Skills-based organizations: the shift from jobs to skills (Deloitte, Mercer, IBM)

60-Second Summary

The 'skills-based organization' (SBO) replaces the job as the primary unit of work with skills, projects, and capabilities. People are matched to work via what they can do, not what their job title says.
Deloitte's Skills-Based Organization research (2022–2024): 63% of executives believe skills-based work creates more value; 98% want to move in that direction; only 20% have made meaningful progress.
Best-publicised examples: Unilever's 'flex experiences' platform (internal gig marketplace), IBM's skills inferencing (no resumes, AI-derived skills), Schneider Electric's Open Talent Market, Mercer's Workday-based skills cloud.
Honest caveats: the evidence is largely vendor-driven. Reductions in time-to-fill are real (30–50%); productivity claims are softer. Skills taxonomies decay fast (~20% per year) without continuous curation.
Start small: skills inventory for one function, internal marketplace for one type of project, hiring without job descriptions for one role family. Don't try to dismantle jobs across the company in year 1.

The job as we know it — a fixed bundle of duties, attached to a fixed person, sitting in a fixed box on the org chart — was invented for industrial-era work. In knowledge work, it increasingly fits badly. The Skills-Based Organization is the most-discussed alternative in 2024–2026, and the most-misunderstood.

What an SBO actually is

Job-based vs skills-based

Traditional job-based org

Person ↔ Job (1:1)
Hiring against job descriptions
Compensation tied to job grade
Career = promotion ladder within job family
Org chart shows boxes (jobs) and lines (reporting)

Skills-based org

Person ↔ Skills ↔ Work (many-to-many)
Hiring against required skills (sometimes no JD)
Compensation tied to skills depth + market value
Career = skill acquisition and project portfolio
Org map shows skills clouds, project networks, capabilities

An SBO doesn't necessarily eliminate jobs — it makes them less load-bearing. Hiring, deployment, comp, development, and mobility are all anchored to skills first, jobs second.

Why now: the four forces

AI compression of tasks. AI does the first draft of many job duties; what remains is task fragments that don't cleanly map to traditional jobs.
Skills half-life shortening. World Economic Forum estimates technical skills now have a ~2.5-year half-life. Job-based hiring locks in last decade's skill profile.
Talent scarcity for specific skills (vs jobs). AI/ML, cybersecurity, climate, regulatory affairs — companies can't hire enough; rotational and internal-marketplace models are the only way to staff.
Workforce expectations. Gen Z and millennials report higher job satisfaction in roles framed around capabilities + projects than fixed JDs (Gallup 2023, Deloitte 2024 surveys).

The evidence (and the hype)

Read the evidence carefully

Most published outcomes come from vendors (Workday, Eightfold, Gloat, SAP SuccessFactors) and consulting firms (Deloitte, Mercer, BCG) with skin in the game. Independent peer-reviewed evidence is thin. The reductions in time-to-fill and increases in internal mobility are well-documented; the productivity and innovation claims are softer.

Claim	Source	Honest reading
63% of execs believe SBO creates more value	Deloitte 2022 Skills-Based Organization study	Self-reported belief, not measured outcome
30–50% reduction in time-to-fill via internal marketplace	Unilever, Schneider, Mercer client data	Genuine and replicated; the strongest evidence
3.5x more agile workforce	Deloitte composite metric	Composite, not directly observable; treat as directional
98% want to move toward SBO	Deloitte 2024	Aspiration, not action
Skills inferencing accuracy 70–85% (IBM)	IBM Watson Career Coach data	Real but precision varies by job family; long-tail skills under-detected

Real examples: Unilever, IBM, Schneider

Three published case studies

1
Unilever — Flex Experiences
Internal gig marketplace. Employees take short assignments (10–20% of time) outside their formal role. Reported >540k hours logged, ~1M USD saved in external hiring. Started 2018; expanded post-COVID.
2
IBM — skills inferencing
Resumes deprecated for internal mobility. AI infers skills from project history, learning activity, code commits. Used in promotion, lateral moves, and external hiring. ~75% of IBM's open roles can be matched against inferred internal skills.
3
Schneider Electric — Open Talent Market
Internal marketplace for full-time moves, projects, mentoring, learning. Powered by Gloat. 2.5M+ matches since launch; 18,000+ internal moves attributed. Cited by Josh Bersin as the most mature large-scale deployment.

The building blocks

Skills taxonomy: a canonical list (200–2,000 skills depending on company size), maintained continuously. Open-source starts: Lightcast Open Skills, ESCO, O*NET.
Skills inventory: who has which skills, at which level. Sources: self-attestation, manager validation, certifications, project history, AI inference from artefacts (code, docs, presentations).
Skills-based hiring: JDs become 'required skills + project descriptions'. Some companies drop JDs entirely for internal moves.
Internal marketplace / talent platform: where work meets people. Gloat, Eightfold, Fuel50, SAP SuccessFactors Opportunity Marketplace, or built-in-house.
Skills-based comp: pay differential for in-demand skills, regardless of job grade. The most politically hard piece — touches the comp philosophy directly.
Skills-based learning: L&D programs designed around skill gaps in inventory vs. demand from work, not generic curricula.

How to start without burning the org down

Pick one function (often Engineering, Data, or Product) and one use case (often internal mobility or project staffing). Don't try to skills-ify the whole company.
Build the taxonomy from existing sources (Lightcast, ESCO) rather than from scratch. A bespoke taxonomy is 18–24 months of work that can be skipped.
Start with self-attested skills + manager validation. AI inference is year-2, not year-1.
Run an internal-marketplace pilot for 6 months with 10–20% time on gigs. Measure mobility rate, time-to-fill on internal roles, attrition of marketplace users vs control.
Don't change comp in year 1. Skills-based pay is the hardest piece and the most likely to fail. Get the other building blocks proven first.
Don't dismantle jobs. Make them less central. Keep the org chart; add the skills overlay.

FAQ

Frequently asked questions

Does this work in unionised environments?

It's harder. Unions typically negotiate around job classifications; SBO disrupts that. Several European unions have engaged constructively with skills-based pilots (notably IG Metall in Germany), but expect 12–18 months of consultation before any meaningful change.

What about job titles for external markets?

Companies running SBO internally still maintain external job titles for hiring, LinkedIn, comp benchmarking. The shift is internal-first; external markets aren't ready for 'Senior Distributed Systems Skills Cluster Member'.

Is this just internal mobility re-branded?

Internal mobility is one application. SBO goes further: it changes how work is described, how people are paid, how careers are mapped, and how external hiring is done. Mobility is the easiest place to start, which is why most case studies highlight it.