• Member
• November 12, 2026

From Islands to a Value Chain An Internal Competence Exchange for the University

Nov 10, 2026 · @Shailendra Jaiswal

The observation: capability islands

Observations 

•       Research on teaching and learning stays in education, while engineering, pharmacology and other departments teach without drawing on it.

•       Design-of-experiments competence stays in mathematics and statistics, although engineering and pharmacology would value it.

•       Computer science strengths stay largely within computer science.

•       Applied units (the sciences, computer science, nursing) do a good deal of applied work without involving the internal experts in the relevant field.

•       Insight from literature, media, sociology, economics and psychology is not used by faculty elsewhere to upgrade their teaching and research.

These are not failures of individuals. They are what a loosely coupled system produces when nothing carries ideas across the boundary. The question for this paper is how to make a university's internal supply of knowledge findable and usable by its own members, without damaging the autonomy that makes the university productive.

Position in brief

Universities do not coordinate like industry chains because nothing forces them to. In industry, the object being transformed moves from stage to stage, so each stage depends on the one before. An academic output, usually a paper, is a finished product read by peers in the same field. Loose coupling is right for discovery and wrong for the stages that convert knowledge into value for the institution's own members. Interdisciplinarity is therefore a necessity for the system and an option for the individual. The remedy is an internal competence exchange: an output-input matrix of what each unit can supply and what each needs, supported by interfaces, brokers and credit, built around autonomy rather than against it.

Seven claims carry the argument:

1.      Industry chains coordinate because of input-output dependence, shared interfaces, integrator roles and end-customer pull. Departments have none of these between them.

2.      Autonomy at the discovery stage should be preserved. Synthesis, teaching and application should be run as a real chain with owners, handoffs and credit.

3.      Internal competence should be made findable through capability passports and need passports, searched from either side of one matrix.

4.      Demand that people cannot see (they do not know what they are missing) must be surfaced through methodological audits and expert clinics at approval gates that already exist.

5.      Incentives should stay small and qualitative, so that coordination does not breed gaming.

6.      The exchange plugs into two steps of the chain each researcher already runs: sourcing and process design (finding colleagues' competences) and the market (internal uptake recorded alongside citation).

7.      Universities should borrow industry's diagnosis, not its prescription: adopt the chain view, interfaces, brokers and demand signals, and leave hierarchy, quarterly targets and efficiency as the measure of scholarship.

Industry chains and the academic knowledge value chain

Industry chains coordinate because four mechanisms force it, and universities lack all four between departments.

Forcing mechanism	In an industry chain	In a university today
Input-output dependence	One stage's output is the next stage's input, so delay or defect shows downstream	Colleagues rarely need my output to do their own work
Shared interface	Specifications, drawings, units and contracts let functions work on one object	Each discipline has its own vocabulary, with no common specification
Integrator roles	Systems engineers, logistics and programme managers, standards bodies	Few synthesizers or brokers, and no career path for them
End-customer pull	A paying customer sends demand back up the chain	The audience is the peer group, so no demand signal crosses departments
Reward	Pay and promotion follow chain performance	Reward follows individual, discipline-denominated output

In global value chains, lead firms govern suppliers anywhere from arm's-length market links to full hierarchy (Gereffi's typology, cited from general knowledge and not checked for this paper). Departments sit near the arm's-length end: publish, and hope to be cited. A university department's idea travels mainly through people, such as students, postdocs and moves between institutions, which is slow and accidental.

A shared object changes this. At Arizona State University, a team that grew to more than 100 people met daily for 100 days to develop Arizona's first saliva-based COVID-19 test (ASU University Design Institute). The cause was the object and the goal, not a different kind of academic.

Mapped as a chain, the academic route from discovery to application looks like this.

academic knowledge value chain · 7 stages, 3 with no clear owner

Each stage has a different owner today, or none. Synthesis, application and feedback have no clear owner, and the loop from learners and users back to research agendas is rarely closed.

The researcher's own chain

Each researcher already runs a complete supply chain alone, and the exchange can plug into two of its steps.

