Kagi Stopped Building Orion for Windows. Narrowing a Bet Is a Strategy, Not a Retreat.

Two software companies pulled back from a market and a beloved product in October. How to tell a deliberate narrowing from a quiet stall.

Matthew Stublefield

6 min read

Person walking past a red building with arched windows

On October 2, Kagi announced it was ending development of its Orion browser for Linux and Windows. It's open-sourcing both, handing them to the community, and putting its handful of developers back on the Mac and iPhone versions. The Windows version had been planned for release this year. It won't come from Kagi.

The reasoning in Kagi's announcement is short and unusually candid. "When we decided to take Orion beyond the Apple ecosystem, we knew it was a big bet for a small, independent company." And then: "One thing we hear often from members is that they want us focused. They're right."

Three weeks later, The Browser Company announced that its Arc browser will stop working on December 22. Different company, different situation, same month, and between them they make a useful pair for anyone running a product company. Pulling back from a market or a product is a strategic decision, and like any strategic decision it can be made well or badly. What most teams get wrong is when they make it, and whether they make it at all.

Two different kinds of pulling back

Kagi's decision is a narrowing. It expanded into new platforms, decided a small team spread across three of them was spread too thin, and chose to focus on macOS and iOS. The open-sourcing gives the work a future without Kagi paying for it. The message to customers is specific: what's ending, what's continuing, and why.

Arc's story is a longer one. TechCrunch reported that The Browser Company stopped active feature development on Arc in 2025, after conceding that while early adopters loved the browser's novel interface, the "novelty tax" and its complexity limited mainstream adoption. Arc kept running for another year while the company built Dia, its AI-first browser. This October it set an end date, citing the cost of keeping a browser secure. Atlassian acquired The Browser Company for $610 million last year, and Arc's users are being pointed to Dia.

Both decisions narrow the company's bets. The difference is that Kagi made its call before the Windows version ever shipped, while Arc's played out over a year of maintenance after feature work had already stopped.

Beloved by early adopters isn't the same as a market

Arc's story carries a lesson that's easy to miss when you have passionate users. Enthusiasm from early adopters is real, and it feels like product-market fit. The Browser Company's own phrase for what stopped Arc from going further, a "novelty tax," is a precise description of the gap. Early adopters will pay that tax gladly. Most people won't, and a product that asks everyone to learn a new way of working is asking for exactly that.

The same thing happens in B2B software. A feature or product line wins over a devoted group of power users, the team reads their enthusiasm as demand, and the product keeps getting investment long after it's clear the broader market isn't following. The power users are right that it's good. They're just not a market by themselves. A useful test is whether the enthusiasm is spreading. If the people who love it are bringing in colleagues who come to love it too, a market may be forming. If the same devoted group keeps asking for more while new users bounce off, it isn't.

Most moves outside the core don't pay off

Kagi's choice also fits what the long-running research on growth says. Bain's work on adjacency found that only about one in four moves into a new market, segment, or product produced profitable growth. That doesn't mean expansion is wrong. It means the odds are against any given move, and the companies that do well are the ones that notice early when a move isn't working and pull back without drama.

That's harder than it sounds, because every expansion creates people inside the company whose work, and sometimes whose jobs, depend on it continuing. Saying "this isn't working, we're going back to what we're best at" requires someone with the standing to say it and the willingness to absorb the disappointment that follows.

Why most teams stop too late

I know the late version of this from the inside. At Adaptavist, I built an e-learning product, the line that eventually became Learn for Jira. Before we started, I wrote down eight risks to it. One was that the company that made the underlying platform would start investing in its own training and could outspend us without noticing. By the time the product was shut down, all eight risks had materialized, and I wrote about why naming them wasn't enough.

The risks weren't hidden. The document existed and it was accurate. What was missing was a moment where someone decided, on purpose, that the evidence was in and the bet should narrow. Without that moment, a product doesn't stop. It drifts, with less investment each quarter, until the decision gets made for you.

The people who pay for that drift are the customers. Months after I'd left Adaptavist, the CEO of one of our customers messaged me on LinkedIn: "Hey, what happened? We were relying on it." I got on a Zoom with him and walked him through the whole thing. A product that drifts to a stop leaves its customers finding out after the fact. A deliberate pull-back tells them first.

That's the difference between Kagi's announcement and a quiet stall. A deliberate narrowing has a date, a reason, a plan for the people affected, and a clear statement of what the company is doing instead. A stall has none of those. It just has a line item that keeps shrinking.

Questions to ask before you pull back

If you're looking at a market, a platform, or a product line that isn't working the way you hoped, a few questions help separate a narrowing from a retreat.

What will the narrowed bet buy? Kagi's answer was focus for a handful of developers on the platforms it knows best. If you can't name what you'll do better with the freed-up people and money, you're cutting cost, not narrowing a strategy.

Is the enthusiasm coming from a market or from a group of early adopters? Look at who's adopting and whether the group is widening. A devoted base that isn't growing is a signal, however loud it is.

How long ago did the evidence arrive? If you can point to the month the answer became clear, count how long it's been. Most of the cost of a late stop is in that gap.

What happens to the customers you're leaving? Kagi open-sourced its code. Arc set a date and offered a path to Dia. A pull-back done well treats the people who relied on you as the reason to do it carefully.

Who has the standing to make the call? Clear evidence doesn't stop a product on its own. Someone with authority has to decide it should, and say so.

Strategy is a filter, and a filter is defined as much by what it removes as by what it lets through. Kagi used theirs in public, with a clear reason and a date. Narrowing a bet like that isn't a retreat, and it's a decision most companies make far later than they should.

–Matthew

More Field Notes

  1. Consulting Operations

    '300 Million People Is a Big Market on Paper.' Envision's Shutdown Is a Lesson in Market Sizing.

    A market sized by everyone with a problem tells you little about a business. Envision's founders said so about their own 300-million-person market.

  2. Consulting Operations

    Two-Thirds of Software Growth Now Comes From Existing Customers. Retention Just Had Its Worst Year.

    Software companies lean on existing customers for most of their growth just as retention weakens. What that means before you bet on a new market.

  3. Consulting Operations

    92% Say Their Research Works. 85% Have Caught It Being Wrong.

    Most research professionals trust their research, and most have caught it being wrong. Four checks to run before you build on a finding.

Want help running a sharper practice?

The reading and synthesis behind your client work, handled – a living deliverable kept current, so more of your time goes where your name is actually on the line.