What Is SaaS Sunsetting? The CoScreen Shutdown Explained

What is SaaS sunsetting, and what does the CoScreen shutdown show?

SaaS sunsetting means the tool your team depends on every day can vanish the moment its parent company decides it no longer fits a bigger roadmap. You get little say in the timing, the notice, or what happens to your data. CoScreen is the clearest recent proof of that risk, and the details are worth knowing before it happens to your own stack.

What happened the day CoScreen disappeared?

In February 2022, Datadog acquired CoScreen, a real-time screen-sharing and pair-programming tool built for incident response and technical collaboration. The acquisition announcement promised continuity: existing users could keep collaborating on CoScreen while it folded into Datadog's broader platform. For a while, that held.

Then, on July 31, Datadog sunset the standalone CoScreen service entirely. Roughly four years after the acquisition, the workflows product teams had built around it simply stopped existing. No new owner picked it up. No in-house Datadog feature replaced it for the teams who relied on it as a dedicated tool. It was just gone.

Is the CoScreen shutdown a one-off?

What makes CoScreen's shutdown worth writing about isn't that it was unusual. It's that it followed a pattern collaboration-tool veterans could see coming. CoScreen co-founder Jason Thomas said as much when he announced the product's open-source successor, laying out the same lifecycle repeating across the category:

• Screenhero, acquired by Slack in 2015, sunset as a standalone app on December 1, 2017, with its technology folded into Slack Calls.
• Multi (formerly Remotion), acqui-hired by OpenAI in June 2024, shut down about a month later, on July 24, 2024, with user data deleted after that date.
• Around, acquired by Miro in October 2022, shut down fully on March 31, 2025, with user recordings becoming permanently inaccessible after closure.
• CoScreen, acquired by Datadog in February 2022, sunset on July 31, roughly four years later.

Four companies, four acquirers, one ending: buy the tool, absorb what's useful, let the standalone product die.

Why do platforms buy and kill collaboration tools?

This isn't malice. It's incentive structure. A larger platform acquires a collaboration tool to buy a team and a piece of technology that strengthens its core product, not to run a separate product line indefinitely. Thomas's own framing of the CoScreen shutdown makes this explicit: the standalone tool became tangential to Datadog's core mission, even after years of investment in it.

That's the quiet part of every acquihire press release. The acquirer's roadmap, headcount priorities, and product strategy can shift years after the deal closes, and the standalone tool that justified the acquisition becomes the first thing cut once it stops aligning with wherever the platform is headed next. Customers who built workflows around that product never had a vote. They just inherited the consequence.

How did the Petal response rebuild what the team never owned?

After the shutdown, four members of the original CoScreen team spent months rebuilding the product from scratch and released it as Petal, a free, open-source pair-programming and screen-sharing tool. Thomas was direct about why: "If we lose interest or get hit by a bus, you can self-host it or fork it, and nobody can sunset it on you."

That sentence is the whole argument in miniature. The team that built CoScreen had to spend real engineering months recreating functionality customers had already paid for once, because the only copy of it lived inside a company that decided to stop running it. Nobody writing a CoScreen subscription check in 2022 budgeted for paying that cost again in time, migration pain, and lost continuity. That's the hidden line item SaaS sunsetting adds to a tool's real lifetime cost, and it never shows up on an invoice.

What Does Sunsetting Actually Cost a Team?

CoScreen's story has a clean before-and-after. Most sunsets are messier, and the data on how messy is worth knowing before it happens to you.

A 2026 cross-provider study of ten documented SaaS and cloud shutdown events between 2015 and 2026 found a median shutdown notice period of 311 days for cases with a clearly announced, planned sunset. That sounds like reasonable warning until you look at the outlier: Firefox Send was pulled offline before its shutdown was even formally confirmed, giving users effectively negative notice. Notice periods are a courtesy, not a guarantee.

Even with generous notice, the same study found the real cost of exiting a tool isn't the data export. It's what the researchers call the portability gap: the distance between what a vendor lets you download and what you can actually restore operationally, including workflow logic, identity and access configuration, integrations, and managed-service behavior that doesn't travel in a CSV file. "A long shutdown notice period does not prove recovery capability," the report notes.

