The Hire You Think You Need Is Usually a Tool You Haven't Built

There's a pattern I've watched play out across dozens of companies: something breaks, or growth stalls, or the founders feel overwhelmed. The reflex is always the same. "We need to hire someone." And then three months and $40,000 later, the problem is worse because it was never a people problem. It was a systems problem wearing a people-shaped mask.

Brian and I run Kief Studio as a two-person operation that consistently delivers at the pace of a much larger team. Not because we're working 80-hour weeks — we're not — but because we spent years building systems before we needed them. That choice has shaped everything about how we think about growth, staffing, and the reflex to hire.

The Expensive Default

Hiring is treated as the default solution to capacity problems, but default solutions are often just the most familiar ones, not the most effective. The costs are staggering once you look at them honestly.

DockYard, a well-known software consultancy, was paying $35,000 per month — over $400,000 annually — for a Boston office that maybe five people were regularly using. That's not a staffing cost; it's a consequence of staffing decisions that created overhead nobody questioned until the numbers became impossible to ignore.

Andrew Field, CEO of PFL.com, put it bluntly in a 2023 interview with Inc.: "All of my worst mistakes have been around hiring — making a bad hire, then not making the call when you know." That second part is what costs the most. Not the bad hire itself, but the organizational inertia that keeps a bad hire in place for months because firing feels harder than tolerating the drag.

And it's not just bad hires that cause damage. Even good hires in the wrong role create problems. COO hires brought in from big blue-chip companies commonly fail at startups and leave within six months, according to a 2022 analysis by First Round Capital. The skill set that works at a 5,000-person company doesn't transfer to a 12-person team, and both parties figure that out around month four — after onboarding, context loading, and the first round of "how we do things here" conversations have consumed enormous amounts of organizational energy.

The Scaling Trap

Era Solutions documented what they call the Scaling Trap in a 2024 case study that deserves wider attention: the pattern where growth does not automatically mean progress. Revenue increases. Headcount increases. And margins shrink. The company is bigger but less efficient, and the additional people create coordination costs that eat into the productivity they were supposed to add.

This isn't theoretical. Every person you add to an organization increases communication paths combinatorially. A team of 3 has 3 communication paths. A team of 6 has 15. A team of 12 has 66. The context-switching cost alone is brutal — and it's invisible on a P&L statement because it shows up as diffused slowness rather than a discrete line item.

The question isn't whether to grow. It's whether hiring is the right mechanism for the growth you need. Sometimes it is. A lot of the time, it isn't.

What Tools Actually Solve

Most of the tasks companies hire for fall into categories that tooling handles better: data transformation, report generation, monitoring, scheduling, notification routing, compliance checking, content formatting, and system integration. These are tasks that require consistency and repetition — exactly what humans are worst at and software is best at.

When I look at a capacity problem, the first question isn't "who should we hire?" It's "what is the actual bottleneck, and does it require human judgment or human labor?" Those are different things. Judgment is worth paying for. Labor that can be systematized is a tool waiting to be built.

The two-person studio model works because we've been ruthless about this distinction. Every repeating task either gets automated or gets questioned. If we're doing the same thing three times, the third time is the last time we do it manually. That's not a philosophy — it's an operating constraint that forces better decisions.

And the tools don't have to be sophisticated. Some of the highest-value automation we run is embarrassingly simple: scripts that format data, templates that generate documents, monitoring that pages us before clients notice problems. None of it is cutting-edge AI. Most of it is conditionals and cron jobs. But it runs 24 hours a day without vacation time, sick days, or the need to be onboarded.

When You Actually Need a Person

To be clear: there are things only humans can do. Relationship building. Nuanced negotiation. Creative direction. Strategic judgment in ambiguous situations. Client communication that requires empathy. These are real, irreplaceable human capabilities, and no amount of automation substitutes for them.

The framework I use is straightforward: if a task requires judgment that changes based on context, it needs a person. If a task requires consistent execution of a known process, it needs a tool. Most roles people hire for contain both types of work — and the right answer is usually to automate the process work and hire only for the judgment work, at fewer hours and higher rates.

