Project Processes via Corporate Selection

Here’s an exciting yet boring question: Is it useful to consider the improvement of project processes caused by corporations beating other corporations in the market. I’d say no because the topic is boring. I’ll explain why it is exciting and controversial at the end.

Consider Company X who performs projects for customers and uses a set of standard internal processes called P, Q, R, and S. For example, process P could be to use Microsoft Project for project management, Q could be to contract at a fixed price to the customer, etc. It doesn’t really matter what they are. Now imagine that these are used in projects A, B, C, and D with the following financial results:

  • Project A: price to customer = $1000K, cost to run = $700K
  • Project B: price = $2000K, cost to run = $1900K
  • Project C: price = $500K, cost to run = $500K
  • Project D: price = $3000K, cost to run = $2000K

The project manager of Project D thinks Process R is inefficient and changes it to R2, resulting in Project D with a new cost to run of $1500K. But Company X doesn’t like change and doesn’t make R2 standard.

So, the project manager leaves and starts Company Y. He copies the same processes but though Process R was inefficient and creates a new process, R2. This company then under bids on the same project contracts, wins them, and gets the following project results:

  • Project A: price = $900K, cost to run = $500K
  • Project B: price = $1900K, cost to run = $1500K
  • Project C: price = $400K, cost to run = $300K
  • Project D: price = $2900K, cost to run = $1500K

By using the same project letters I mean to say that, for instance, Project A is an open bid from a customer for a certain set of requirements and deliverables, and the same with B, C, and D.

Company Y competes with Company X for customers and offers a lower cost so wins the bids. Company X goes out of business. As a result, Process R is gone out of existence.

There are two ways to look at what happened here. First, you could do the accounting by project. In Company X, Projects A had a net benefit (profit) of $300K. In Company Y, Project A had a net benefit of $400K. For Project B it is $100K (X) vs $400K (Y), Project C is $0 (X) vs $200K (Y), and Project D is $1000K (X) vs $1400K (Y). From this you can show that by switching from Process R to R2 you improve net benefit in each project. Any individual project benefited by changing from R to R2, even when it was done inside Company X. Hence, you can say Process R died out because it was inferior to Process R2 for individual projects.

The other way you can look at it is by company. Company Y underbid Company X on every project and so won the competitions. Furthermore, Company Y also was more profitable even if they both won parallel contracts, making $2400K total profit vs Company X’s $1400K total profit. One might then say that Process R2 succeeded in surviving over Process R because it was more prevalent in Company Y who beat Company X in competing for customers.

Is this second version of value though? It seems to me this loses some important information. Company Y beat Company X because R2 was better than R on each individual project. This second description might make sense if R2 was actually worse at the project level but somehow allowed Company Y to still beat Company X to win customers and be more profitable. For example, if R2 made Projects A, B, and C cost more and lose money but made Project D significantly more profitable. This would allow Company Y to still underbid Company X, put them out of business, and be overall more profitable. But if that’s the case, why not use Process R for Projects A, B, and C, and R2 for Project D? Then you get the best of both worlds. If Company Z did that, it’d even beat Company Y.

Looking at it from the level of the company seems to muddy the water and make descriptions awkward. What’s really going on is that the survival of processes is competed at the project level, even within a company. How they affect the overall company performance is important when competed against other companies. Even good processes can die out if the company competes poorly overall. But ultimately it comes down to what is the best process for the individual project.

So why is this exciting and controversial? Because this is essentially the argument going on about group selection versus inclusive fitness (kin selection, selfish genes) for explaining evolutionary behaviours. In a prior post I pointed to Steven Pinker’s great essay on the folly of group selection and now there’s a great set of responses (The essay is above them.) giving all of the arguments on both side, including Pinker, Richard Dawkins, Daniel Dennett, Jerry Coyne, Jonathan Haidt, and many other well-known names in the area of the mechanics of natural selection and evolution in terms of genetic, memetic, physical, psychological, and phenotypical components.

I’ve used processes, projects, and companies above because people are more familiar with them and the benefits and costs are easier to see in dollars. The arguments can be difficult to follow if you are unfamiliar with neo-Darwinian terminology and what it means. 

My corporate analogy above essentially uses processes for genes (information carriers), projects for individual organisms, and companies for groups. Some might note a few things left out of the analogy but I don’t think they affect the result. For instance, I didn’t spend a lot of time looking at the competition of processes (genes) within a company (group) through projects (individuals). That would look something like projects competing for human resources to run them, and the most profitable projects winning so those processes surviving. But that effectively makes something like the brief mention of Company Z compete better if it drops Project C as least profitable if it doesn’t have the resources to do them all. It’s still the best process for each project that wins regardless of which company it is in, even if the company eventually goes bankrupt.

The problem of group selection thinking is that it hides what is really going on and leads to awkward explanations. Genetically caused behaviours don’t survive and spread because they help groups that have them survive better than other groups. Genetically caused behaviours survive and spread because they help individuals that have them survive (and reproduce) better, including forming cooperative groups. The success of the group in competing is the result of, not the cause of, the genetic policies of the individuals maximizing their own chance of survival which includes forming groups, sharing risks, and cooperative behaviour up to the point that it becomes a net cost over a net gain. Process R2 does not survive and spread better than R because it helps Company Y (or Z) to beat Company X. Company Y (or Z) beats Company X because individual project managers seeking the best process for their individual project adopt R2 when it is best.