Org Rot

You've seen it before whether you know it or not.

Ever try to hop on a call on Google Meet, the enterprise Google Chat, previously called Duo, which replaced Allo, the replacement of Hangouts, which was the rebrand of +Hangouts, that replaced Talk and Voice? Maybe you tried to join an interview on Microsoft Teams with (For Home and School), only to find out you need to download the separate version (For Business), but for some reason (Classic) works, but definitely not (New).

This process all begins when incentives aren't aligned internally in large companies, and it creates a feedback loop that degrades the org over time.

Executives have ideas for things they want to do to drive the business. Line level employees want to maximize the value they capture per unit of effort. Middle managers want to keep things stable by keeping employees and executives happy so that things don't get too chaotic. Each party has a different incentive structure.

When an executive brings in a new idea and the middle manager can see it doesn't fit into the roadmap without their team being upset about the increased workload, they will say they can't do it right now.

The executive will ask them to quantify how much work it is, so that they can figure out how much of an investment it will be.

The middle manager says fine, and gives a kind of high estimate because they don't really want to deal with it.

The executive thinks their new idea will drive $5M/year and the middle manager thinks the new idea will take 8 man years to deliver. Some napkin math later, 8 * $150k = $1.2M means this is a great financial investment and they should make it as soon as possible.

At this level of abstraction, it seems logical.

Since the middle manager says there isn't enough bandwidth, and the executive sees a very clear return on investment, they tell the middle manager to hire 8 engineers and start the new project asap. The middle manager knows their estimate is kind of nonsense, man-hours are not real, work is rarely parallelizable in that way. Hiring a whole team is slow and annoying, but getting another team expands their fiefdom, which makes them look more important to the machinery that determines middle manager comp. Maybe if they do this a couple more times they'll go from manager to director, or director to VP.

They're incentivized to agree and hire them anyway.

As the org keeps doing this, the friction in communication and alignment slows things down even more (friction is exponential with the number of people coordinating), so estimates for man hours keep growing. This leads to executives getting more and more annoyed at the difficulty of getting things across the line. If they are not thoughtful enough to understand what is actually happening to their org, they will just keep saying to hire more engineers to get what they asked for.

As headcount keeps growing, their payroll does in line. When upper management finally begins to put pressure on driving down the bottom line, middle managers have to care more and more about keeping average comps down, and carrots start being taken off the table.

The best people tend to churn off first because better options exist elsewhere, and the remaining people aren't skilled or industrious enough to run the org with the original team size. Levels of production fall off a cliff because producivity is pareto distributed (the sqrt of people do half the work), then this spiral begins to happen organization wide. It is possible to hold this off, but it requires a level of discipline and clarity that most companies aren't capable of.

Actually fixing that underlying rot in a large org is so hard that it might really be impossible. I've never seen it fixed successfully. A CTO stuck too deep in this will start to feel defeated and check out, hoping to just keep the ship upright and not sinking, rather than even really trying to go anywhere. That's what went wrong with Twitter before, and why firing 90% of staff didn't even really change the trajectory of the company.

But who knows, maybe this all is a good thing.

Companies get old, they rot, and they die, giving room for new entrants to fill the void.