If you need motivation, maybe the organization is designed badly.

It was once said of the Roman legions "The Legion is not composed of heroes. Heroes are what the Legion kills." Field Marshall the Viscount Slim, who commanded in the China-Burma-India theater in WWII, once wrote "Wars are won by the average performance of the line units." He wrote negatively on various special forces type units, preferring to use regular infantry and training them up to a good, but not superhuman, standard. Arthur Imperatore, who had a unionized trucking company in New Jersey, is profiled in "Perfecting a Piece of the World" (1993) for how he made his trucking company successful despite a very ordinary workforce.

There's an argument for winning by steady competently managed plodding. The competently managed part is hard. Steve Bechtel, head of the big construction company that bears his name, once said that the limit on how many projects they could take on was finding bosses able to go out to a job site and make it happen. Failure is a management problem, not a worker problem.

This post is talking about very small companies. At that 20+ person department, it's true. Once you have a team where the founder doesn't know everyone, the average matters a lot more.

If you have 15 people, you can hire 15 people and they will be able to organically organize if you hire well. If they have a question, they know what everyone is working on. The code base is small enough that everyone can just figure it out even if the documentation is bad.

The larger that group is, the more effort it takes to make sure everyone has the context they need to get their job done. That's where management matters.

And honestly, when I was the first manager (team of 17) brought in, I was writing code and on my own project in addition to starting to build up the "what do we need to do to scale?" You bring someone like me in at 17 people because you're going to need to scale soon and someone needs to build the first set of processes that solve the problems of the next stage, and figure out the onramp because done wrong, they make everything worse.

in the military we had a saying "you don't go to war with the army you want, but with the army you have." consequently, there was a lot of effort given to training and planning. the nature of most combat arms roles is such that you need most of the team operating at a decent level. i think the idea behind so much training is that if you can raise the performance of the worst performers, you might be able to improve the overall unit performance dramatically.

to put it in marketing manager speak, for many tasks in a combat arms unit, individual performance is a satisfier, not a a delighter. if one person in the unit does a bad job, the unit will fail. if everyone in a unit does an "okay" job, the unit will not fail. the outcome between the two cases is dramatic. but if you have a unit where everyone is "okay" and then expend effort to make everyone in the unit exceptional, you will not notice a concomitant increase in performance.

flipping this over to software development... you have a lot more control over whom you hire to be in your unit. but everyone has a bad week or a skills gap, so training (which could be as simple as giving people time to read up on a subject or write a few test programs) will eventually be important. like line military units, everyone needs to be hitting on all cylinders for the dev team to work in accordance with plan. investing in upskilling existing developers who are competent but underperforming may be a better strategy than uber-skilling your best developers or firing them and hoping you can replace them with someone with better ability to figure out how to be productive on the team.

as a humourous aside... at amazon my manager discovered i was prior-service, saying "Oh! You were a MARINE!? I want to manage my team like a military outfit." unfortunately, my response was "WHAT!? You want to spend 80% of your budget on training and logistics!?" that was probably not the best thing to say in that situation.

also... if we're talking about applying military metaphors to product development, it's worth it to look up the various OODA talks by John Boyd. i don't know if i agree with all of it, and it's not directly applicable. but there's enough there to justify at least reading about it. Boyd was a friend of my dad's, so i remember thinking he was crazy when i met him as a child, but again, he may have been crazy, but he was definitely an intellectual outsider who hit more than he missed.