22 Jul 2026

Fixing the broken job evaluation process

Traditional job evaluation systems are no longer fit for purpose. Sticking with them is now the riskier choice.

Job evaluation is supposed to give you confidence that pay decisions are accurate, consistent and defensible. In practice, most job levelling and job evaluation systems do the opposite. They eat up enormous effort, produce outcomes that are hard to explain and quietly drift as the organisation changes around them.

With pay transparency legislation advancing, and equal pay litigation on the rise, the gap between what these systems were built for and what organisations now need has become impossible to ignore.

Outdated vendor systems

Traditional job evaluation tools were designed last century, for stable hierarchies and slow-moving job structures. Talk to the organisations using them and you hear the same frustrations: they're complex, time-consuming and demand significant resource just to stand still. The methodologies are often black box. They're difficult to explain, easy to game and hard to defend when someone asks why a role landed where it did.

They also depend on scarce expertise. Running these systems properly requires deep subject-matter knowledge and often paid training, which makes them nearly impossible to scale. What you end up with is a one-size-fits-all framework administered by a small group of specialists who become a permanent bottleneck and a long-term dependency on the consultancy whose methodology it is.

Trading one problem for another

Many organisations have responded by building their own levelling frameworks in spreadsheets, with job evaluation judgment sitting in one or two experts' heads. These offer flexibility but they struggle elsewhere. Without systematic governance, levelling drift and grade inflation go undetected.

Roles creep up over time and no one can say when or why. Consistency collapses at scale too: a framework applied by different people, in different regions, using different interpretations, stops being one framework at all. There's also a single point of failure question worth asking honestly. What happens when your levelling expert leaves?

Evaluation process is broken

Even where the methodology is sound, the process feeding it often isn't. Job evaluation depends on accurate job information, which is notoriously poor. Ask any HR team about the endless back-and-forth with managers just to clarify a role's scope before it can be assessed.

It slows down hiring and grading decisions, and one estimate suggests that around 85% of levelling and evaluation exercises have to be redone because the input information was inaccurate.

The data problem makes it worse. Levelling frameworks sit in one place, job profiles and descriptions in another, architecture in a third and outcomes in a fourth, often behind separate logins. Disconnected data makes governance harder and calibration slower and small errors compound until the structure no longer reflects reality.

Work changes faster than systems

All of this might be tolerable if jobs were stable. They're not. Organisations restructure near-constantly and AI is expected to reshape the skill requirements of almost every role.

Levelling built for static job models simply can't keep up with work that is continuously redesigned. A framework that takes months to apply and gets revisited every few years will always be describing an organisation that no longer exists.

What fit-for-purpose looks like

The requirements are clear from the failures. Job evaluation needs to be transparent rather than black box, so decisions can be explained to employees, panels and courts. It needs governance and audit trails built in rather than bolted on, capturing factor scores, rationale and ownership for every decision.

It needs to be simple enough for non-specialists to apply consistently, so an HR business partner or line manager can complete a defensible evaluation without weeks of training and the practice can scale without an army of experts.

It also needs connected data, so frameworks, job descriptions, architecture and skills sit in one place and roles can be calibrated side by side in real time rather than reconstructed in Excel after the fact.

Above all, the methodology needs to live in a system rather than in someone's head, so the knowledge stays when people move on and the framework keeps pace as the organisation evolves.

Status quo is now the risky option

For years, the safest choice in job evaluation was to keep doing what everyone else did. That logic has inverted. Systems that can't produce an audit trail, explain their decisions or keep pace with change are now a source of exposure, not protection.