And the conditions that precede a sunset can arrive with no warning at all. Miro, the visual-collaboration company that acquired and later shut down Around, was itself valued at $17.5 billion in a January 2022 funding round. In September 2026, Bending Spoons acquired Miro for $1.355 billion, a markdown of more than 90 percent from peak. A valuation collapse of that size is exactly the kind of strategic shock that precedes product consolidation and sunsetting, as Around's own fate under Miro's roof already demonstrated.

Why is vendor lock-in the mechanism and sunsetting the event?

Sunsetting feels sudden, but the exposure to it builds slowly. Vendor lock-in is the dependency that accumulates as a team's workflow logic, data model, and institutional habits get embedded inside a rented tool. The deeper that embedding goes, the higher both the switching cost and the sunset exposure climb, because more of your actual working process is sitting inside someone else's product.

As one framing puts it, a SaaS subscription is a rental, not a purchase: you rent access to the tool, the data model, and the workflow logic built on top of it, and you lose all three the moment the vendor stops the service, whether by choice, acquisition, or shutdown. CoScreen customers weren't just renting a screen-sharing window. They were renting the incident-response and pairing habits their engineering teams had built around it. When the service went away, those habits had nowhere to run.

The mechanics are identical to the lock-in no-code AI builders quietly sell for $20 a seat: convenience up front, dependency that compounds, and a decision about whether your workflow keeps running that sits with someone else's roadmap. It's also the same exposure behind why renting platform access through someone else's API leaves a business one legal letter away from zero: the acquirer or platform owner, not the customer, holds the decision rights over whether the dependency keeps working.

What's the Alternative to Renting Mission-Critical SaaS?

The honest fix isn't a better vendor-selection checklist. It's recognizing that no amount of due diligence protects you from a tool's acquirer changing its mind years later. Self-hosted and open-source tools can't be unilaterally sunset by someone else's M&A decision, because there's no single company holding the kill switch. That's the same lesson running through Cricut's remote-locked cutters and Google's quietly disappearing Git tags: if you don't control the software layer, you don't really own what's running on top of it. Petal's own design choice, self-hostable and forkable from day one, is the structural answer to the exact failure mode CoScreen's customers just lived through.

This is also why the local-first, open-source movement in developer tooling has gained real traction beyond nostalgia: teams are rediscovering that owning the code, even imperfectly, beats renting a polished product someone else can discontinue on a quarterly earnings call.

What Should Teams Do Before Their Tool Gets Sunset?

You cannot predict which acquihire will end in a shutdown notice. You can audit your exposure before it happens.

1. Inventory which tools are "tangential features" inside a bigger acquirer. If a tool you depend on was bought mainly for its team or its tech rather than as a durable standalone product, treat it as higher sunset risk, regardless of how good it currently is.
2. Demand data and export rights now, not after a shutdown notice. Ask what's actually portable: raw data, yes, but also workflow configuration, integrations, and access logic. The CoScreen study's "portability gap" is the right lens here.
3. Separate core-workflow tools from commodity tools. A tool that holds your team's actual working process, pairing, incident response, documentation of record, deserves more scrutiny than a tool doing something interchangeable.
4. Weigh owning versus renting for anything in the first category. If a tool is central enough that losing it would set your team back months, as it did for the engineers who had to rebuild CoScreen as Petal, that is the signal to build or self-host rather than subscribe.
5. Watch the acquirer's financial health, not just the product's. Valuation collapses and ownership changes, like the one that reshaped Miro, tend to precede the product decisions that affect you.

None of this guarantees safety. It does shift the odds. CoScreen's engineers didn't get a vote when Datadog decided the product was tangential to its mission. The only reason their customers have a path forward at all is that four people who understood the product were willing to spend months rebuilding it as something nobody could sunset again. Most teams won't have that option. The lesson from CoScreen isn't that Datadog did something wrong. It's that renting mission-critical software was always a bet on someone else's roadmap staying stable, and that bet keeps losing.