Step	In research	Who supplies or receives
Sourcing	Literature survey	The global literature
Process design	Research methodology	The field's own methods, learned in training
Production	Experiment, fieldwork, analysis	The researcher's lab, data and time
Finished good	The paper	Peers in the field
Quality control	Peer review	External reviewers
Market	Publication	Journals and conferences
Demand signal	Citation	Other producers in the same field

Academia therefore looks like a cottage industry, not a factory. Each researcher runs every step alone, and the university supplies the campus, brand and treasury, not the production line.

Two features explain why the institution sees so little flow:

•       Both ends sit outside the institution. Inputs come from the global literature and customers are peers elsewhere, so knowledge travels through disciplinary networks and bypasses the university.

•       Citation measures uptake by other producers, not delight. Students, other departments and administrators do not cite, so internal demand leaves no trace.

The author also finds little knowledge management for competence in institutions of learning. Libraries and repositories hold finished outputs, while a method sits in the methods section of a paper written for peers, so it is rarely captured as a reusable asset.

The exchange does not have to invent a chain. It plugs into two steps of the one each researcher already runs:

1.      Sourcing and process design. Extend the search for inputs from the literature to colleagues' competences, such as experimental design or qualitative methods.

2.      Market and demand signal. Add internal customers (students, other departments, the institution) to the external peer market, with an internal equivalent of citation. The Contribution Ledger logic in the Srijan Sanchar Exchange could record who used whose method, course or insight.

An internal uptake record can drift toward rewarding the popular over the rigorous, so it should inform judgment and not replace it.

What loose coupling protects, and where it fails

Loose coupling is the right design for the front of the chain and the wrong one for the back. Ajay Mohan Goel's essay A University Is Not a Corporation describes the university as a loosely coupled system in which trained experts do the core work largely on their own. It also notes that institutional work such as curriculum redesign, accreditation and mentoring behaves like a public good that each person has reason to leave to someone else.

What the low-coupling design protects:

•       Independence of judgment. A reputation anchored in a field and a network of peers is portable, so a weak department rarely sinks a strong individual (the essay's point).

•       Long horizons and tolerated failure. No one has to show quarterly results, so slow and uncertain work can survive.

•       Many parallel bets. A failed line of inquiry in one unit does not cascade into others (this paper's argument).

•       Quality control outside management. Disciplinary peer review sets standards that internal politics cannot override.

•       Low coordination cost. Solo scholarship needs no meetings to align.

Where it fails is the back of the chain: synthesis, teaching and application exist to combine and deliver knowledge, so leaving them to chance is a design failure, not a virtue.

The essay adds a design clue. In the corporation it describes, interdependence channels most self-interest into cooperation but leaves openings for gaming. In the founder-funded foundation, politics stayed limited when little was at stake for individuals (one author's observation, not a study). The implication is to add interdependence where conversion happens while keeping personal stakes low and reversible.

Rule: loosely couple discovery, and couple the stages that turn knowledge into value.

Learning from industry: mechanisms, not logic

Universities should borrow industry's diagnosis, the chain view, and leave its prescription of hierarchy and targets.

Industry practice	Verdict	Why
Seeing work as a chain with owners and handoffs	Borrow	Makes the conversion stages visible
Interfaces and shared vocabularies between functions	Borrow	Lets different functions work on one object
Integrator roles such as brokers and synthesizers	Borrow	Someone owns the joins
A demand signal that flows back to producers	Borrow	Closes the feedback loop
Process discipline in teaching delivery, application and administration	Borrow	These are conversion stages, not discovery
Tight coupling of discovery, command and control, quarterly targets	Leave	Removes variety and invites gaming
Treating every stakeholder as a customer to be satisfied	Leave	Pulls research toward what is easy to measure
Standardized methods	Leave	Variety is the asset in discovery
Efficiency as the core measure of scholarship	Leave	Rewards speed over depth (author's reading)

The essay already names the failure mode: institutions that adopt corporate-style targets for admissions, placements and rankings, without corporate authority and without academic security, get the worst of both. Any borrowed practice can be screened with three questions:

1.      Does it couple the conversion stages or the discovery stage? Only the first is safe.

2.      Does it work by pull and opt-in, or by mandate?

3.      Are the personal stakes low enough that it cannot be gamed?

Borrowing also runs the other way. Peer review, open science and loosely managed labs exist partly because tight management does not suit discovery (author's reading). The competence exchange follows the pattern: it takes the chain view from industry and leaves the autonomy of discovery alone.

