I’ve been recommending the Mythical Man Month a lot…this month, so I figured I’d circle the drain on the topic for a bit. People don’t scale the way that you think they do. Having two people on a project doesn’t make it 2x faster. Sometimes it makes it 1.5x slower. Why? Communication between two people is always (ALWAYS) inefficient. It’s the equivalent when you compress audio, you’re losing something in return. You always have to pay a cost, whether it’s the compute resources or the actual data. It takes energy to communicate thoughts and ideas and it’s a “lossy” process (think of the game of telephone).
I constantly find myself in this quandary of deciding between delegating engineering and product development tasks versus doing it myself. “But Tony, that’s not a good use of your time.” Oh yeah? You think so? Did you do the math on it? Got that good good calculus down don’t cha? Yeah, didn’t think so. Have you tried drawing a picture, but instead of drawing said picture, you hired someone else to draw that picture for you? How about if we made it a bit more interesting? What if you hired two people to paint your picture while you described a part to each of them? Let’s add some additional complexity and make it so you had to describe the portion of each picture separately to each hired painter. Now let’s really throw a wrench into your process and hire someone to design the picture for you, so you can just describe generally how you would like the picture. If that’s your only option because you can’t paint and can’t design, then I suppose that math makes sense. Three heads to help one person realize their vision. That’s an expensive venture you’ve got going there. Let’s break down the math then. How does the delegated painting operation make sense? If:
traditional wisdom model = (2 paints + 1 designer ) * (time * inefficiency_coefficient * production_speed)
is less than:
self-sufficient model = (1 you) * (time * production_speed)
What does the Mythical Man Month teach us? That adding headcount to a project does not directly correlate with decreasing the estimated timeline of a project. Why? Well, a myriad of network inefficiency issues, such as communication inefficiencies and the cost of delegating work. As a result, the inefficiency_coefficient noted above will likely always exceed 1.
“Tony, there’s no way that’s the case in every scenario, otherwise companies would just consist of 1 person”
That’s correct. We’re assuming a lot of things, like the one man show actually has all of the skills of the other specialists and can work at the same speed or faster than the specialists. That’s a large assumption and unlikely the scenario in most cases. There’s also the additional math around whether the one-man show is better served tackling some other task which yields greater value while the other specialists work on the delegated project.
My point of all the above? Sometimes it makes sense to just roll up your sleeves and get to work painting, even though you don’t want to or even if popular wisdom says that it’s not a “valuable” use of your time. Sometimes the inability to efficiently delegate is actually a competitive advantage, since it would take your competitors no less than 2 + n more headcount to accomplish the same task, and not to mention the cost that’s associated with such headcount. I’m tired of hearing that “delegation is key” for every scenario. Sometimes that’s just an excuse for laziness.
