Executive Summary
The World Economic Forum expects 39% of workers' existing skill sets to be transformed or become outdated between 2025 and 2030, names skill gaps as the main barrier to transformation for 63% of employers, and finds 85% of employers planning to prioritise upskilling their workforce (The Future of Jobs Report 2025, 7 January 2025). Yet most people approach skill development reactively: chasing trends rather than building architecture. This article introduces the Core/Adjacent/Aspirational (CAA) Framework for strategic skill portfolio management, inspired by Steve Jobs' philosophy of "connecting the dots".
Why Your Career Needs Architecture, Not Just Skills
Consider two professionals entering the job market in 2026:
Professional A has 15 certifications across trending topics: AI, blockchain, cloud, agile, cybersecurity. Their LinkedIn profile is a wall of badges. They apply to dozens of roles matching keyword filters.
Professional B has 4 deep competencies that form a coherent narrative: data science applied to healthcare, health policy analysis, medical AI ethics, and patient outcome modeling. They do not match as many keyword filters, but every application tells a story.
Why Professional B is better positioned is not a proprietary statistic but the shape of employer demand: the WEF ranks AI and big data, networks and cybersecurity, and technological literacy as the three fastest-growing skills to 2030, and reports 70% of employers expecting to hire staff with new skills (The Future of Jobs Report 2025, 7 January 2025). Engagement and compensation multiples previously printed here were withdrawn in round 219: the surfaces credited for them publish no dated edition.
The difference? Architecture.
The Core/Adjacent/Aspirational Framework
Borrowed from product portfolio strategy (and used by Google, 3M, and Apple in their innovation planning), the CAA Framework applies remarkably well to personal skill management. The 70/20/10 split below is this desk's illustrative allocation; a planning heuristic, not a measured statistic:
| Layer | Allocation | Definition | Example |
|---|---|---|---|
| Core (70%) | 70% | Skills that define your current professional identity and earning power | Financial modeling, regulatory compliance, Python |
| Adjacent (20%) | 20% | Skills that expand your core into neighboring domains | Data visualization, ESG reporting, AI for finance |
| Aspirational (10%) | 10% | Exploratory skills in fields unrelated to your current work | Behavioral psychology, creative writing, UX design |
Why 70-20-10?
The 70/20/10 allocation is this desk's framework, not an empirical finding: it encodes the judgement that depth pays the bills, adjacency creates the next role, and a small deliberate reserve buys optionality. Percentages previously attributed to an OECD longitudinal study were withdrawn in round 219: no edition of an OECD publication we could open prints them. What the OECD does publish is the return to depth: a 56% earnings premium for tertiary-educated 25–64 year-olds (Education at a Glance 2024, 10 September 2024).
Steve Jobs and the Art of Connecting Dots
In his famous 2005 Stanford commencement speech, Steve Jobs described how a calligraphy class he took at Reed College, seemingly useless at the time, became the foundation for the Macintosh's revolutionary typography. "You can't connect the dots looking forward," he said. "You can only connect them looking backwards."
This is the Aspirational layer of the CAA Framework in action. The illustrative 10% allocation to seemingly unrelated skills is not indulgent: it is strategic optionality. A cross-domain earnings premium previously credited to the OECD was withdrawn in round 219; the OECD's own framing of resilience in the green and digital transition is in its Skills Outlook 2023 (6 November 2023), and the measurable premium it publishes is the 56% return to tertiary attainment.
Modern Applications of Jobs' Philosophy
- A software engineer who studied philosophy becomes the ethics lead for AI products: a role created by the same AI demand the WEF ranks first among fast-growing skills to 2030
- A marketing manager with data science skills becomes a growth engineering leader: one of the hybrid roles created where analytical and commercial work meet
- An architect who learned parametric programming leads computational design teams reshaping how buildings are specified and built
What a stranger can verify is what a stack is worth
Two published records describe the same labour market and disagree about where the scarcity is. Employers expect a large share of existing skill sets to be transformed or become outdated this decade, which is a statement about demand. Eurostat records what adults did about it: in 2022, 47% of adults aged 25 to 64 in the European Union took part in education and training in the previous 12 months, and 64.2% took part in informal learning. Source: Eurostat, Adult learning statistics, data extracted in May 2026.
