Having been a contributor (mostly to Thunderbird, though) for a decade, I hold that Mozilla's biggest mistake was effectively killing off embedding and stable extension APIs. They merrily took the opportunity afforded by the Gecko 2 transition to kill all the compatibility, which is understandable, but the task of providing a simple frozen or at least not-going-to-change-soon API kept falling off the radar. It meant that technology experiments like Node.js or Electron--which is the sort of thing that was entirely in line with Mozilla's philosophy--couldn't be effectively built on Gecko and SpiderMonkey, at that means ceding a large captive market.
The extension ecosystem was always one of the strongest distinguishing facets of Firefox. I think Mozilla panicked when Chrome threatened Firefox's market share, and started trying to compete hard on things like being at the forefront of new technology adoption at the expense of the existing extension ecosystem. And in the process, Mozilla lost control of the messaging and it came to be seen as constantly doing little more than aping Chrome. The Web Extensions stuff is an example of the marketing issue. Mozilla new as far back as 2010 that extensions were the biggest stumbling block to getting multiprocess working. What they should have done was start developing the new extension APIs that far back, and made Firefox use multiprocess mode only if no legacy extensions were installed.