HR operating model: why a new org chart isn't a new model

At a glance

Most HR reorganizations move boxes and change titles. Six months later the same decisions take the same amount of time, and nobody can say what improved. The org chart changed. The operating model didn't.

An HR operating model is the structural blueprint for how an HR function organizes its people, processes, technology, and governance to deliver value to the business. It decides who owns what and how work actually gets done. Five established model types exist, each suiting a different kind of business, and your own employee feedback data is the fastest way to tell whether the one you have is working.

An HR operating model answers questions about authority and delivery, not reporting lines. Who approves an off-scale pay request. Whether a local team can buy its own learning platform. Where an employee goes first with a question.

Those answers are the model. Change them and you have changed how HR works. Redraw the chart without changing them and you have changed almost nothing.

Whatever type you land on, the model has to resolve four things:

Governance is the one most redesigns skip. It is also the one that determines whether the new model feels different to the people using it.

These get used interchangeably in the same meeting, which is how redesigns go wrong. Your HR strategy is what you intend to achieve with people. Your operating model is how the function is organized to deliver it.

The org chart is the thinnest of the three. AIHR makes the point directly in its guidance: reporting structures change often as roles and teams shift, and a change in reporting line is not a change in operating model. Treat the model as a deep change in how services are delivered, not a box-moving exercise.

Five model types are broadly established in HR practice. They sit on a spectrum from highly centralized to highly decentralized, and the right one depends on how centralized the wider business already is.

The most prevalent model, and the one pioneered by Dave Ulrich in the late 1990s. Work splits across shared services for day-to-day transactions, centers of excellence for specialist expertise, HR business partners advising line managers, and an HR leadership team. It is often called the three-legged stool.

It positions HR as a strategic partner while capturing efficiency through shared technology. The known failure mode is business partners going native, identifying with the business rather than with HR, which creates an internal divide. Mercer data cited by AIHR puts prevalence at roughly 75% in North America and 44% in Europe.

HR organizes around specialisms such as recruiting, compensation, and learning, all reporting to a central HR leader. Decision-making runs top down. This is the common shape for smaller organizations.

It builds specialist depth quickly and is cost-efficient. It also tends toward silos, and the business often experiences HR as fragmented with too many separate touchpoints. It struggles as organizational complexity grows.

A close relative of the business partner model, with business partners, shared services, and centers of expertise still present. The difference is localization: each line of business crafts its own strategy within shared principles, and can run its own technology where genuinely needed.

It pushes decisions closer to the business and moves faster. It also duplicates services, needs heavy governance to keep front and back aligned, and can get expensive if that governance slips.

Structurally similar to front-back, with one meaningful difference in direction of influence. In front-back, the hub sets strategy and the spokes localize it. In hub and spoke, the spokes influence the hub.

The hub provides shared resources, consistency, and technology, while spokes localize by geography, business unit, or vertical. It scales well and speeds up learning across the function. It fails when spokes do not know when to draw on the hub, and duplication creeps back in.

The fully decentralized option. HR operates as independent lines of business by region or business unit, each setting strategy within agreed principles, with corporate HR mostly coordinating.

Accountability and regional responsiveness are the payoff. Without tight governance it produces conflict, duplication, and rising cost, and it has a tendency to become everything to everyone rather than holding a focused direction.

Agile HR is a direction of travel rather than a sixth model. It describes HR moving toward multi-disciplinary, project-based work with less functional separation, drawing on automation and data skills alongside traditional HR capability.

You will also hear target operating model, or TOM. That is simply the model you are working toward rather than the one you have. Transitions typically run somewhere between six months and six years depending on size, complexity, and readiness, which is why the target version gets discussed far more often than the current one.

Design runs in five stages. The order matters, because most of the expensive mistakes happen when structure gets chosen before service expectations are defined.

Start with evidence about the model you already have, not a blank page. You want to know where decisions stall, which services managers avoid, and where employees give up and route around HR.

This is where employee surveys earn their place in a structural project. Manager experience of HR service delivery is a direct read on whether the current model works, and it is rarely captured anywhere else.

List what HR is actually accountable for delivering, to whom, and at what standard. Separate transactional services from advisory and specialist work, because they belong in different parts of any model.

