Should HR be centralized, or handled by senior staff directly?

Why not an internal HR department

The HR essay is direct. HR is overhead by design, and that is correct: revenue is created when skilled professionals sell and deliver work, and anything that does not directly contribute to those activities is overhead. In a healthy boutique, full-time employees should be billable, overhead should be kept lean, and non-billable functions should be fractionalized and outsourced wherever possible, which is why disciplined firms treat HR the way they treat finance, IT, legal and benefits administration: essential, but not internal.

The essay gives the reasons an internal department rarely makes sense. The volume of work does not justify the headcount, the complexity does not warrant the cost, and the dollars are better spent on billable talent or growth. Founders who internalize HR too early usually reach the same conclusion, just more expensively. The overhead answer on this site places HR with operations, marketing, finance, IT and legal in the overhead block, funded from gross margin before sales and EBITDA.

The power chapter of the 2020 book describes the governance model most scaling boutiques adopt, and it points the same way: a managing partner with a small staff who outsources most of the overhead functions, naming IT, HR and legal.

Why not leave it to senior staff

The HR essay describes what happens when people decisions are made case by case by whoever is closest. In the first era, HR was manual, intuitive and reactive. Hiring was driven by urgency rather than planning, development was informal, promotions came late, and succession was acknowledged and ignored. Selection prioritized speed over quality, onboarding was compressed or skipped, deployment filled gaps rather than matching skills, and founders filled the gaps themselves, so the firm became more dependent on them, not less.

That is what senior staff handling HR directly tends to produce. Each leader makes reasonable decisions for their team, and the firm ends up with different hiring standards, different review practices and different promises about pay and promotion in each corner. The essay warns that contractor-heavy and generalist models introduce inconsistency where precision matters most, and the same is true of HR distributed across partners. In a small firm the costs arrive fast: a single mis-hire is paid for twice, turnover is contagious, and single points of failure trap the founder in the day-to-day.

Split the work by what it is

In effect the essay separates HR into three kinds of work, and the answer to the question follows from putting each in the right place.

Compliance and administration. Policies, documentation, benefits administration and regulatory compliance are necessary, and neglecting them creates legal, financial and reputational risk. But they are defensive, and none of them improve how the firm scales. The essay says this work should be handled by fractional specialists, the way firms rely on benefits brokers or outside counsel: done correctly, consistently and out of the head of the founder. Founders buy it as a total service, recruiting support, onboarding, performance management, compensation guidance, benefits and compliance in one bundle, and the essay calls that rational.

Governing the talent supply chain. The strategic value of HR is in how talent flows through eight stages: recruit, select, onboard, deploy, develop, retain, promote and succession. In the third era the essay assigns roughly 80 percent of this to AI: continuous capacity forecasting from projected demand, pattern analysis across recruiting, selection and attrition, time-to-productivity modeling, deployment optimization, early detection of attrition and burnout risk, promotion readiness signals and visibility into succession risk.

Human judgment. The remaining 20 percent is judgment and tradeoffs, cultural reinforcement, coaching and accountability, leadership decisions and contextual interpretation. The essay says the humans supplying HR expertise need not be internal, and that the fractional partner should validate the outputs, apply benchmarking across comparable professional services firms and handle compliance consistently. As an inference, coaching, accountability and culture are also the daily work of whoever manages people, so that part of the 20 percent sits with senior staff by necessity. The difference from leaving HR to them is that they now act inside one system, with one set of standards and one view of the data, rather than inventing their own.

The essay makes one more point that bears on the question. HR does not create forward visibility on its own; it receives it from how the firm sells, prices, packages and delivers work. When demand becomes predictable, capacity planning moves from speculation to projection and hiring from reactive to intentional. As an inference, that is another reason HR cannot be left to each senior person: the signal it depends on comes from the whole firm, and only a system that sees the whole firm can use it.

So what is centralized

As an inference, the useful answer is to centralize the system and distribute the judgment. Centralize the standards, the data, the policies and the compliance, through the AI layer and a fractional specialist. Centralize the decisions that must be consistent across the firm: the power chapter of the 2020 book puts salaries and bonuses with a compensation committee made up of the board, the managing partner and the executive team, not with individual partners. Distribute the coaching, the feedback and the day-to-day leadership to the senior people who do the work, and hold them to the shared standards.

The essay adds a condition on who supplies the fractional layer. Generalist providers without deep professional services specialization cannot add value on top of AI, and their judgment collapses back into compliance and commentary.

What we do not prescribe

Collective 54 names no HR provider, sets no headcount at which an internal HR hire becomes sensible, and publishes no ratio of HR cost to revenue. The published positions are HR as overhead by design, fractionalization over an internal department, compliance as necessary but not strategic, the eight-stage talent supply chain, the 80/20 division between AI and human judgment, specialization over generalist providers, and compensation decided by a committee.

When this answer flips

If the firm is small enough that the founder manages everyone, the founder is the senior staff, and the risk the essay describes is the founder becoming the single point of failure; the fractional layer matters more, not less.

If you already employ someone in an internal HR role, the model does not require removing them, but as an inference their work should move from administration, which the fractional and AI layers can carry, toward the judgment and coaching support the 20 percent describes.

And if a regulatory or client requirement demands an internal owner for certain people processes, keep that ownership narrow and place everything else in the fractional model.

The short answer

Neither an internal HR department nor HR improvised by each senior person. The published position is that HR is overhead by design, that full-time employees should be billable, and that non-billable functions should be fractionalized, so an internal department rarely makes sense at boutique scale, while leaving HR to senior staff produces the reactive, inconsistent people decisions that kept earlier firms small. Split the work instead. Give compliance and administration to a fractional specialist with professional services expertise. Let AI carry roughly 80 percent of governing the talent supply chain, from capacity forecasting to attrition and succession risk. Keep the human 20 percent, coaching, accountability, culture and leadership decisions, with the senior people who manage the work, acting inside one set of standards, and put pay decisions with a compensation committee. Collective 54 names no provider and sets no threshold for an internal HR hire.