One way is to look at the organization as a random graph. Relations between people - edges - describe business dependencies. If the growth has been completely o

One way is to look at the organization as a random graph. Relations between people - edges - describe business dependencies. If the growth has been completely organic to that point, there are chunks of undocumented and unnoticed connections. Unnoticed, until they break, I presume. The complexity grows as a square of the number of people. With 40 people there are 40 x 39 random edges (1560). With 20 (20×19) the number is 380. With 10 (10 x 9) 90. So, as back of the envelope estimate a company with 40 people is 17 times more complex than a company with 10. I'm just blowing hot air, but it seems to me people - engineers especially - are not discussing dependency graphs in relation to this problem as much as they could. Also, root cause analysis suddenly looks like a much more powerfull tool to discover the connections in the graph (if those hidden dependencies cause major pathologies).