Will more people make it later?

"Adding manpower to a late software project makes it later." A newcomer isn't productive on day one — they ramp up, and while they do they pull a mentor off the work. Worse, every extra person multiplies the communication pathsA team of n has n(n−1)/2 pairs who may need to coordinate. The cost grows with the square of the team, so each head added taxes everyone already there. the whole team carries. Past a point, the drag outweighs the hands — and the finish date slides the wrong way.

The project
A late project hasperson-daysThe work still to do, measured in person-days: how many days one fully-productive person would need to finish it alone.left and a team of. Addpeople, each takingdays to get up to speed.
While ramping, each newcomer takes% of amentor'sA teammate has to answer questions, review work and pair with the newcomer instead of shipping. That mentoring time comes straight out of delivery.day, and every extra head adds% coordination drag.
Finish date
33.5days with +3

Adding 3 persons ships it 10 days earlier than the day 43.5 you'd hit with nobody added.

48.7 days
2040
add 0+5+10+3+12
43.5add nobody
+12sweet spotThe number of people to add that finishes soonest. Often zero on a nearly-done project — the ramp-up never pays back.
—turns later atThe first headcount addition that finishes no sooner than adding nobody. Beyond it, every extra person makes the project later.
Here the extra hands pay off: 3 more persons pulls the finish in by 10 days. There's enough work left to absorb the ramp-up.
a day-by-day model of ramp-up, mentoring and coordination drag