This is the difference between knowing when to automate and when to hire. The mistake isn't hiring. It's hiring for a full-time role when you need 10 hours a week of human judgment and 30 hours a week of process execution that a script could handle.

Institutional Memory as Infrastructure

One of the underrated costs of hiring is the knowledge fragmentation it creates. Every new person becomes a repository of institutional knowledge — knowledge that walks out the door when they leave. Building institutional memory that survives turnover is possible, but it requires deliberate architecture that most companies never build.

Tools, on the other hand, are institutional memory by default. A well-built internal tool encodes decisions, processes, and edge cases in a form that doesn't degrade over time, doesn't need to be retrained, and doesn't reinterpret instructions based on mood. It's not better than a person — it's better than a person at the specific task of remembering and consistently executing a defined process.

When I look back at my own hiring mistakes, the common thread is that I hired people to be systems. I needed consistent output, reliable process execution, and scalable operations, and I tried to get those things by adding humans to a workflow that should have been a pipeline. The humans were fine. The decision to use humans for that work was the error.

The Math Nobody Does

Before your next hire, do this math: take the fully loaded annual cost of the role (salary, benefits, equipment, office space, management overhead, onboarding time) and divide it by the number of hours of irreplaceable human judgment the role actually requires per week. If that number is above $200/hour, you're paying for a full-time person to do part-time judgment work and full-time process work.

Then ask: what would it cost to build or buy a tool that handles the process work, and hire a fractional expert for the judgment work? Usually less. Often significantly less. And the tool keeps working after the expert logs off.

The engagement model is part of the strategy. A retainer relationship with a specialized studio or contractor, combined with internal tooling, often outperforms a full-time generalist hire at lower cost and higher quality. Not always — but often enough that it should be the first option you evaluate, not the last.

Async-first operations make this even more viable. When your systems are designed for asynchronous collaboration, you can work with specialists across time zones and schedules without the coordination overhead that makes distributed teams fail. The constraint becomes the advantage.

The Real Question

The next time you feel the pull to hire, pause. Write down exactly what you need done. Separate the judgment work from the process work. Cost both independently. Then decide.

Sometimes the answer is still a hire. But when you've done the analysis honestly, the answer is often a tool — one that pays for itself in months, scales without overhead, and frees up the humans on your team to do the work that actually requires being human.

Frequently Asked Questions

Doesn't this approach just work for tiny companies?

The principle scales in both directions. Larger organizations have even more process work that could be systematized — the coordination costs just make it harder to see. A 200-person company where 40% of work is automatable process execution has the equivalent of 80 people doing work that tools could handle. The savings at scale are proportionally larger, not smaller.

What about roles that seem like they require a person but might not?

The most commonly over-hired roles are reporting analysts, data entry specialists, QA testers (for regression testing specifically), scheduling coordinators, and compliance checkers. Each of these involves some genuine judgment work, but the majority of hours are spent on process execution. The right move is usually to automate the process work and redistribute the judgment work to existing team members or fractional specialists.

How do you know when you actually need to hire instead of building a tool?

Three signals that genuinely indicate a hiring need: the work requires building new relationships (sales, partnerships, client management), the judgment component exceeds 60% of the role's time, or the domain expertise required would take years to encode into tooling. If all three are true, hire. If only one or two are true, look harder at the tooling option.

What about the cost of building internal tools?

Internal tools have real costs — development time, maintenance, and the opportunity cost of building them instead of something else. The difference is that those costs are front-loaded and finite, while hiring costs are ongoing and compound. A tool that takes 200 hours to build and 20 hours a year to maintain costs less over three years than a single full-time hire costs in three months.

Related reading

• The Two-Person Studio Model: Why We Don't Scale Headcount
• When to Automate and When to Hire
• What I Got Wrong About Hiring
• Building Systems Before You Need Them
• The Real Cost of Context-Switching