Your Product Can Be Delisted On A Date You Didn't Pick

Google removed every remaining Manifest V2 extension from the Chrome Web Store. That included uBlock Origin, probably the most respected content blocker ever shipped for a browser, along with uMatrix, NoScript, and a long tail of smaller tools nobody wrote headlines about.

The Chrome Web Store team notified affected developers. Google's note was clinical: extensions already installed on Chrome 138 or earlier stay installed, but they cannot receive updates and cannot be reinstalled from the store once removed.

So the software still runs. It just cannot be found, cannot be fixed, and cannot be handed to a new user.

This was the last step in a migration Google has been signposting for years. MV2 was disabled for ordinary users in 2025, the remaining flags and workarounds disappeared in Chrome 150 and 151 over the summer, and August 31 closed the door. The stated reasons are security, privacy, and tighter limits on what an extension can reach. Given how much access an extension has to somebody's entire browsing session, those are real problems worth solving.

The Hacker News thread hit 524 points and more than 400 comments within a few hours. Not because anybody was surprised by the timeline. Because watching it finally happen felt different from reading about it on a roadmap.

Because the mechanism has nothing to do with ad blocking.

Strip the specifics. A developer builds something people love. The thing depends on a platform capability. The platform decides that capability creates risk it no longer wants to carry. The platform announces a migration. The developer either rewrites to fit the new rules or stops existing on that surface. On a date the platform picked.

Nobody cheated. There was no ban, no appeal, no policy violation, no bad review cycle. The rules changed underneath a working product, and the working product lost.

Raymond Hill, who builds uBlock Origin, was not offered a negotiation. Neither will you be. Google did not need to be hostile to end a distribution channel. It only needed to change its own architecture.

Most early companies have exactly one of these dependencies and have never written it down. Chrome Web Store. Apple's App Store. Shopify's app directory. Slack's marketplace. Zapier. A Google Workspace add-on. Every one of those is a place where a version number change can quietly become a shutdown notice.

This distinction is the whole story, and almost nobody tracks it.

Your installed base is the set of people who already have your product. Your reachable base is the set of people who can still get it. On August 30 those two numbers were roughly the same for MV2 extensions. On August 31 the second one went to zero while the first one barely moved.

That gap is where companies die slowly. Revenue looks fine for a quarter. Retention looks fine. Nothing breaks in the dashboard. But new installs flatline, and the existing base decays the way every install base decays: new laptop, browser reset, corporate reimage, a colleague who asks where to get it and gets no answer.

So track them separately. Two numbers, one chart, reviewed monthly:

1. Active users acquired through each distribution surface in the last 30 days.
2. Whether each of those surfaces can deliver a new install today.

If the second one ever goes to zero for a surface carrying more than a quarter of your growth, you are not in a growth problem. You are in a survival problem with a delay built in, and the delay is the dangerous part because it feels like time you have.

You do not need a consultant. You need a table and two hours of discomfort.

List every way a new user can arrive at your product. For each row, fill in four columns: percentage of last quarter's new users, who controls the surface, what the published deprecation or policy timeline says, and how long a rebuild would take if that surface disappeared on Friday.

Then sort by the first column and read the fourth.

Most founders find something ugly in row one. The largest channel is usually the one with the least owned infrastructure behind it, because the channels that grow fastest are the ones somebody else already built an audience for.

This is the kind of exercise that dies in a Slack thread unless it lives somewhere with the rest of your planning. Some teams keep it in a spreadsheet tab next to the financial model, some in Notion, some in a planning tool like Foundra that keeps go-to-market assumptions and risks in the same workspace as the plan they support. The tool matters much less than whether the table gets revisited after the first time somebody makes it.

One more column worth adding: the date you last read that platform's developer changelog.

There are four, and they vary wildly in cost.

Migrate to the new rules. Least glamorous, usually correct, sometimes impossible. Plenty of extensions moved to MV3 and are fine. uBlock Origin Lite exists. The question is whether the new rules still permit the thing that made your product worth using. For a content blocker that depends on dynamic filtering, they arguably do not, which is why the full version could not follow.

