Job Levelling: The Foundation of a Fair Pay Strategy

Job levelling is one of those things most HR teams know they should have done properly. In my experience, roughly a third of mid-market organisations actually have. That matters, because almost every pay activity downstream, from benchmarking to pay reviews to difficult conversations with your people, depends on it.

‍

With 42% of UK employers planning flat-rate pay awards in 2026, the questions employees ask about pay are getting harder, not easier. A levelling framework is what gives you the evidence base to answer them.

‍

In this guide, I'll cover what job levelling is, how to build a framework, and how to connect it to live market data

‍

HR DataHub's levelling feature is discussed in this article; HR DataHub is its publisher..

At a glance

• Job levelling organises every role into a consistent hierarchy based on scope, accountability, problem complexity, and autonomy, providing the structure that fair pay decisions depend on.
• Without a levelling framework, salary benchmarking compares roles that aren't equivalent, producing data that misrepresents the market in either direction.
• Most mid-market UK organisations manage well with five or six levels; operationally heavy sectors like retail and logistics may need more granularity.
• The two most common failure points are overgrading during deployment, driven by management politics, and grade drift after implementation.
• Job levelling is the practical first step toward pay transparency, giving you a defensible basis for connecting individual roles to external market data.

What is job levelling?

Job levelling is the process of organising every role in an organisation into a structured hierarchy based on factors like scope, accountability, problem complexity, and autonomy. It brings consistency to pay decisions, career progression, and salary benchmarking.

‍

Job levelling differs from job evaluation in that evaluation assigns a numerical score to each role's size, whereas levelling places roles in the right position within the hierarchy without quantifying absolute value.

Why job levelling is the foundation of a fair pay strategy

The practical impact runs further than most teams expect. Without levelling, almost every downstream pay activity, from benchmarking to pay reviews to transparency reporting, is built on shaky ground.

Fair, consistent pay decisions

Levelling removes individual interpretation from pay decisions and replaces it with a shared standard. Without it, pay ends up being shaped by who argues loudest in a calibration meeting rather than by the actual scope and complexity of the role.

‍

The politics are the hardest part of any levelling project. Line managers will push for a higher level for their teams because a higher level means a higher pay range.

‍

If you don't believe in your methodology and you don't push back consistently, the system breaks down. Some managers argue harder than others, and that inconsistency is exactly what creates equal pay risk.

‍

A solid framework gives HR the evidence base to hold those conversations. You can point to the criteria, show where the role sits, and explain the decision in a way that holds up to scrutiny, whether from the employee, the manager, or a pay equity audit. Our guide to pay equity covers how to run that audit once your levelling architecture is in place.

Accurate salary benchmarking

Without consistent levelling, credible external benchmarking is almost impossible. You're comparing roles that aren't truly comparable, and the data comes back either significantly above or below the actual market.

‍

What often happens without levelling is that pay ends up guided by whoever has the most confidence in the room: the hiring manager, the agency, or the candidate. Levelling gives you the defensibility to push back. You have a framework to point to rather than just a gut feel.

‍

The most important principle: level first, benchmark second, and keep those processes entirely separate. Our salary benchmarking guide walks through how to connect your levels to live UK market data once you're confident in the framework.

‍

One nuance worth flagging: being on the same level doesn’t mean you’re paid the same. Two roles at the same level can carry very different market rates. An AI engineer and a software engineer might sit at the same internal level, but the AI engineer carries a meaningful market premium right now. Your benchmarking process picks that up; your levelling process doesn't.

Career clarity and employee retention

A levelling framework gives employees a clear picture of what progression inside the organisation actually looks like. That matters a great deal in sectors where the job is broadly similar across employers and the barrier to switching is low.

‍

When someone can see what a step up involves, how scope and accountability change at each level, and how the pay range shifts, they have something concrete to work toward. Without that clarity, career conversations tend to be vague and unsatisfying, and people look elsewhere.

‍

In operationally heavy sectors like retail, logistics, and transport, pay differentiation between employers is shrinking as the National Living Wage continues to lift the floor. The progression framework and the clarity of the internal structure are often the retention levers you have left when you can't compete purely on base pay.

