Your system will break where you go cheap

Therefore, make deliberate choices

This is one of the "lessons in an article" that I could have decorated with the big words "Systems Thinking" or the grandeur of "looking at the whole". However, I find it much more adequate to very non-philosophically come to the key lesson that any software developers worth their money can tell: Your system will break where you go cheap. Let's go through some classical options:

Operations (or Ops)

This is based on the idea that running a complicated system "can't be that hard". Have you ever done it? No? Thought so. Here, the usual tragedy kicks in: Success is based on nothing terrible happening; but if nothing happens for too long, people wonder whether it is worth paying for high quality. Operations done right is like a nicely working digestion: Too many people only realize their value when they are gone.

Ideas like DevOps cannot heal this mistake: If Ops is not paid enough, there is an incentive to hand over as much as possible to Dev.

Support

There is the idea that "anyone can do 1st-level support". This thinking can be corrected quickly by being on the other side of a support call and having someone who obviously has no clue. A competent support colleague knows how to diagnose problems and keep a human touch while doing so. Only paying the bare minimum means people switch jobs as soon as they can; hence preventing the necessary knowledge from being built up. Plus, the first line has an incentive to do the bare minimum before forwarding tickets to the 2nd or 3rd line.

Software Development (including Testing)

Nothing against using external software developers - I was one for almost a decade! The tragedy starts when you pay too little, using externals as an "extended work bench". Low salary leads to low morale, low responsibility, and ultimately low quality. The internal colleagues need to compensate for this; leading to their skills and knowledge being consumed in chasing and fixing defects of others instead of creating high-quality work on their own in the first place. The coordination costs incurred are usually hidden.

Consulting choice

It is not the consulting itself that is cheap - on the contrary, it is often ridiculously expensive. But the choice which company to engage and what advice to accept seems to go the path of least thinking and least experimentation. Of course, asking the usual big consulting companies and going for the obvious ("what everyone else is doing or saying you should do") is very unlikely to solve deeply-rooted long-term problems.

Management thinking

Managers' salaries are anything but cheap; but the thinking and behavior that is incentivized in hiring and promotion decisions can be extremely conventional. Who will go for the new if the traditional path leads you to success? This is very effective in holding back a department or even a whole company.

What to do instead

Although people understand that "a chain breaks at its weakest link"; this wisdom is too often forgotten for the systems that are our workplace. So much about the widely touted "end-to-end" view and responsibility!

But it does not have to be that way. You always got a choice!

Option 1: Do not go cheap. Pay decent or above-the-ordinary salaries in the countries where you hire. Then you can attract and bind people who really want to do the job. Alternatively, hire one person with a good salary instead of several with low ones.

Option 2: Make a decision and watch the results. Think of circuit breakers. If you know where to expect the break, it will be easier to detect and fix it. One lesson from the book Team Topologies is that friction can be a signal for necessary change. If you see the workflow breaking down at the place where you expected it, you can make a choice whether the cost of cleaning up is higher than the cost of preventing the problem; hence whether it is worth spending more money.

Attempts at lowering costs can turn out very expensive. They get paid in one of the hardest currencies which is abbreviated TYS - "Told You So" (coming back at the software developers).

Cheap Trick: Surrender