The Shirky Principle: Every HR Team Needs the Problem It Was Built to Solve

60-Second Summary

Clay Shirky (2010, coined by Kevin Kelly summarising Shirky's writing): institutions preserve the problem to which they are the solution, because their existence depends on the problem's persistence.
Not conspiracy. Selection. Teams whose measured outputs are 'programs run' rather than 'problem solved' cannot rationally act to make themselves unnecessary.
In HR this shows up as: DEI teams that measure participation not composition change, L&D teams that measure course completion not skill acquisition, ER teams that measure cases handled not conflict prevented.
The counter-design is problem-elimination incentives: teams whose success is measured by making their own function smaller, coupled with graceful redeployment paths so success does not mean unemployment.
Rare in the wild. Requires CEO-level design, sunset dates on function charters, and a comp structure that rewards the team for eliminating the problem it was chartered to solve.

Clay Shirky writes about the sociology of the internet. Kevin Kelly, summarising Shirky's writing in a 2010 blog post, distilled it to nine words: 'Institutions will try to preserve the problem to which they are the solution.' It is one of those sentences that reorganises how you read every quarterly HR-programs review you have ever attended, forever. It also happens to be true.

Where the principle comes from

“Institutions will try to preserve the problem to which they are the solution.”
— Kevin Kelly (2010), summarising Clay Shirky

The formulation appeared in Kelly's blog and was quickly picked up because it named a pattern people recognised immediately. It has deep roots in institutional economics — Niskanen's budget-maximising bureaucrat, Tullock's rent-seekers, Wilson's mission-drift — but the compression is what made it viral. It is not the claim that people in institutions want the problem to persist. It is the claim that institutions structurally select for actions that preserve the problem, whether or not any individual intends it. Iron Law adjacent, but distinct: Iron Law is about self-preservation of the institution; Shirky is about self-preservation of the problem.

The Shirky Principle running in HR

Six HR functions the Shirky Principle predicts will drift

1
DEI teams
Chartered to shift composition. Measured on training hours, ERG participation, and survey inclusion scores — all of which can rise while composition doesn't move. See Schelling article: the mechanism producing homogeneity is structural, but the interventions measured are individual.
2
Learning & Development
Chartered to close skill gaps. Measured on courses run, completion rates, and NPS. Skill acquisition is rarely measured because it's hard. A skill-closed workforce would require the L&D team to shrink; the incentive to actually close skills is inverted.
3
Employee Relations
Chartered to resolve and prevent conflict. Measured on cases handled and turnaround time. A conflict-free organisation makes the ER team unnecessary. Structural prevention (better manager training, clearer role design) is under-invested because it reduces case volume.
4
Engagement / Culture
Chartered to raise engagement. Measured on survey participation, action-plan completion, and pulse coverage. A structurally engaged workforce doesn't need annual surveys; the survey program becomes its own reason for existing.
5
Talent Acquisition
Chartered to hire. Measured on requisitions filled, time-to-fill, pipeline health. Internal mobility programs that would reduce external hiring are chronically under-invested because they cannibalise TA's core metric.
6
Compliance / Ethics
Chartered to prevent violations. Measured on training completed, cases investigated, policies published. A violation-free culture is structurally invisible; the compliance function's visibility depends on cases existing.

Why it isn't fixed by good people

This is the point at which HR leaders reasonably object. 'My DEI team wants composition to change. They would celebrate becoming unnecessary.' Individually true. Structurally irrelevant. The mechanism operates through what gets measured, budgeted, and promoted — not through what individuals want. A team member who says 'we should shrink because the problem is solved' becomes an unpromotable member of a shrinking team. The system selects against them. Over time, the promoted members are those who successfully argue for more scope, more headcount, more programs. This is not cynicism about individuals; it is realism about incentive structures.

The uncomfortable test

Ask every HR sub-function leader: 'What is your success metric such that hitting it makes your team smaller?' If they can't answer, they are structurally selected to preserve their problem. Almost none can answer without prompting.

Design patterns that beat the principle

Sunset charters. Every HR program has a 24- or 36-month expiry, renewed only with evidence of business outcome, not activity.
Success-as-shrinkage metrics. DEI success = composition change per dollar spent, not participation hours. L&D success = measured skill acquisition per learner, with a plan for what the team does after gap closure. ER success = case-volume reduction over time, not case-throughput.
Redeployment guarantees. The single largest structural resistance to shrinking a function is that the leaders will lose their jobs. Guarantee redeployment to line roles for leaders who successfully make their function smaller.
External benchmarks with teeth. Compare your HR function's cost and outcomes to the 25th-percentile peer, not the median. Median comparisons preserve the status quo; percentile comparisons reveal drift.
Business-side scorecards. HR programs are scored by the line managers they serve, not by HR-internal committees. A DEI program with a 40% adoption rate among line managers is failing regardless of how many trainings it ran.
The 'what would we build from scratch' audit. Annual. Kills the accumulated programs no rational new function would create today.

The rare example that worked

Netflix's early people function was structured explicitly against Shirky: the team was small, tied to business metrics, and Patty McCord routinely killed programs (annual reviews, PIP processes, formal performance ratings) rather than expanding them. The result was 8-10 person people team supporting hundreds of engineers effectively for years. Rare because it required a CEO (Hastings) who protected the design.

FAQ

Frequently asked questions

Doesn't this argue against having HR functions at all?

No. It argues for designing HR functions with charter clarity, sunset dates, and success-as-shrinkage metrics. The problem is not the existence of the function; it is running it as a permanent institution with expanding scope.

What about DEI specifically — isn't the problem far from solved?

Correct — and this is precisely why Shirky-aware design matters most there. If the problem is structural (see Schelling article), then measuring participation in training is the wrong instrument. Composition change per intervention dollar is the right one. Most DEI functions measure the former because it makes the team look successful; the latter often reveals that the current programs are not the highest-leverage.

How does this relate to the Iron Law?

Iron Law: bureaucracies drift toward self-preservation. Shirky: institutions drift toward preserving the problem. Iron Law is about the institution; Shirky is about the problem the institution feeds on. They compound.

Takeaways

Institutions structurally preserve the problem that justifies them. Not because individuals want it — because the system rewards actions that preserve it.
In HR, this shows up as programs measured on activity rather than problem elimination.
The counter-design requires charter sunsets, success-as-shrinkage metrics, redeployment guarantees, and business-side scorecards.
The test: can each HR sub-function name a metric such that hitting it makes their team smaller? If not, you are running Shirky-shaped programs.
The healthy state is not the absence of HR functions. It is HR functions whose success would visibly reduce their own scope, and whose leaders are rewarded for that reduction.