Interdisciplinarity: a necessity for the system, an option for the individual

Whether interdisciplinarity is necessary depends on what is being done.

Situation	Interdisciplinarity is	Why
A puzzle inside a mature discipline	An option	Disciplines are efficient divisions of labour, and forced crossing adds cost
A problem, user or phenomenon that ignores boundaries (health systems, climate, cities, AI governance)	A necessity	No single field holds the whole problem
Educating graduates	A necessity	Graduates meet integrated problems, not departmental ones
The institution's own capability	A necessity at system level	Specialists can stay specialists, but the institution must be able to combine them

The industry analogue is modular specialization with standard interfaces, not everyone doing everything. This paper also argues, without having tested it, that as within-field answers get cheaper (including through AI tools), value shifts toward integration and judgment across fields.

ASU's experience adds a caution. Its own faculty say that restructuring alone can create new silos, and that interdisciplinarity stays an aspiration unless assessment for tenure and promotion changes, because current systems still weigh individual contribution over shorter periods (ASU University Design Institute).

The Competence Exchange Matrix

The exchange is an output-input matrix of the university's own knowledge: rows are the competences units can supply, columns are the units that need them, and each cell records a flow of a given kind. Competences, not departments, are the unit of analysis.

Reading by row answers "who could use what I have?" Reading by column answers "who here can help me with this?" The column view is the matrix inverted: start from a need and find the supply.

In input-output economics, the Leontief inverse gives total requirements, direct plus indirect:

L = (I - A)^{-1}

Here A holds the direct flows between units and L the total flows. Statistics may not feed nursing directly, but it may feed education research, which feeds nursing pedagogy. Computing the indirect paths shows which competences are hubs whose absence would damage the whole institution, and which units are isolated.

Illustrative cells, drawn from the author's observations and to be confirmed through passports:

Competence	Home	Likely users	What users gain
Design of experiments	Mathematics and statistics	Engineering, pharmacology	Valid, efficient studies
Learning and teaching research	Education	Every teaching department	Better pedagogy, evidence on what works
Computational methods	Computer science	The sciences, nursing	Scale, automation, reproducibility
Qualitative and behavioural methods, media, sociology, psychology, literature	Humanities and social sciences	Nursing, applied sciences, engineering	Understanding of users, patients and behaviour
Evaluation of cost, uptake and policy	Economics	Applied units	Judgment on cost and adoption

Three objects make the matrix work:

•       Capability passport. The competence in plain language, a worked example, the level of engagement offered (consult, co-teach, co-supervise, co-author, train, share a tool), and a contact.

•       Need passport. What a person is trying to do and where they are stuck, in their own words.

•       Matching at three levels. Attribute matching on keywords comes first. Structural matching comes next, for example two problems that both compare treatments under noise. Analogical matching comes last. This follows the architecture of the Srijan Sanchar Problem Solver–Solution Seeker Exchange, applied at the scale of one institution.

Profiling systems such as VIVO can feed the supply side of the matrix. They do not by themselves reveal need, which is why the next section matters.

Mechanisms: surfacing demand, credit and guardrails

A directory solves the visible need but not the unknown unknowns: weak methodology is invisible from inside a field. Six mechanisms address this.

1.      Methodological audit. An expert reviews a sample of theses or applied studies from a unit and shows concretely where competence would have helped. The evidence persuades more than any catalogue.

2.      Expert clinics at existing gates. Place them where need surfaces anyway: proposal approval, ethics review, course design approval and thesis vivas. CTSA biostatistics units serve on review boards and grant panels, run walk-in clinics of 15 to 20 minutes, and join interdisciplinary study-design studios (Perkins et al.).

3.      Credit. Count institutional work (curriculum design, synthesis, brokering, consulting) as scholarship, with workload credit, named authorship and promotion weight. The same paper says consulting is undervalued at many institutions and can hurt promotion, and that pay should follow effort while authorship follows scientific contribution.

4.      Capacity and funding for hub experts. The CTSA paper recommends mixed funding, and notes that free cores often cap free hours (for example 5 to 10) before charging. UCSF's consultation service pays consultants only for hours worked, which lets it keep a large pool available (UCSF CTSI).

