Why understaffed teams stay understaffed

That distinction matters more than anything in this article.

A problem has a cause you can address. A loop feeds itself: low retention produces understaffing, which consumes capacity to onboard. Weak onboarding can add to retention risk. Each cycle gets faster because the ratio of people who know how things work to people who do not is deteriorating each time round.^[2]

Loops do not stabilize on their own. They accelerate until something external stops them, and in this case the external thing is usually a bad quarter followed by an expensive realization.

The reason this deserves its name is that it responds badly to the obvious intervention.

The instinctive response to being short-staffed is to hire more. This is the response to almost every other operational shortage, and it is the wrong response to this one. Onboarding capacity is not a function of headcount. It is a function of headcount and available time, and that is exactly the resource the loop is consuming.

So when you double your hiring rate, you have not doubled capacity to bring people up to speed. You have halved the attention each new person gets, which means they take longer to become useful and are more likely to leave, which reduces headcount further.

There is a threshold worth knowing roughly where yours is, beyond which hiring faster produces fewer productive people than hiring more slowly would. Most companies do not know their threshold. Discover it by crossing it.

At this size, a company may not have a dedicated HR function. A People lead may also be doing office management and recruitment coordination. A founder may be doing it between other things. In SME research, onboarding at larger firms involved handoffs between recruitment, HR and IT.^[3]

Which means nobody has a view of the loop. Everybody has a fragment.

A detail that matters for a fintech in particular.

When people talk about onboarding capacity, they tend to mean availability: somebody to have coffee with the new joiner, somebody to explain how the expenses system works. That kind of capacity is fungible. You probably have enough of it.

The capacity you lack is specific. The person who can bring a payments engineer up to speed is another payments engineer who understands your reconciliation logic. The person who can onboard a compliance analyst is someone who knows how escalation judgments actually get made, which is not written down anywhere and differs from the manual.

There might be two people in your company. They are also your engineers and your most trusted analyst, which is precisely why they are pulled into every incident, customer escalation, and architecture decision.

So the bottleneck is not that senior people are busy. It is that the same three individuals are simultaneously your incident response, your knowledge, your interview panel, and your onboarding capacity. Any plan that assumes you can draw on all four at once is a plan that will not survive a month.

There are three ways to respond. They are not substitutes.

Start with capacity you have, not hiring plan you wish you had.

At some point this becomes a budget conversation. It is worth doing arithmetic rather than arguing about principle. You have three options. They are not the same size.

Hiring more slowly for a quarter costs revenue you would have earned if you had people. It is free in cash terms, but hard to measure and expensive to discuss.

A People lead costs more, but gives you something the other two options do not: a person who sees the whole process rather than one part of it. This is the eventual answer, but changes take time to show up.

You can also reduce how much onboarding depends on people who are in short supply. Use software, and do the work of writing down what your experienced people know. Price varies by company size, scope, and employee count.

These options are not the same. The cost difference between a People lead and software is worth considering before deciding you cannot afford to fix this.

Do not take a vendor’s payback model. Use three numbers you already have.

Hiring volume alone does not tell you whether the loop exists. If the problem is documentation, write things down, book conversations, and spend money elsewhere.

When the real cause is compensation or management, no onboarding process fixes it. Poor terms and conditions and limited career progression are among the retention barriers employers report. Find out why people actually left before you buy anything.^[4]

When nobody owns it, a system produces a configured system that nobody uses. This does not have to be a full-time role. It has to be someone’s named responsibility with time attached.

When you are about to reorganize, onboarding built around a structure that changes in three months is work you will do twice.

Not a headcount. Not a payback calculation on its own.

The moment worth acting on is when your engineering manager, recruiter, and CFO are each describing a problem. All three descriptions are accurate. None is the same problem. That is the loop being visible in fragments. It means nobody currently has the whole view.

In a business, there is a second trigger, and it arrives earlier than most founders expect. Once you have obligations tied to who your people are and what they have been trained on, you have an administrative burden that does not scale by working harder at it. That one comes with a deadline. It is not a good deadline to meet at the last minute.