While working in consulting I spent about 4 months with a manager who had 60 direct employees. That was a disaster, but I got to spend 3 /4 months with no responsibilities :)
The rule of thumb at Google is 10 directs in an engineering management role, less if you’re a TLM (an IC with some management responsibilities) since you’re still evaluated on the IC ladder. There are exceptions, but they usually involve “managing” a number of contractors where you have significant program management support for coordinating their day to day activities.
A micromanager will be absolutely crushed by a high number of reports. In fact 10+ years at Google, managers intentionally had a large number of directs so they couldn't micromanage.
If you have a pretty autonomous team, strong tech leads, can delegate, ~20 still isn't great, but it's not horrible. It's not the steady state you want, but it's doable for a few quarters.
I think the real solution here is to find the micromanagers and train them or fire them, rather than hampering all the managers in order to cut the micromanagers off at the metaphorical pass.
Personally I’d quit in such a situation. If I’m going to be managing 20 people and creating informal team structures under me, then some of my directs deserve a management title and I would want a title befitting a manager of managers. The whole thing has a whiff of “how can we avoid promoting a bunch of people” to me.
My manager has about 25 reports and he thinks it's ok, because we're all organized in smaller teams that manage daily work themselves. That's true, but the effect is that the person who should care about people aspects has no idea about you (how you're doing, how you want to evolve, whether you're performing, what you need to improve, whether you deserve a raise or a promotion, whether you're burnt out, whether you're looking for a new position...).
And nobody tells him, because he's the manager and you don't go to your manager with problems, especially if he doesn't maintain a close relationship with you.
For instance, has she had to reduce the frequency, duration, and/or quality of her 1:1s? I’m guessing the answer is “of course” and maybe some ameliorating fact like having more 1:1 time with the team lead. The “so what?” here is that 1:1s are how you head off problems before they get serious. So either the tech lead is picking up that slack and is your de facto manager or you’re simply coasting on the strength of past 1:1s, but inertia will only get you so far.
Also, this tech lead you mention is likely doing less actual technical work because they are picking up slack left by your manager not having enough time in the day. This might not be a problem right now, but the first question I would be asking in your manager’s shoes is: “is my TL happy? Is this sustainable for them?” The second would be something like “What long term technical objectives are being sacrificed to deal with this management issue?”
It’s often the case in the situation you’re describing that the TL is interested in moving into management, and this is a “try out” effectively. Personally I like to be open about situations like this, although there are valid arguments against that. In any event, I think most people would agree that it’s best do such “try outs” when the stakes are relatively low, and the TL can have an easy escape if they need it. The situation you’re describing doesn’t fit my definition of low risk, although ymmv — maybe conditions are such that this is a very temporary state and the machinery to fix it is already in motion.
The best way I’ve seen to accomplish this is have a job title that’s purposefully ambiguous. Usually I’ve seen this as a “lead engineer”, which can come with or without direct reports. This lets people try out management while giving them space to change their mind without losing title or status.
I’m good friends with one of the high school teachers at “an elite private school in the Bay” (HRS) and it’s fun comparing and contrasting our jobs, but they are not at all the same job.
They're saving a bunch of money just having one useless person.
It would be rude to describe my anatomical response to having 25 reports, but suffice to say it wasn’t good. I generally consider 6-7 the ideal max, and 8-9 the “I’m dropping all the unimportant balls in order to keep juggling” line. 25 is so far beyond the pale I can’t imagine how any manager of managers’s worth their salary would let it get that bad. 10 reports is “you put this off a bit too long” territory, 25 is in the “someone up the chain should be fired” territory.