Skip to content
Thomas Stockham
Back to writing

Maintenance Mode Is a Decision You Didn't Make

productscultureopinion

I've watched a lot of layoffs over the last few years, some up close and some from a distance. One pattern keeps showing up. Companies cut the people and keep the product.

A team gets laid off. The PM, the engineers, and the designer who knew why every screen worked the way it did. The part of the product they owned stays exactly where it was. It gets a new label: maintenance mode.

What maintenance mode means

Maintenance mode sounds like a plan. It means nobody owns the area and nobody improves it. Customers still use it. Support still fields tickets about it. Other teams still build around it. The cost stays after you stop paying the people who understood it.

Here's my hot take. If a role isn't valuable enough to keep, about 90% of the time the product area isn't valuable enough to keep either.

That number is a gut call from years of working in product. I think most people who have inherited an orphaned feature would land in the same place.

Maintenance mode racks up three kinds of debt, and only one of them shows up on an engineering roadmap.

Tech debt

The first is tech debt. Code that nobody touches still ages. Dependencies drift out of date. Security patches pile up. A platform migration that should take a quarter takes two, because one corner of the codebase has no one who understands it. The people who could have explained it are gone. The remaining teams tiptoe around it, and every change near it gets slower.

Experience debt

The second is experience debt. Customers feel this one. The rest of the product keeps evolving with new design patterns and new navigation. The maintenance mode area stays frozen. A customer clicks from a modern screen into one that looks like it came from a different company. Behaviors don't match. Settings live in different places. Bugs get triaged and then sit, because no team exists to fix them. Nobody announced that this part of the product was abandoned. Customers can tell anyway.

Decision debt

The third is decision debt, and I think it costs the most. Every future decision has to route around the thing you kept. Want to change the pricing model? Somebody has to work out what happens to the orphaned feature. Redesigning the data model? It has to keep working for that one area. Launching something new that overlaps with it? Now you're debating whether to integrate, replace, or ignore it, and nobody in the room owns the answer. This tax shows up in every planning meeting, and nobody puts it on a slide.

Why companies cut the people and keep the product

So why do companies cut the people and keep the product? Cutting people feels like a cost decision. Cutting a product feels like a strategy decision. Cost decisions are easy to make in a spreadsheet. Strategy decisions mean admitting a bet didn't pay off, telling customers something they don't want to hear, and doing the work to wind something down. Maintenance mode lets you skip all of that. Nobody has to say no out loud.

The 10% that belongs in maintenance mode

Some cases do belong in the other 10%.

Compliance and regulatory features have to stay. If the law says you need it, you need it. Contractual commitments work the same way. If you promised a customer something in writing, you can't turn it off because the team that built it is gone.

Some product areas are done. They're stable, they make money, and they need very little ongoing work. A mature feature that customers rely on and that would cost more to remove than to keep is a fair case for a skeleton crew.

Sometimes the area was overstaffed. If five people were doing a job that two could do, cutting three of them is a staffing correction.

I'd push on all of these. "It's stable and needs little work" is the easiest argument to tell yourself, and it's the one I trust least. Some things look stable until the dependency underneath gets deprecated or a big customer finds an edge case. If you claim an area is in the 10%, say why in one sentence. A named person should still own it. Without an owner, the area is in decline and the label hides it.

What to do before the next layoff

Here's the question I wish more leaders asked before a reduction in force: if we're cutting the people, are we willing to cut the thing?

If the answer is yes, plan the sunset. Tell customers early and plainly. Give them a timeline that fits how they use the product, with a migration path if one exists. Then remove it. Delete the code, kill the settings page, and take it out of the docs. A "deprecated" feature that still sits in the product is maintenance mode with extra steps.

If the answer is no, name one person as the owner. Give them time to keep the area healthy, and revisit the decision on a schedule. Keeping something on purpose is fine. Keeping it by default leaves you with a product that's half alive.