5.      Pull, not mandate. Units opt in, and leaders act as conveners. Heads of departments have authority without much power, so persuasion has to do the work.

6.      Guardrails against gaming. Keep incentive weights small, judge outcomes qualitatively, and avoid targets such as counts of joint papers or projects. The essay's corporate examples show that numbers someone can shape get shaped.

Benchmarking: five comparable efforts

Five efforts each cover part of the exchange: cataloguing supply, delivering expertise to other units, making one department's research usable by others, pairing faculty across fields, and restructuring. Figures are as reported on the cited pages on 2026-11-10, and Bio-X and UCSF figures are self-reported by the programs.

Effort	What it does	Mechanism that matters	Reported evidence	Lesson for the exchange
VIVO	Open-source software and ontology that represents scholarship, for expert finding and discovery across disciplines	Draws on local systems of record (HR, grants, courses, faculty activity) or publication databases; member-supported	Adopters include US federal agencies and universities worldwide; the pages reviewed give no collaboration-impact figures	Cataloguing supply is a solved technical problem. It shows what people do, not what they need, so pair it with need passports and audits (author's reading)
CTSA biostatistics and research-design units, e.g. UCSF and Perkins et al.	Embedded expert consultation and collaboration on study design and analysis for investigators across an academic health centre	Core funding, free-hour caps then fee-for-service, walk-in clinics, design studios, written agreements on funding, lead time and authorship	UCSF: more than 15,000 hours of advice over four years (undated newsletter). Across 62 CTSA sites, the share supporting over 300 projects a year rose from 9.1% to 24.2% between 2011 and 2013. Demand often exceeds supply	Closest analogue to competence flowing from statistics and design to applied units. The hard design work is funding, prioritization, credit and authorship rules, not the directory
Simon Initiative, Carnegie Mellon	Builds a learning-engineering ecosystem so faculty can use learning science in their own teaching	Infrastructure plus human support; overhaul of centerpiece courses, including Statistical Methods and Introduction to Computer Science; tools offered to other institutions	The page states mission and goals, with no outcome figures	Makes education research usable by other departments through support and shared courses as common ground (author's reading)
Stanford Bio-X seed grants	Seed funding for faculty teams across disciplines on high-risk ideas in bioscience, bioengineering and biomedicine	Two-year grants of $200,000 per project, about $4 million every other year	312 projects and over 478 faculty from all seven schools through round 13 (2026); over $400 million in external funding credited to the program	Small, time-limited money pulls people across boundaries. It is a pairing tool, not a way to discover competence
Arizona State University transdisciplinary schools	Reorganizes departments into transdisciplinary schools and institutes	25 new transdisciplinary schools since 2002; Biodesign Institute (2002) with 150 tenured faculty and $760 million in research funding; a 2014 charter	The COVID-19 test team of over 100 people; interviewees warn that restructuring can create new silos and say promotion biases persist	Restructuring is the costliest route and its own participants say incentives lag. Add an exchange layer and fix credit before redrawing boxes

Read across, as this paper's judgment: the three efforts that work through consultation, infrastructure and small funding (CTSA, Simon, Bio-X) each pair a service with credit or money. The structural route (ASU) is the most expensive, and its participants report that incentives remain the binding constraint. None of the five, as described in the sources reviewed, offers a two-sided matrix of capability and need across all departments.

Pilot, measures and risks

A one-semester pilot can test whether the matrix changes behaviour before anything is restructured.

1.      Choose about 6 departments and 10 to 12 transferable competences.

2.      Collect capability passports and need passports from both sides.

3.      Run two or three methodological audits.

4.      Station expert clinics at two existing gates, such as proposal approval and course design.

5.      Track outcomes for one semester, then review and extend.

Measures to watch, judged qualitatively and not set as targets: joint studies or co-taught modules started, methods changed in applied studies, hours of expert advice delivered (as UCSF reports), repeat requests, and the load on hub experts.

Risks and responses:

•       A catalogue nobody uses. Pull demand through the gates instead of relying on announcements.

•       Hub experts overwhelmed. Build in capacity, credit and workload relief.

•       No credit. Without it the matrix runs on goodwill until the goodwill runs out.

