7 min readPriya Raghunathan

Updated on

Job architecture without a consulting project

Levels, titles and families, built in six weeks by two people.

Job architecture without a consulting project

Job architecture sounds like a project with a steering committee. At under a thousand people it is a spreadsheet, a few hard conversations and about six weeks.

You need it before you can do pay bands, career paths, internal mobility or pay transparency compliance properly. Everything else waits on this.

Start with families, not levels

Group work by the kind of contribution, not the department. Engineering, product, design, commercial, operations, corporate functions. Six to ten families is normal. Departments change every reorg; families do not.

Define levels by scope, not by years

Four to six levels for individual contributors, three to five for management. Describe each by the size of problem someone owns, the amount of ambiguity they absorb, and who they influence. Never by years of experience, which is a proxy for nothing and creates age discrimination exposure in several jurisdictions.

Map people before you publish

Take the whole population and place them. This is where the architecture meets reality. You will find three or four people who do not fit, two titles that mean different things in different offices, and one team where everyone is somehow senior.

Resolve those before you announce anything. A framework published with obvious anomalies in it will be judged on the anomalies.

Separate title from level

Titles are market facing and emotional. Levels are internal and structural. Keeping them separate lets you give a sensible external title without distorting your pay bands, and lets you fix title inflation later without demoting anyone.

Publish the ladder, hold the mapping

Show everyone the levels and what each requires. Individual placements stay between a person and their manager. This is the balance that gets you the benefit of transparency without a week of public comparison.

Then leave it alone for two years. Architectures that get revised constantly teach people that levels are negotiable, which is the opposite of the point.

Sources

  1. 01Directive (EU) 2023/970 on pay transparencyEUR-Lex
  2. 02Research and benchmarkingSHRM
  3. 03People and organizational performance insightsMcKinsey and Company
  4. 04Human resources research and insightsGartner
ShareLinkedInXEmail

Read next in compensation and benefits

  • The Two-Speed Workforce Plan: Separating Core Capacity From Elastic Talent

    Traditional headcount planning fails when market volatility collides with multi-year business goals. A two-speed workforce model splits predictable core operations from elastic talent networks to protect capacity while controlling fixed employment costs.

  • Modeling Headcount for AI-Augmented Roles

    Traditional headcount planning relies on fixed output ratios per employee. When artificial intelligence alters task completion speed, workforce planners must rebuild capacity models using task-level decomposition.

  • Scaling Skills-Based Architecture Beyond the Pilot Phase

    Moving from an opt-in talent marketplace to a skill-based enterprise requires re-engineering job architectures, navigating European co-determination laws, and aligning with US pay transparency rules. Here is how forward-looking organizations are executing the transition.

  • Why headcount planning fails without structural design

    When you automatically backfill roles, you bake in the inefficiencies of the past. Learn how to identify structural debt, balance global jurisdictions, and design reporting lines that drive actual revenue.

  • Sequencing workforce reductions when headcount targets drop

    As boards mandate margin preservation next quarter, talent leaders must execute phased reductions across external spend, flexible labor, and permanent staff without triggering compliance failures in North America or Europe.

The newsletter

One edition roughly every two weeks: new articles, and what changed in hiring that is worth your time.

Back to all articles