AMO hides its long tail

Popular add-ons get recommended on AMO while new add-ons are very hard to discover. This creates a certain “apple app store”-like dynamic that might work for some commercial partners who have a whole marketing department behind them, but hinders free software innovation. Why? Because small projects will remain niche for a very long time. The main distribution platform works against them.

I already ranted about this here a while ago. Imho the situation has worsened since then. We now have a “tightly curated” list of extensions which “may only feature up to a few hundred, or less, at a time.” Every project that is not on this list gets a warning message slapped on its AMO page:

This is not a Recommended Extension. Make sure you trust it before installing.

I understand the intention, but what most users will read is simply this: Mozilla does not recommend installing this! Do not trust this piece of code. It could be anything. Use what everyone else is using.

Things like this will further increase the gap between the popular and the unknown, the known and the new.


I too think that above warning sends a misleading signal. Mozilla has removed the yellow color, which helps a tiny bit. But I still think most people will read above as:

“Mozilla has analyzed this extension, and concluded they wont recommend installing this”

I think that a slight re-phrase like the following will be a big step in a better direction:

This is extension is not part of Mozilla’s “Recommended Extensions” program. Make sure you trust it before installing.

I was also surprised by these warnings. It’s basic English: “not recommended” always means “I don’t recommend this”, “I recommend against this”. Nobody ever says something is “not recommended” when they mean that they just don’t know anything about it, or haven’t tried it. Everyone is familiar with the concept of “recommended items”, like “Amazon’s choice” etc., but they don’t go putting stickers on everything else saying “this is not Amazon’s choice - beware!”. As far as I’m concerned, continuing to use this wording is not a recommended course of action.

It would be more accurate to say something like stig suggested, or “This extension has not been reviewed by Mozilla”, etc. It should be made clear whether it has been reviewed and failed/rejected, and therefore not recommended (recommended against), or just not reviewed, or not as extensively reviewed.


Re-phrasing the warning might better the situation, but for me it feels a bit like fighting symptoms. The real issue I see is a strategic one.

Does Mozilla see the world of browser add-ons as a space where innovative and ux-driven free software development happens – as some sort of open lab or creative playground? Does it support smaller initiatives, so they can grow? Is it proud to be a host of experimental projects with often no financial backup and outlook at all?

Or does it want to get rid of those sometimes amateurish, un-save or even potentially malicious “hobby projects”, so a professional publisher-driven “store” can arise?

Well, I asked about the current install rates for the non-Recommended Extensions and how they were impacted by the messaging in place right now. Caitlin Neiman was kind enough to respond. Mozilla is monitoring these numbers and apparently they are remaining stable. See that post here. Even though I think the current wording is problematic, that tells me they do care and are monitoring the situation. Even if we disagree, I hope we can keep from assuming bad faith on anyone’s part and work to a great solution.

The recommended numbers grow while the non-recommended stagnate. That means that the gap widens.

I am not assuming bad faith! It’s a matter of strategy – and I am eager to know what mid and long-term goal the Mozilla add-on team is pursuing.

Popular extensions getting the most attention is not new to AMO. I’ve been working on add-ons for over ten years, and the top dozen or so extensions have always had orders of magnitude more users and downloads than the rest. Most people are looking for an ad-blocker, a privacy tool, or a video downloader. The rest of the extensions are pretty niche. That doesn’t mean they aren’t valuable, but it does mean their numbers won’t grow in the same way. Giving everyone in the ecosystem an opportunity while giving users what they most likely expect is a difficult balancing act that exists in many contexts (like the Apple App Store, as mentioned above).

Having staff dedicated to select which extensions are Recommended, instead of only going for the most popular ones, allows us to at least have some balance in the kinds of extensions that end up getting the most attention, and signal-boost some less-popular extensions we think users will be especially interested in.

It is unfortunate that such a small number of extensions can be offered without the “trust” message, but we are looking into various ways of expanding that number. It is in everyone’s best interest that more extensions can be confidently installed by end users.