•       Status gaps. Units may treat humanities and social science competence as soft. Early worked cases flowing in that direction help.

•       Gaming. It appears when outcomes become targets, so keep weights small.

Lineage: this paper as a microcosm of the chain

This paper's own development ran the same chain it describes, and each version was a handoff. Lineage here means the record of where each idea came from and how it changed on the way.

Idea	Origin	Version
Loose coupling and the professional bureaucracy	Goel's essay, which uses Mintzberg's term	1
Industry value chain versus the academic chain	The author's question on why academia does not coordinate like industry; Gereffi's typology from general knowledge	1
Interdisciplinarity as a system necessity and an individual option	This paper's argument, tested against ASU's experience	1
The output-input matrix and its inverse reading	The author's proposal; the Leontief method from general knowledge	1
Capability and need passports, three matching levels, Contribution Ledger	Srijan Sanchar's Problem Solver–Solution Seeker Exchange	1 and 2
Five benchmarks	Pages on ASU, CTSA units, Bio-X, the Simon Initiative and VIVO	1
The researcher's own chain	The author's analogy: literature survey, methodology, paper, peer review, publication, citation	2
Mechanisms versus logic, and the three-question test	This paper's synthesis, prompted by the author's question on whether universities should learn from industry	2

Read as a chain, the paper itself maps like this:

Chain step	In this paper	Lineage
Sourcing	Goel's essay and the benchmark pages	The essay of 6 Nov 2026 posed the question; the pages were opened and checked
Process design	The value-chain analogy and the output-input matrix	The author's observation of capability islands in university work
Production	The position paper, versions 1 and 2	Each question from the author reshaped the structure
Quality control	Figures checked against opened pages	A stronger wording about ASU was softened to match what the source says
Market	The author's own university work, where the islands were seen	The use the paper is put to
Demand signal	The author's next question	The researcher's chain and the industry lessons came from replies

The same record, kept for competences, is what the exchange needs: where a method came from, who has used it, and how it changed.

New explorations and open questions

Four explorations would test this version of the argument:

1.      Chain audit. Take one department and walk it through the seven steps of the researcher's chain, marking for each step whether inputs and customers are internal or external.

2.      Transfer note. Publish a short note beside a paper, written for a reader in another field, and see whether any internal request follows.

3.      Internal uptake record. Trial a light ledger of who used whose method, course or insight, and check whether it surfaces hub competences.

4.      Mechanism test. Apply the three-question test to practices an institution has already adopted, such as admissions, placement or ranking targets, and see which fail.

Questions that remain:

•       Which stage of the chain is weakest in a given institution: synthesis, translation or feedback?

•       Can an interface vocabulary be built between two departments without flattening either?

•       Does credit for synthesis change behaviour, or does it only create a new box-ticking category?

•       Does the ASU COVID-19 test team show that a shared object matters more than incentives? Comparing fields with and without large consortia could test it.

•       Who should own the matrix: the library, the research office or the teaching and learning centre?

•       How far can a competence be described in plain language before its meaning is lost to a user in another field?

•       If citation measures uptake by other producers, what is the internal equivalent, and who should count it?

•       Where does a researcher's chain lean most on internal inputs today, and where on external ones?

•       Is the researcher the right unit of the chain, or is it the lab, the course or the programme?

•       Would making methods reusable change the paper itself, for example through a transfer note published beside it?

•       Could an internal uptake record reward the popular over the rigorous, and how would that be detected?

•       What would show the position wrong? One test: the passports and clinics produce no new requests across departments.

Sources

Pages opened on 2026-11-10:

•       A University Is Not a Corporation, Ajay Mohan Goel, 6 Nov 2026

•       Embedding Interdisciplinarity at Arizona State University, University Design Institute, Dec 2021

•       Stanford Bio-X Seed Grants

•       UCSF CTSI Consultation Service, CTSA newsletter article, undated

•       Best Practices for Biostatistical Consultation and Collaboration in Academic Health Centers, Perkins et al., The American Statistician 70(2), 2016

•       About VIVO

•       Simon Initiative mission, Carnegie Mellon

Not sourced here: Gereffi's governance typology and the Leontief inverse are cited from general knowledge. The Problem Solver–Solution Seeker Exchange is Srijan Sanchar's own framework.