Pay transparency compliance

Job levelling is, in my view, the first practical step any organisation needs to take before it can move toward meaningful pay transparency.

‍

The UK's direction is clear, even if the legislation is still moving. The Employment Rights Act 2025 requires employers with 250 or more employees to publish gender equality action plans alongside their gender pay gap reports, with voluntary compliance from April 2026 and mandatory from Spring 2027. A government call for evidence on salary range disclosure and pay information rights closed in June 2025; further developments are expected.

‍

But the real reason to get levelling in place isn't the legislative timeline. If you can't explain to an employee how their level was determined and how it connects to the external market, publishing a pay range doesn't mean much to them. Levelling is what makes transparency meaningful rather than just procedural.

How HR DataHub assesses job levels

HR DataHub's levelling tool assesses roles against four dimensions: Autonomy and Supervision, Scope of Impact, Problem Complexity, and People/Resource Accountability. Each job posting processed by the platform is evaluated against these criteria using a language model trained on a manually levelled dataset. The output is a level assignment alongside a confidence score of High, Medium, or Low.

‍

High confidence means all four dimensions are clearly evidenced in the job posting. Medium means two or three dimensions are clear and the rest are inferred. Low means there are significant gaps in the spec and the level is a best estimate.

‍

I want to be upfront about what our approach doesn't do. Our methodology depends on the quality of the job posting. Thin or generic specs give us less to work with, and the confidence score reflects that.

‍

We also don't factor in organisational context in the way a formal evaluation system does. Other frameworks consider company size, international presence, and supply chain complexity. We don't.

‍

What we do provide is a live market connection and a framework usable without a specialist reward team or a six-figure implementation budget. For mid-market organisations that need a credible starting point, that's often exactly what's required.

The six levels in practice

HR DataHub uses a six-level framework spanning the full hierarchy from entry-level roles through to senior leadership.

‍

‍

The framework also runs parallel People Leader and Individual Contributor (IC) tracks at Levels 2 through 5. This matters for benchmarking: an Engineering Manager and a Principal Engineer can sit at the same level if scope and impact align, and both paths are treated as equivalent for compensation purposes. It means you can pay a senior specialist competitively without forcing them into management, which the market increasingly reflects anyway.

‍

Five or six levels is the sweet spot for most mid-market organisations. The right number depends on how many distinct layers of decision-making authority exist in the business, not on what any framework provider recommends.

How levelling maps to industry standards

If your organisation already uses Willis Towers Watson, Hay, Mercer, or Radford, HR DataHub's levels are mapped to sit alongside those frameworks. The intention isn't to replace an established system. Once you've invested in WTW or Mercer, the benchmarking data is calibrated to it, and that alignment is a significant part of what you're paying for.

‍

What the mapping lets you do is use HR DataHub as a live market data layer alongside your existing framework rather than choosing between the two.

‍

‍

For organisations that can't yet justify a WTW or Mercer implementation, HR DataHub's levelling can function as the standalone framework. It won't be as formally calibrated as a traditional job evaluation system, but it gives you a solid architecture without the significant upfront cost.

Building a job levelling framework: the key steps

Most of the difficulty in building a levelling framework isn't technical. It's political and organisational. Here are the steps I'd recommend, and where things most commonly go wrong.

Step 1: Define your career tracks

Before you assign a single level, sort out your organisational structure. This is the step most teams want to skip, and it's the one that causes the most rework later.

‍

Look honestly at your org. If 80% of your roles are doing essentially the same job under five different job titles, that needs resolving before levelling can work. Harmonise job titles first and map out the career tracks you actually have, deciding which functions need their own track and what the general hierarchy looks like across them.

‍

Rolling out a levelling system on a flawed structure means four or five iterations of the exercise rather than one clean pass. The prep work up front is the biggest time-saving investment you'll make in the project.

Step 2: Set level criteria

With a clean org structure in place, define the criteria that will distinguish one level from another. These should cover what actually determines a role's scope: the autonomy to make decisions, the breadth of responsibility, the complexity of problems faced, and the accountability carried for outcomes.