Half a labour market in some form of learning is not a skills shortage in the ordinary sense. It is a verification shortage. When almost everyone is learning something, the scarce thing stops being the learning and becomes the evidence that a particular combination of it works, because an employer cannot price a claim it has no way to check.
The selection literature is unusually direct about which evidence gets weighed. Structured interviews reach a validity of .51 and work samples .54, both of them measuring demonstrated capability rather than declared interest. That is the mechanism behind every recruiter behaviour candidates find arbitrary: a method that shows the work is worth more to the person hiring than a method that describes it, so the item in your architecture that carries an artefact outranks the item that carries a certificate.
The same asymmetry appears in the one return the OECD actually measures. A 56% earnings premium for tertiary-educated 25 to 64 year-olds is a premium on a qualification a stranger can confirm in one step, at a stated institution, on a stated date. Source: OECD, Education at a Glance 2024, 10 September 2024.
And the group least able to produce that kind of proof is exactly the group the European Skills Agenda set a target for. The EU objective was that by 2025 at least 20% of unemployed adults aged 25 to 64 should have a recent learning experience. In 2025 the share was 15.9%, against 9.3% in 2015, and 8 member states met the target. Learning capacity, in other words, is unevenly distributed exactly where a legible stack would matter most.
So the practical rule follows from the documents rather than from advice. Build the architecture so that each layer names the artefact that proves it: an analysis a sceptical stranger could read, a system in use, a measured improvement, or a qualification on a public register. A stack of six items with two artefacts is legible. A stack of twelve items with none is a list, and a list is priced as a list.
56%earnings premium for tertiary-educated 25 to 64 year-olds
The one return on depth the OECD measures, and it is paid on a qualification a stranger can confirm
OECD, Education at a Glance 2024, 10 September 2024. Read it as the price of verifiable attainment rather than as the price of study: the premium attaches to the record, which is why an unverifiable layer of a stack earns nothing comparable.
Source 2 Organisation for Economic Co-operation and Development
Compound Skill Stacking: The Mathematics of Career Advantage
Scott Adams, creator of Dilbert, popularized the concept of "talent stacking" in How to Fail at Almost Everything and Still Win Big (2013): the idea that being in the top 25% in two or more skills can make you more valuable than being in the top 1% in any single skill.
The mechanism is combinatorial scarcity: a pairing that is common in neither of its parent fields has few substitutes, so it is priced as a specialism rather than as a commodity. A profile-analysis statistic previously printed here was withdrawn in round 219 for want of a dated document. The selection literature supports the underlying point differently and checkably: structured interviews reach a validity of .51 and work samples .54, because both measure demonstrated capability rather than declared interest (Schmidt & Hunter, Psychological Bulletin 124(2), 1998).
The Compounding Effect
Skills do not add: they multiply. A professional with:
- Financial analysis: the core
- + Python programming
- + Machine learning
- + ESG and sustainability knowledge
Creates a profile whose value is not the sum of its parts but the scarcity of the intersection: few competitors hold all four, and the roles that need all four cannot be filled by any one of them. The specific multipliers previously printed here were withdrawn in round 219: they were arithmetic, not measurement.
🎯 For Junior Candidates: Building Your First Architecture
- Define your core (3-4 skills). What can you do well enough today that someone would pay you for it? Be honest. This is your foundation.
- Map 3 adjacent skills. What skills would make your core 2x more valuable? Look at job postings for roles 2 levels above yours: what skills do they require that you don't have?
- Choose 1 wildcard. Something completely unrelated that interests you. Art, music, psychology, philosophy, biology. This is your future dot to connect.
- Create a quarterly review cadence. Every 90 days, assess: Have your core skills deepened? Have your adjacent skills created new intersections? Has your wildcard revealed unexpected connections?
- Build publicly. The WEF recommends creating a "learning portfolio": visible evidence of your skill development. Blog posts, open-source contributions, project documentation, conference talks.