Find an aligned distributor. Brave decided to host four MV2 extensions on its own backend and let users switch them on inside the browser: AdGuard, uBlock Origin, uMatrix, and NoScript. That is a competing platform turning your problem into their positioning. Somebody in your category is looking for exactly that trade right now. Go find them before you need them.

Ship direct. Self-hosting, sideloading, an installer on your own domain. Higher friction, worse conversion, total control. Worth building as a fallback even if it does nothing today, because a fallback you build under pressure is a fallback you build badly.

Change surface entirely. The browser extension becomes a desktop app, a proxy, a service. Expensive and slow, but the underlying value usually survives even when the delivery mechanism does not.

The useful move is deciding which of these you would pick, in writing, before the notice arrives. Under deadline you will pick whichever one you already understand.

Two things, and they are both cheap when nothing is on fire.

The first is a direct line to your users that does not route through the platform. Email addresses you own. A newsletter people actually open. A Discord, a forum, a status page they have bookmarked. When a store removes your listing, it will not tell your users where you went. If your only relationship with them lives inside somebody else's product, that relationship ends with the listing.

The second is a documented second channel that has delivered at least one real user. Not a plan for a second channel. A second channel that has been switched on, has produced signups, and has a known conversion rate, even if it is a bad one. The difference between a tested backup and a theoretical one shows up in the first week.

Both of these look like busywork when the main channel is working. That is exactly when they are affordable.

One caution: do not read this as a reason to avoid platforms. Distribution through somebody else's store is often the fastest growth available to a small team, and refusing it on principle mostly means being unknown. The point is to know the exit before you need it, and to keep one door that only you control.

• Google removed all remaining Manifest V2 extensions from the Chrome Web Store on August 31, 2026, including uBlock Origin. No rule was broken by anyone.
• Installed extensions keep running but cannot update and cannot be reinstalled from the store.
• The Chrome Web Store also feeds other Chromium browsers, so the removal reaches users who never chose Chrome.
• Track installed base and reachable base as separate numbers. The gap between them is the early warning.
• Audit distribution with four columns: share of new users, who controls it, the published timeline, and your rebuild time.
• Brave now self-hosts four MV2 extensions, a reminder that an aligned distributor is sometimes the fastest escape hatch.
• Own an off-platform channel to your users before you need one. Email you control beats a store listing you do not.

Does this only affect Chrome users?

No. The Chrome Web Store is the main extension marketplace for Chromium browsers generally, so people using Brave and other Chromium builds discover and install through it too. Removal from the store reaches them even when their browser still supports MV2 technically.

Can existing users keep the extension?

Yes, for now. MV2 extensions installed on Chrome 138 or earlier remain installed. They will not receive updates and cannot be reinstalled from the store once removed. That means no bug fixes, no filter list changes shipped through updates, and no path back after a clean install.

Was Google wrong to do this?

There are positions on both sides. Manifest V3 restricts what extensions can do with network requests, and extensions have enormous access to browsing activity, so tightening the model addresses actual security and privacy risk. Critics point out that the tightening also weakens ad blocking on a browser built by an advertising company. Both can be true. Neither changes what a developer needs to do next.

What if the platform is my only realistic channel?

Then build the user relationship anyway. Collect email at first run, publish somewhere you own, and keep a direct-download path working even if almost nobody uses it. You cannot always diversify distribution. You can almost always diversify contact.

How often should I re-run the audit?

Once a quarter, plus any time a platform you depend on ships a major version or a policy update.

1. Google Has Removed Manifest V2 Extensions From the Chrome Web Store, Including uBlock Origin (Web Iterate)
2. Hacker News discussion: Google Has Removed MV2 Extensions from the Chrome Web Store
3. Chrome Developers: Manifest V2 deprecation timeline
4. Brave: Shields and Manifest V3
5. uBlock Origin v1.74.0 Is the Final Version for Chrome Before Google's Delisting (Security Online)
6. The last lifeline for uBlock Origin in Chrome is almost gone for good (PCWorld)