenance, upgrades, manual operations.

What surprised me was how invisible it had been. Without a number, all of that work feels normal. Fixing incidents is normal. Updating things is normal. Doing manual work is normal. So you never stop and step back to automate. Then you start measuring it, and you realize you have a problem.

Since then, KTLO is one of the first things I measure in every team. I’ve written about how I measure it. This post is about what comes next: what number to aim for, and what to do when your team goes over it.

Why 20% is the baseline

My baseline is 20% KTLO, and part of the reason is psychological.

Nobody enjoys KTLO, especially the reactive kind. Incidents and manual work are stressful. Feature work is different. It moves the team closer to its goals, lets engineers see their impact on users and on the business, and gives them a sense of purpose. Reactive KTLO keeps them in survival mode.

At 80% features and 20% KTLO, the team is in a healthy place. When KTLO grows, it eats into the 80% that keeps people motivated.

The opposite direction is also a trap. Pushing KTLO below 20% is very expensive, and the return on that investment is low. Getting to 10% or 5% is sometimes impossible. There will always be an incident. A rule will change and make your systems non-compliant again. There will always be an edge case you haven’t automated yet. Your systems need to keep scaling and evolving, and that work never ends.

So I treat 15% to 25% as the healthy range, with 20% as the baseline.

Why KTLO snowballs

The real danger of KTLO is that it compounds. You measure and you’re at 20%. Next time, 25%. You do nothing. The next measurement isn’t 30%. It’s more like 32%, then 40%, then 50%, then 70%, and by then you’re in serious trouble.

It snowballs because KTLO feeds itself. Incidents interrupt the work. Patched systems make every new feature harder to build. Less time for features means more pressure, and more pressure means more shortcuts.

How a team ends up there

I saw this with one of my teams. We were launching a lot of features that year, and we consciously made trade-offs to be fast in the market. We did manual work instead of fully automating some parts. We knew the platform might not scale well later. We traded quality for speed.

That was a valid decision. The problem is that a shortcut works like a loan, and we never paid it back. We kept launching features with the same aggressive deadlines. But incidents kept interrupting us, manual work kept growing, and development got harder because we were patching over a system that didn’t scale. To keep the pace, we took even bigger shortcuts, which made the long-term damage worse.

At some point it became unsustainable. That’s how a team goes from 20% to 80% in the blink of an eye.

We had to do a hard stop: several weeks without launching anything, just paying back the debt we had accumulated. And like any loan, the longer you wait, the more interest you pay. Something that would have taken one week to fix or automate at the beginning can take three weeks later, because by then the code is so buggy that you have to rebuild entire parts of it.

Three levels, three roadmaps

To stop the snowball early, I use three levels. Each one tells me how to split the roadmap between features and strategic KTLO, meaning the work that reduces recurring KTLO, such as automating a manual operation or fixing the root cause of repeated incidents.

  • 20% KTLO: healthy. The roadmap is 100% features.

  • 40% KTLO: act now. The roadmap is 50% features and 50% tech debt, automations, etc. If you don’t do something here, you’re heading for big trouble.

  • 60% KTLO: in real trouble. The roadmap is 100% tech debt, automations, etc. until you’re back down.

These are guidance, not strict rules. Between the levels it’s a gradient: the higher the KTLO, the more of the roadmap goes to bringing it down.

Some teams use a fixed split instead, like 70% product and 30% technical debt every sprint. I don’t like that. A fixed split ignores where the team actually is. I prefer to let the KTLO level decide how much we invest on each side.

When product says no

This model is also a very easy way to negotiate with product. But what if you’re at 40%, you propose a 50-50 roadmap, and they tell you the deadline is fixed?

In the end, it’s a negotiation, and it depends on your relationship. Do you work for product, or are you peers who make decisions together? Either way, there are a few things I do.

First, raise the risks. Be very clear about the snowball and about the consequences of adding more shortcuts on top of the ones you already have.

Second, be flexible about when you pay the debt back. It doesn’t have to start tomorrow. You can agree to launch what needs to be launched over the next two months, and then make the third month focused on KTLO to bring it back to 20%.

Third, use numbers. If your team has 10 engineers and you’re at 40% KTLO, four of them are doing nothing but KTLO. In a month it could be six, and a few weeks later eight. At that point, you can’t launch anything. That’s the argument to put on the table.

I’ve never had a problem negotiating this with product. When the numbers are clear, the conversation is easy.

Does this apply to platform and SRE teams?

It depends on the company. In some companies, the SRE team acts as the global firefighters for everyone else. I prefer the model where SRE and platform teams build the platform that product teams build on top of. In that model, the product engineers are their customers, and the same levels apply.

The takeaway

Measure your KTLO, treat 20% as the baseline, and watch the trend, not only the number. KTLO doesn’t grow in a straight line. Let the level decide your roadmap: all features at 20%, half and half at 40%, all strategic KTLO at 60%. And when you negotiate with product, talk in engineers, not percentages.

About the Author

Angel Anson

Angel Anson

Angel Anson has spent over 20 years in the tech industry, working across startups, scaleups, and large organizations such as King (Candy Crush), N26, Scytl, RentTopApartments, and Emte Sistemas. Throughout his career, he's held roles like Software Engineer, Tech Lead, Engineering Manager, Head of Engineering, Tech Director and Director of Engineering, and has also founded two companies — one as a co-founder and one independently. His experience spans fintech, gaming, government technology, and enterprise software, where he's led teams building products used by millions of people worldwide. Today, he serves as Director of Engineering at N26, leading more than 100 people, and is the founder of Engineering Manager School, an initiative that helps professionals improve their management and soft skills and lead with clarity and confidence. Over the years, Angel has mentored more than 100 professionals in tech and product roles, helping them grow into stronger, more intentional leaders.