Looking at your comments here it seems that you're not really that interesting in keeping him.
Someone suggested 20% time, your response was to say that it was on hold for a year. Someone else suggested letting him work on something he finds interesting for a while, you say that there are a few things he'd like to do, but they're not 100% aligned with what the company wants to do.
Ultimately, you need to decide just how much this person means to you. Do you want their attention 4 days a week, with 1 day being spent on something they find interesting, or do you want them 0 days a week. Either is valid, but you're going to have to choose.
We just don't want to have special policies for special people. All developers are equal and it will look strange to give 20% free time back for a single developer and hold this for all other developers.
Are you going to go through this dilemma for all your developers when they consider jumping ship at various points of their careers? Probably not, so you're already doing something special for this person.
Since you seem to be intent on being a robot with regard to personnel decisions, there is only one question to answer: Is your company going to go down the tubes if he/she leaves? If so, do what you can to keep them. If not, cut ties and move on.
As an aside, I find it humorous that you're consulting an anonymous online message board for this type of advice. Surely you've handled more difficult problems than this in growing to a 50 person company?
One of the things you realize after you get some experience as a manager is that not everyone is equal, so not everyone needs to be treated equally. It's a tech company, not the drive thru line at McDonalds.
People leave for a lot of reasons, and not having opportunities to grow is one of them.
You want to be equal? Everyone with the title Senior Staff Engineer gets to work on super awesome forward looking new space car. Or maybe just 20% time and a ton more stock options.
In which case let them leave and replace them with another equal developer. Though I am confused as to why you call them one of your best and yet say all developers are equal.
There are lots of ways to make people feel good while keeping pay, benefits, and professional relationships professional. Not knowing where OP is from or where their company operates, I'd give them wide deference in terms of ensuring they follow the law, but talking to an HR professional (who should be involved at this point anyway if there are dozens of folks) can give a lot of good ideas on how to make people feel special without breaching contracts.
As a practical example, Dynamighty listed every full time Dynamighty worker in alphabetical order on CounterSpy, but then called out individuals by their major contributions within the longer credits. It's not a perfect system, but it provided a good balance between giving due credit to folks who worked long hours for equity and those who were making normal salaries the whole time.
It's not unusual for upper management to have discretion and a budget. Forget side projects -- work with him to find an important and interesting thing you won't micromanage at all.
Perhaps importantly: maybe "lead" roles no longer have any meaning, but if you are determining the engineering schedule months in advance, in enough detail for him to know it won't be interesting to work on, he could be frustrated by the micromanagement (whether it is real or only perceived is something I cannot know).
Yeah, unfortunately the high salary is still not proportional to abilities. You get maybe 30 or 40% more for being 2 to 3x as useful (we'll forget about the mythical 10x-ers for now.)
This has always perplexed me. You could hire a junior on say 10,000 credits who could produce 10 units of work per week. You could then hire someone more experienced, on 25,000 credits, but they could produce 50 units of work per week. You could then hire seniors and pay them 40,000 credits, but they're producing 200 units of work per week.
Why is it fair to increase the wage so slowly? Either less experienced developers are not worth the money to begin with or someone, somewhere has an idea of what the max cap of a developer is, and sticks to it.
It is really hard to measure developers productivity and nobody knows how to do that with a good precision. But in general I agree that top developers should be paid good money, like 2 times less then CEO.
Equating 20% time to murder seems like a bit of an overreaction. Do you do that often?
Anyway, you're asking a room full of developers what they would want and they're telling you. 20% time is a solution that works. Toiling away on the same problem day after day sucks and it gets boring. 20% time breaks up the drudgery.
I suspect that you don't want to solve the developer's problems though, just your own.
I agree, I overreacted. We had 20% rules for few years with mixed results, but put this practice on hold till Summer. Maybe we will not see such problems when it will be back.
I hope for the best for you with this. In my experience, the mixed results are part of the process. Think of it more as R&D costs rather than "loss" per se, and it may be easier for you (and your investors) to think about.
Management is hard. We're all learners when it comes to working with people, especially when some of the people are smarter than we are. Keep learning!
There's a good comment down-thread about 20% time. You don't have to say "go and do whatever you want." It can be about "non-billable" R&D. Perhaps this developer sees an issue, or a potential optimization, or even a new product or service, but never has time to develop it. This arrangement allows both of you to get what you want.
As far as "special treatment" goes, why not extend this to all the leads? It seems reasonable to have these folks pushing the technical boundaries for the company.
Important that 20% time can be beneficial to your entire team. And this doesn't have to be absolutely free, it can be tangentially related to something that at least sort of aligns to your company - whether internal (support, communication, internal fun) or external (product, trying new platforms / technologies).
And it doesn't have to start at 20%; if your core product can't afford everybody to take 20% time toward interesting things, then 10%.
Someone suggested 20% time, your response was to say that it was on hold for a year. Someone else suggested letting him work on something he finds interesting for a while, you say that there are a few things he'd like to do, but they're not 100% aligned with what the company wants to do.
Ultimately, you need to decide just how much this person means to you. Do you want their attention 4 days a week, with 1 day being spent on something they find interesting, or do you want them 0 days a week. Either is valid, but you're going to have to choose.