🎓 For Academia: Curriculum Architecture Implications
The CAA Framework challenges the rigid departmental structure of most universities. The OECD recommends that institutions:
- Create structured cross-disciplinary pathways that combine a major with a minor from an unrelated field
- Implement "skills architecture" advising where career counselors help students build coherent skill narratives
- Develop micro-credential programs that allow students to add adjacent skills without full-course commitments
- Partner with industry to create project-based learning that develops compound skills in real-world contexts
Universities that build coherent cross-disciplinary pathways are aligning with measured demand rather than fashion: the WEF finds 63% of employers naming skill gaps as the main barrier to business transformation and 85% prioritising upskilling (The Future of Jobs Report 2025, 7 January 2025). A graduate-employment differential previously credited to Education at a Glance 2024 was withdrawn in round 219; that report's measured figure is the 56% earnings premium for tertiary attainment.
Three layers of a stack, and what a recruiter can actually check
| What the layer is | The proof it can offer | What the published record says about that proof |
|---|---|---|
| A qualification on a public register | A named institution, a stated date, one lookup | The OECD measures a 56% earnings premium for tertiary-educated 25 to 64 year-olds |
| A produced artefact: an analysis, a system in use, a measured improvement | A work sample the employer can set and score | Work samples reach a validity of .54 in the selection literature, above every self-report method it compares |
| A described interest or a completed course with no output | A claim, checked by asking the candidate about it | Structured interviews reach .51 because they test demonstrated capability; a declared interest is not tested at all |
Sources 2 Organisation for Economic Co-operation and Development · 4 American Psychological Association · 6 Eurostat
Your 12-Month Skills Architecture Plan
Months 1-3: Foundation Audit
Map your current skills using the OECD Skills for Jobs database. Categorize each as Core, Adjacent, or Aspirational. Identify gaps and redundancies.
Months 4-6: Core Deepening
Invest heavily in your top 2 core skills. Aim for top-quartile competency. Take on stretch projects that push your boundaries.
Months 7-9: Adjacent Expansion
Begin systematic development of your 2 highest-value adjacent skills. Look for projects that combine core + adjacent.
Months 10-12: Integration and Signaling
Create visible work that demonstrates your unique skill combinations. Update your professional profiles. Publish or present on your intersections.
Sources
- World Economic Forum, « The Future of Jobs Report 2025 », 7 January 2025
- OECD, « Education at a Glance 2024: OECD Indicators », 10 September 2024
- OECD, « OECD Skills Outlook 2023: Skills for a Resilient Green and Digital Transition », 6 November 2023
- Schmidt, F. L. & Hunter, J. E. (1998). « The Validity and Utility of Selection Methods in Personnel Psychology », Psychological Bulletin 124(2), 262–274
- Jobs, S. (2005). 114th Stanford University Commencement address, 12 June 2005
📚 Skills Mastery Series: Continue Reading
This article is part of CareerOn's 5-part flagship series on skills strategy, backed by OECD, WEF, and LinkedIn data.
Withdrawn figures
On 10 September 2026 this page was audited against its own reference list, document by document. The claims withdrawn are "10 new skills every 5 years", LinkedIn's 3.4x recruiter engagement and 27% compensation gap, the 50,000-profile analysis and its "3.2 unusual combinations", the "+340% in postings", Glassdoor's 42% ethics premium, the "US$12 trillion construction industry", the 1.8x/2.3x/1.5x/6.2x stacking arithmetic, and the OECD figures for 40% shorter searches, 22% higher satisfaction, a 15% cross-domain premium and 34% higher graduate employment: none of which appear in any edition of an OECD publication we could open. The 70/20/10 allocation is this desk's framework, not an empirical finding. No document published by a named body, at an edition a reader can open, was found printing them, so they have been withdrawn rather than left standing on a reference nobody can check. Every figure that remains above is credited, in its own sentence, to a document a reader can open at a stated edition. Salary ranges are this desk's indicative ranges for France and are labelled as such; they are not figures taken from a cited document.