When the Whole Business Runs on One Person

Key Person Dependency - When a Business Runs on One Person

It's 7:30 a.m. on Tuesday. Your production planner is out sick. The schedule still exists in the ERP. But nobody else trusts the dates on screen. Only that one person knows which supplier always slips three days, which job can be split mid-run, and which customer will accept a partial shipment. Without them, purchasing stalls. Quotes wait. Invoices pile up.

In a lot of companies this feels normal - especially during growth, when speed beats documentation. Right up until the delays turn into real cost.

It's called key person dependency (opens in new tab), and it's a recognised business risk, not a minor inconvenience - the same kind of "single point of failure" you'd find in engineering. The only difference is that here the failure point is a person, with their own holidays, health, and life plans.

Why the problem looks small until production stops

A bottleneck is easy to spot - work pauses the moment one employee is unavailable. Updates need a phone call, not a process. Decisions change depending on who you ask. And the customer feels the slowness, because order status depends on chasing one person instead of checking a stable workflow.

The first cost is obvious - output drops. What hurts longer is what nobody budgets for. Supervisors get pulled into routine clarifications. Approval cycles stretch. A new hire needs weeks to learn the exceptions nobody wrote down. Without clear process, small delays multiply across quoting, planning, purchasing, and accounts payable.

Why holidays become a threat to the business

Leave should be routine. In many growing companies, it isn't. Annual leave turns into stalled approvals and schedules nobody updates. Sick leave is worse - there's no time to hand anything over.

And here's the point that's easy to miss: if daily work runs on memory, there is no real cover. A plan doesn't need a thick binder. It needs clear steps, a named backup, and rules a person can follow even under pressure.

"We don't need a system - we have someone who knows everything"

We hear this constantly. It sounds practical. But it isn't a solution - it just means the risk hasn't gone anywhere. It's quietly sitting on one person's shoulders.

Key person dependency rarely starts with bad management. It starts as a sensible response to pressure: orders grow, customers want answers today, and the fastest path is to let your most capable person handle it. For a while, it works. The planner keeps the rules in their head, the estimator remembers which job needs more margin, the finance lead knows which supplier invoice always breaks the usual flow.

Map almost any quoting process and you'll find the same thing: the real logic isn't in the ERP. It lives in one person's inbox, memory, and judgment. Everyone assumes there's a system. There is. It's a person.

A related version of the same friction - a sheet metal manufacturer whose quoting ran on manual back-and-forth - emails, files, and one slow estimate at a time. The fix wasn't more people, it was turning quoting into a self-service flow that prices a part the moment the customer configures it.

The same goes for the "master Excel file" - when the entire business logic is packed into one spreadsheet only its author understands (opens in new tab). While they're working, everything runs. When they leave, the logic walks out with them.

Process first, technology second

This is the most common mistake: software gets bought before anyone defines the actual rules. But the truth is simple, a bad process scaled by software is still a bad process. Just faster, and more expensive.

So the order matters: process first, technology second. In practice that comes down to four steps - whether you do it yourselves or bring someone in.

First, map the real process - not the one drawn on the whiteboard. With the actual triggers, stops, approvals, and the spots where people decide "from memory" because the rules are vague.

Second, capture the rules, exceptions, and handoffs - turning what used to be one person's knowledge into a shared, repeatable method. Exceptions matter most here, because they usually create the heaviest workload.

Third, decide what should stay manual. Not everything should be automated. An experienced buyer will catch a risky substitution faster than any form. Automation helps when rules are clear and outcomes repeat, and helps less when the decision depends on experience or a client relationship.

Fourth, automate only what's stable and repeatable. Sometimes that's simple approval routing. Sometimes invoice capture. Sometimes - no new system at all.

What it actually delivers

The strongest proof isn't that software exists. It's that daily work changes.

In our invoice automation (opens in new tab) project, we mapped the real flow first - the handoffs and the awkward exceptions people handled from memory. Once the process was clear, automation made sense: invoice processing became roughly 40% faster, software spending dropped about 30%, and each person could process ten times more invoices.

We saw the same pattern building our proposal tool. The real blocker wasn't a slow interface - it was key person dependency. One experienced person carried the rules and judgment in their head. We moved that logic into a system anyone could follow. Proposal creation (opens in new tab) dropped from 3-4 hours to 30-60 minutes. Just as important, the work became easier to repeat, easier to delegate, and far less fragile when that person was away.

A practical first step

The response doesn't have to be dramatic. Start with the five processes that matter most to your output, cash flow, or customer service. For each, ask simple questions: Who actually knows how it works? Where do exceptions pile up? What happens if that person is off for a week? Is the process written down, or is it just habit and memory?

That review alone will show you where your business quietly depends on one person, and where automation might help later.

If the process is still unclear, fix that first. Then decide whether a lightweight tool, automation, or no new technology at all is the right move. If you'd like a second pair of eyes on those five processes, get in touch (opens in new tab) and we'll talk through the practical next step.