‍

Write the criteria in plain language that a non-specialist can follow. They need to work in a conversation with a line manager, not just in a Reward team meeting. If a manager can't follow the reasoning when you explain why their team member is a Level 3 rather than a Level 4, the criteria aren't clear enough.

‍

Set the criteria before you look at the people. Level the roles, not the individuals. Working backwards from a specific person introduces exactly the bias the system is designed to eliminate.

Step 3: Map your existing roles

With criteria defined, apply them consistently across every role. This is where the politics typically arrive.

‍

Managers will argue for a higher level. If you let the outcome vary based on how hard someone pushes, you lose the consistency the whole project is built on. Hold the line.

‍

A grading panel, where decisions are reviewed by more than one person before being confirmed, helps maintain consistency and creates a checkpoint against grade inflation. Document every decision and the rationale. If a decision is challenged later, that documentation is your evidence base.

Step 4: Connect to salary bands and benchmarks

Once roles are levelled consistently and you're confident in the output, connect the framework to the external market. This is where a live benchmarking tool becomes genuinely useful.

‍

Map each level to external market data. Find the median for each role at that level, identify any premiums by role type or location, and use that to build salary bands that reflect the actual market. Our salary benchmarking guide covers how to do this in detail.

‍

If you're building band structures for the first time, our salary bands guide covers how to set range widths and midpoints at each level.

‍

Don't overgrade a role to reach a market data point you prefer. If the benchmark comes in lower than expected, that's market information.

‍

If there's a business reason to pay above median, make it a conscious decision: name the justification, document it, and pay upper quartile deliberately. Overgrading to chase market data corrupts the framework and creates grade drift that's time-consuming to unwind.

FAQs

What is the difference between job levelling and job evaluation?

Job levelling and job evaluation are related but distinct processes. Job levelling places roles in a relative hierarchy based on scope, complexity, and accountability. Job evaluation takes that further by assigning a numerical points score to each role using a structured analytical framework.

‍

Systems like Hay, Mercer IPE, and WTW's global grading use detailed factor analysis to produce those scores. Job levelling is generally faster to implement and more accessible for mid-market organisations; job evaluation provides greater precision for large organisations managing substantial reward programmes.

How many levels does a job levelling framework need?

A job levelling framework for most mid-market UK organisations needs five or six levels to cover the full span from entry-level to senior leadership. Organisations with significant operational workforces, such as those in retail, logistics, or transport, may need more granularity. Smaller or more specialised organisations, particularly in technology or professional services, can often manage with four or five.

‍

The right number reflects how many distinct layers of decision-making authority exist in the business. When you start thinking about Willis Towers Watson or a similar formal system, 500 employees and above is roughly where that level of investment begins to make sense.

Does job levelling apply to small organisations?

Job levelling benefits any organisation above around 50 employees. The key question at smaller sizes is proportionality.

‍

A 150-person business doesn't need a six-figure implementation, but a lightweight framework that groups roles by scope and accountability gives you a fair basis for pay decisions and career conversations. HR DataHub's levelling tool is designed to give smaller and mid-market organisations a solid framework without the investment that traditional survey-linked systems require.

How often should a job levelling framework be reviewed?

A job levelling framework needs ongoing governance, not just a periodic overhaul. Grade drift, where roles gradually migrate upward through small individual decisions, is a consistent problem even in well-run organisations. Most large organisations run architecture reviews every three years or so, supported by grading panels that review individual decisions as they're made.

‍

Without that governance, the framework inflates quietly and resetting it is significantly more painful than maintaining it. Think of it the same way you'd think about pay drift: it needs an active management process, not just a calendar reminder.

Start with the right foundations

Levelling is where the hard work of a fair pay strategy actually begins. No framework survives contact with a real organisation without some pragmatic adjustment, and that's fine. What matters is that it's consistent enough to be defensible, practical enough to be used, and connected to live market data so that every level has an external reference point.

‍

If you want to see how HR DataHub connects your job levels to real-time UK salary data, book a short demo and we'll walk you through how it works for your sector and size.