@Fox Thanks a lot for taking the lead on bringing up this topic, and sorting it out! Huge kudos for that, it is indeed one of our main challenges currently. And it’s true that we have focused on increasing the visibility of the ad, but there are quite a few other things that we can explore to help fix that.
And thanks for all the insightful comments and ideas btw – it’s a dense thread, which gives plenty to think of and experiment with to solve this. I’ll try to cover the main points brought up so far:
Job ad & attractiveness
The job ad could definitely use a refresh, as it only had a few changes to it over the years. I like the new version a lot more, it gives a much better idea of who we are – and that makes sense, as this is something we have progressively established over the years. :) I’ve also done a pass earlier today (thanks for reviewing my additional suggestions and merging them already btw @Fox , that was fast! :) ).
Btw, note that we haven’t actually had a low number of applications - we have currently ~30 applications per week, which is more than what we had in previous years, and even quite regularly good candidates that we bring in for the trial. But a higher proportion of the candidates we accept seem to pick a different company before signing their contract – maybe because there are so many more companies recruiting remotely lately, so people who want to work remotely have more offers? It might also be more difficult to stand out next to the big name brand corps, who used to be very reluctant to remote work – and who still don’t really provide a good environment for it (or at all), but who provide a shiny reference for the resume and inflated paychecks. Also (or maybe as a consequence?), fewer of the candidates who joined passed the trial, even when they looked very promising in the initial.
To help fight this, working on showing what makes us different, and a nice environment for those who would fit well here, is definitely a good move – so things like improving the job ad, the handbook, or the website are definitely good ones. Actually:
- About the job ad, I’ll update all our job postings next week, once we have had a pass of proofreading done by gramlee
- About the website & handbook improvements, would you want to take on an epic on this @Fox, with maybe @gabor since you suggested it, with @Ali @cassie @larrybotha? This would have priority over the other internal epics UX work, as recruitment is definitely blocking them currently, on the development side.
As a side note, I’m also still regularly looking at expanding the reach of our job ad to new boards – and I’m also going to try something new, which is to use more traditional ads, not specifically job board ads, to reach different profiles – people who might not even know yet they want a job here. 
I’ve reached out to ReadTheDocs to give it a try – we’ll see how it goes, but it could be something to extend to a few places where developers go.
Hiring contractors
+1 to try to fill the gap with contractors – though it’s true that the tricky bit, just like with hiring normal team members, is to find good ones (or for them to find us). But yes, even though we generally prefer to work with full permanent team members, especially for core functions like development, it still could make sense to temporarily ramp up the number of contractors in the team while we sort out the recruitment.
On this, @raul came to me with an interesting proposal – would you like to explain it here?
Bug bounties
That does sound like a useful experiment to try. One thing that I’m a bit wary of with it is that bounties are based off a monetary incentive – and part of the reason why we have third party open source contributions as a prerequisite is that it takes a specific mindset for someone to set aside some of their (or their employment’s) time, which for developers is always being competed for by a thousand different things, and fix a bug not just for themselves, but also publish it publicly for others to reuse, and go through the whole process of submitting it, getting it reviewed and accepted by an upstream. Most people “don’t have the time” (or, at a minimum, will accept to keep working for an employer who doesn’t allow this type of contributions) – that’s precisely what makes someone who still finds the time (and courage) to contribute special. They want to share and contribute, regardless of the external incentive – and that’s part of what makes us who we are, I think.
That said, we can see how that goes. A third party contribution is still an achievement by itself, and we don’t filter out people who did their contributions as part of a previous job anyway, so there is no reason we couldn’t still find good people this way. I’m guessing successful open source bounty hunters do this for more than money, like we do…
What would be good upstream tasks we could start with? Maybe some of the pending release group bugs for Lilac @adolfo? If we can build a list of tickets to submit, I can take care of posting the bounties. Ideas for the amounts to give for each of the bounties, or time estimates for the effort, would also be useful.
Headhunters & reaching out to potential candidates
I’m ambivalent about that one, as headhunters tend to simply spam a lot of people, and both ways (sending also lots of irrelevant candidates). They are also extremely expensive, taking months of salaries for any hire – and many definitely wouldn’t agree to a refund for all of the newcomers we don’t confirm as core devs… If some of you have worked with recruiters you liked (either to recruit, or be recruited), let me know though, we can experiment. I’ll also try to figure out if I can identify someone for this.
Alternatively, we could also start looking ourselves for potential candidates – through our job postings, we already have access to some databases. @paulo @usman @adolfo Is that something you would like to look into, as recruitment managers? Or maybe others would like to help punctually on this?
Organizational development & consultants
Yes, if we had taken the usual path of most normal companies, these might be useful, but for our model we are one of the experts. (Though not the only one, so it could be worth reaching out to other similar companies?)
We could potentially sell our own services on this one day – but yes for the current issue that wouldn’t really help us solve the problem, we have already too much to do :) We also have to show that we can grow with our model first. I don’t have any doubt that we can, we have already passed some pretty difficult growth milestones in the past, but we actually need to tackle and solve those problems – which is actually what we are doing here. :)
For the transmission of our knowledge, it might also make more sense to start by publishing it, either as text on places like the blog or the handbook, or with courses, like we are doing with our knowledge about contribution in the upstreaming MOOC. If those work out, we can develop more like it over time. It also makes sense to release the information freely, rather than asking people to pay individually for us to release our knowledge to them – which sounds like a more proprietary approach ;p
Upstreaming MOOC
Talking about the upstreaming course, this is also part of the strategy for recruitment, although on the longer term, since it will take time to complete it and release it. Part of the idea is to attract people interested in contributing to open source, and maybe getting a job in open source – and, not coincidentally, one of the criteria for completing the course is to have a merged contribution to a third party open source project… precisely the main requirements for recruitment that is the trickiest. It would also hopefully raise our profile within the open source community, and get us more known of open source enthousiasts – especially the ones stuck in horrible proprietary corporations. 
Addressing newcomer feedback
We do reliably gather (and regularly address) the newcomer feedback at Issues · opencraft / documentation / public · GitLab – even when they provide feedback at the exit 121, if it’s relevant I usually ask them to post it there, or in the forums if that’s a more general issue.
It’s definitely a good opportunity to do a pass on those though – even if we didn’t have any tension on the recruitment, as we’ll need to onboard quite a few people over the next few months anyway, and they would benefit from it. @gabor you have a pass coming up on this, as part of SE-3755, right?
That said, iterating on the onboarding has been an important focus from very early on, and generally people who are good are able to work things out, even at times where things weren’t as laid out as they are now – so I’m not sure if this will change our trial retention rate, generally it seems to have been issues with the person/fit, rather than the process? I could be wrong, though.