There is a comfortable lie at the heart of corporate governance

That compliance proves integrity, that efficiency is the supreme value, that measuring the measurable is what matters.

NPM is built on the opposite instinct: the refusal to exercise power you technically have but ethically should not, made architectural. Five invariants, one gate, one core of evidence under every standard you run.

Read the framework See it run
The problem

Each standard is coherent. Together they are not

Every management standard is designed in isolation from the others. Run several and you get parallel governance structures that never meet.

Fragmented scopes

A perimeter per standard. Evidence duplicated, scopes overlapping, responsibilities colliding. The documentation burden grows faster than the governance value it produces.

Binary checklists

A control is in place, or it is not. Convenient for audit, pathological for governance. It cannot tell a control that protects something from a control that only produces evidence of its own execution.

Checklist governance

Under operational pressure, compliance degrades into ritual. Conformity gets signaled. The outcomes the activity was designed to produce quietly stop happening.

The question upstream of every standard goes unasked: what is this organization for, and at what cost to the systems that sustain it?

The deeper problem

Your risk matrix cannot see the crash coming

Probability-by-impact matrices are correlational. They plot which risks co-occur with which impacts and say nothing about the causal pathway between them. Two failures compound:

Causal blindness

The matrix treats each risk as an independent coordinate. It is structurally unable to represent the feedback and second-order effects behind most real organizational failures.

Organizational silence

Even when the risk is known, hierarchy and fear keep the signal from the people empowered to act (Chernov et al., 2023Chernov, D., Ayoub, A., Sansavini, G. and Sornette, D. (2023) Averting disaster before it strikes: how to make sure your subordinates warn you while there is still time to act. Springer Nature.). No better matrix fixes that.

NPM answers with structural causal models as the default framing for governance analytics, and invariants that make risk information flow architecturally guaranteed rather than a matter of managerial discretion.

The answer

One core. Every standard becomes a view

L1
The logical gate
Five ethical invariants. Nothing enters the core without clearing all five.
L2
The Compliance Core
Bounded contexts organized around stakeholder impact. Replaces fragmented ISO implementations.
L3
DCCA, the analytical engine
Eight pipeline stages. Keeps the core honest under drift and measures nothing outside it.

"An organization compliant with NPM is compliant with each of the underlying standards as a byproduct."

From the NPM whitepaperPereira O., J.A. (2026) Net-Positive Management: a governance, risk, and compliance meta-framework. TELOS. SSRN abstract 6867443.

The gate

Five invariants. Checked, not aspired to

Every candidate control, policy, metric or model must clear all five at once. Fail one and it is not admitted. The invariants are minimums: add your own, but none of the five can be removed or weakened.

Invariant 1

Purpose primacy

Every automation, control, metric or model must trace to the organizational purpose and the Net Positive North Star. A candidate that traces only to operational efficiency fails.

The test. A documented chain from bounded context, through a strategic objective, to the purpose declaration.

Invariant 2

No manipulation. Judgment stays yours

No hidden persuasion. No dark patterns, internal propaganda or undisclosed nudging, and no decision support that quietly becomes decision capture. This applies to everything the organization produces.

The test. A reasonable third party, examining the mechanism, would recognize it as informing judgment rather than replacing it.

Invariant 3

Extraction limits

For every material topic, the relevant ecological and social limits are identified, internal boundaries keep impact within them, and repair plans exist before a limit is at risk, not after.

The test. Per topic: a documented limit, a monitoring mechanism, a contingency plan. Thresholds published in advance.

Invariant 4

Accountability for externalities

Unintended and second-order effects count, including those carried by suppliers, contractors and downstream users. Consequences the organization discovers must be acted on, not merely disclosed.

The test. The risk register includes externalized effects with named owners, and at least one evaluation cycle per year addresses them explicitly.

Invariant 5

Auditability of the ethical limit

The limit produces evidence. Decisions, exceptions, compensations and overrides leave persistent records that allow continuous evaluation with verifiable results.

The test. An auditor can retrieve every record without asking the decision-maker.

The method

Causally honest analytics

Eight pipeline stages, each auditable, implementing the DCCA loop. Stages 3 and 4 are blocking gates: bad data halts the cycle and fails loudly, before it can reach a score.

S0
Data preparation
Per-source pre-processing
S1
Ingestion
Sources into the pipeline database
S2
Correction
Deterministic fixes for known quirks
S3
Validation
Blocking gate
S4
Statistical validation
Blocking gate, Simpson check
S5
Signal modeling
Criterion scores with SHAP and PDP
S6
Composite scoring
Effective MCDA weights
S7
DCCA finalization
Drift check, weight update, append-only persist
Control 1
Stability monitoring
Drift index against a fixed baseline. Bands published in advance: stable below 0.10, flagged to 0.25, forced recalibration above.
Control 2
Weight recalibration
A smoothed update every cycle, with the single-cycle shift capped. Large moves leave a clear trail.
Control 3
Data-quality validation
Schema, completeness, consistency, distribution sanity and Simpson's paradox, checked before any weight update.
Control 4
Change logging
Every weight update persists an append-only snapshot. No row is ever modified.

Every dashboard claim is labeled correlational or causal. Causal claims require the causation analyses: structural equations from the business map, difference-in-differences per deployed control, and a Bayesian network once the data supports it. A claim that cannot be supported at the level it is presented is downgraded or retracted. Explore the pipeline ›

The dashboard

Every tile traces to evidence

Read-only to operators, produced by the pipeline, never edited by hand. Claims labeled correlational or causal, and missing data shows as "insufficient data", never as a score.

Adoption

Ten to twelve weeks to a defensible baseline

F0
Direction and scope
F1
Business map
F2
Event discovery
F3
Metric catalog
F4
Causal analysis
F5
Intervention design
F6
GRC translation
F7
Audit and improve
A0 scope and stakeholders A1 obligations register A2 policies A3 risk methodology A4 statements of applicability A5 AI impact register A6 audit programme
10–12

Weeks, minimum, in a small organization. Parallel workstreams in a larger one. No metric reaches executive visibility without traversing all eight phases at least once. See the programme ›

Ground truth

The site follows the whitepaper

Net-Positive Management: A Governance, Risk, and Compliance Meta-Framework

"Organizations implementing multiple management-system standards face structural fragmentation: independent scopes, duplicated evidence, and parallel audit cycles that erode strategic coherence and degrade governance into checklist compliance. This whitepaper introduces Net-Positive Management (NPM), a GRC meta-framework that integrates existing standards without replacing them."

First page of the whitepaper: Net-Positive Management, a governance, risk and compliance meta-framework
Johan A. Pereira O., TELOS
© 2026 TELOS, CC BY-ND 4.0
SSRN abstract 6867443