Do this before choosing a structure. The service list is what makes the structural choice in Stage 3 an argument about fit rather than a matter of taste.

Match the degree of centralization in your model to the degree of centralization in the business. A highly centralized organization is well served by a centralized model, and a group of genuinely independent businesses will fight one.

Size and budget narrow it further. Smaller organizations generally do well with a functional model where a few people cover multiple roles. Let the business strategy lead the choice, since a model designed in isolation from it tends not to survive contact with the business.

Write down who decides what, and where a decision escalates. Name the owner for each service in the Stage 2 list, and set the standing review points where HR and the business look at performance together.

Map the capabilities each role needs while you do this. A common reason operating models underperform is that organizations define the way they intend to work without confirming they have people who can work that way.

Decide the evidence you will judge the model on before you launch it. Useful signals include manager satisfaction with HR service, time to resolve requests, and whether employees know where to go.

Then measure it on a cadence rather than at a single post-launch moment. A pulse survey aimed at managers is a practical way to catch a model that is drifting before the next reorganization gets proposed.

This is the part most operating model guidance leaves out. Consultancies will describe the models and the change program, but few tell you how to read the evidence in front of you to work out whether structure is your problem.

Start with how HR is positioned, because it shapes everything else. In SurveyMonkey research among 269 HR professionals, just over half, 53%, said leaders in their company view HR as vital to a culture of success, while the rest said leaders see HR's role as strictly operational. 

That split tracked with real differences in influence: 68% of HR professionals reported higher expectations from leadership since the pandemic, rising to 73% where HR is seen as vital.

The same study found that where HR sits structurally changes what employees are willing to tell it. 

Overall, 72% of HR professionals said they were concerned about employees providing open and honest feedback about their experiences at work. That concern reached 80% at companies where HR is not seen as vital, against 64% where it is.

Data fragmentation shows up the same way. Among companies reporting a unified view of employee data, 81% said they ask employees for DEI data, compared with 66% where data is scattered and siloed. Siloed information also appeared among the top challenges HR teams named in collecting feedback at all.

Two more figures are worth holding next to each other. Only 32% of HR professionals said their company had been very successful at addressing turnover, and 29% said they do not get enough employee feedback. A function that cannot hear its workforce clearly is poorly placed to explain to leadership why people are leaving.

Methodology: SurveyMonkey research was conducted between August 25 to September 5, 2023 among 269 human resource professionals. Respondents were selected from an online non-probability panel.

The following is a diagnostic approach rather than a research finding. It maps the documented limitations of each model onto signals you can pick up in manager and employee feedback, so you can tell a structural fault from a local one.

A single signal is not a mandate to reorganize. A pattern that holds across business units, and persists after the local fix, is the case for looking at the model.

Consider a 900-person software company with nine people in HR, organized functionally. A recruiting lead, an operations lead running payroll and systems, and a small L&D group all report to the head of HR.

It worked for 300 people. At 900, the head of HR is personally the business partner for every function, and engineering leaders have started hiring their own coordinators to get faster answers.

In Stage 1 the diagnosis is specific rather than atmospheric.

Manager feedback shows a consistent complaint about response times on anything requiring a judgment call, and it holds across three business units.

Employee feedback shows people asking their own managers policy questions because they do not know where else to go.

Stage 2 separates the work: transactional payroll and benefits queries, advisory support for managers, and specialist compensation and L&D design.

Stage 3 points toward a business partner model, because the company is centralized and wants consistent practice, but at nine people it moves in that direction rather than building three full pillars.

Two senior business partners get hired, and self-service takes the transactional load.

Stage 4 names owners and sets an escalation path.

Stage 5 sets the measure: manager-reported time to a usable answer, tracked quarterly, with a repeat of the same diagnostic questions used in Stage 1 so the comparison is honest.

Choosing between five models is not really the hard part. The hard part is knowing which problem you are solving, and that answer sits with the managers and employees living inside the current structure.

SurveyMonkey gives you that input. Use survey software to run the Stage 1 diagnostic and the Stage 5 measurement with the same questions, so you can show leadership what changed rather than asserting it. Teams designing for the HR function can start from the templates and question sets built for human resources teams.

Once the model and its governance are settled, build the measurement into the employee lifecycle by choosing the HR survey questions that track service quality, manager support, and engagement year after year.