Before a company can say what a role should be paid, it has to settle a quieter question: how big is this job compared with the others? Is a payroll lead the same level as a talent acquisition partner? Is a senior test engineer above or below a business analyst? Job evaluation is the structured way of answering that. It looks at the role as designed, not at the person currently in it, and produces an internal order of jobs that later becomes grades and pay bands.

Market data alone cannot do this. Surveys tell you what a “Java developer” earns elsewhere, but many roles have no clean market match, and two companies’ “managers” can do very different work. Evaluation gives internal logic; benchmarking then attaches money to it.

The four classic methods

Textbooks describe four approaches, split by two questions: does the method look at the whole job or break it into factors, and does it compare jobs with each other or against a fixed scale?

MethodLooks atCompares againstTypical use
RankingWhole jobOther jobsSmall firms, quick first pass
Classification (grading)Whole jobPredefined grade descriptionsGovernment and public sector
Factor comparisonFactorsBenchmark jobs, in money termsRare today; mostly historical
Point factorFactorsA scored scaleMost mid-size and large companies

Ranking and classification: fast, and fragile

Ranking lists jobs from most to least valuable, often by comparing pairs. With fifteen roles in a startup, a founder and an HR lead can do it in an afternoon. It breaks down quickly. With 150 roles there are over 11,000 pairs, and the ranking tells you order but not distance: the gap between first and second may be huge or trivial.

Classification writes descriptions for a set of grades first, such as “Grade 3: performs established procedures under general supervision; resolves routine exceptions”, and then slots each job into the best-fitting description. India’s government pay levels work broadly this way. It scales better than ranking and is easy to explain, but grade descriptions are necessarily broad, and borderline roles become arguments about wording.

Point factor, worked through

The point-factor method breaks every job into the same set of compensable factors, defines levels for each, assigns points, and adds them up. Here is a stripped-down plan with 1,000 points in total.

  • Knowledge and expertise: up to 400 points
  • Problem solving: up to 250 points
  • Accountability and impact: up to 250 points
  • Working conditions: up to 100 points

Each factor has five defined levels, and each level earns one fifth of the factor’s maximum. Now evaluate two roles in an illustrative IT services company.

A payroll lead needs solid statutory knowledge (level 3, 240 points), solves mostly defined problems (level 2, 100), carries significant accountability because errors hit every employee’s pay and the company’s compliance (level 3, 150), and works in normal office conditions (level 1, 20). Total: 510.

A talent acquisition partner needs comparable knowledge (level 3, 240), handles more ambiguous problems such as hard-to-fill roles and negotiations (level 3, 150), has narrower direct accountability (level 2, 100), and the same conditions (level 1, 20). Total: 510.

If the grade structure places 450 to 549 points in Grade 4, both roles sit in the same grade despite being very different jobs. That is the method’s real value: it gives a defensible reason why unlike roles share a pay band, and an equally defensible reason when they do not.

The weights are a statement of values. A manufacturing company might give working conditions far more than 100 points; a consultancy might weight problem solving more heavily. Changing the weights changes the order of jobs, which is why the weighting step should involve senior leaders and not be left to a consultant’s default.

Proprietary systems

Most large Indian employers do not build a plan from scratch. They license one. The best known is the Hay method, now part of Korn Ferry, which scores jobs on know-how, problem solving and accountability using guide charts. Mercer’s International Position Evaluation and the WTW Global Grading System are other widely used frameworks. Their advantage is that the resulting grades map directly to the same consultancy’s salary surveys, so evaluation and benchmarking speak the same language.

Where evaluation goes wrong

Evaluating the person. A brilliant incumbent makes a role look bigger than it is. The test is to ask what the job requires of a fully competent replacement, not what the current person happens to do.

Title inflation. Managers rewrite job descriptions to push roles up a grade, adding “strategic” and “end-to-end ownership” to routine work. A panel that evaluates against actual responsibilities, with evidence, resists this better than one that scores the paperwork.

Hidden bias. Factor plans can undervalue work traditionally done by women, such as coordination, care or emotional labour, simply by not having a factor that recognises it. Equal pay for work of equal value depends on evaluation being fair in the first place, so the factor set itself deserves scrutiny.

Freezing the structure. Jobs change. A role evaluated in 2021 may have absorbed automation work, new compliance duties or a team. Re-evaluating when a role materially changes, and auditing a sample every year or two, keeps the hierarchy honest.

Written by the RP Strativa Team. The point plan, roles and scores are illustrative.

Leave a Reply

Trending