ODX Assessment Methodology

Version 1.0 Last updated: 25 July 2026

This document explains how ODX assessments are designed, scored, and interpreted. It is written for governance committees, procurement reviewers, consulting partners, and any executive who wants to understand what the numbers in an ODX report mean before acting on them.

Every figure produced by ODX can be traced to a method described here. Where a figure rests on an assumption rather than a measurement, this document says so.

1Purpose and scope

What ODX measures

ODX measures organizational maturity: how consistently and deliberately an organization performs a defined set of practices. Each assessment covers one domain, such as delegation of authority, governance, or leadership effectiveness.

An assessment produces four things:

1. 1.A score from 0 to 100
2. 2.A maturity level from 1 to 5
3. 3.A gap analysis against the ODX target standard
4. 4.A prioritized action plan

What ODX does not measure

ODX does not measure outcomes. A high governance score does not guarantee good decisions, and a low score does not mean the organization is failing commercially.

ODX does not audit. Responses are self reported by executives. No evidence is collected, verified, or tested. An ODX assessment is a structured self diagnostic, not an assurance engagement.

ODX does not certify. No ODX score constitutes compliance with any regulation, standard, or listing requirement.

Design intent

Each assessment is designed to answer one executive question. For example, the Delegation of Authority assessment is designed to tell an executive whether their delegation framework would survive outside scrutiny from an auditor, an underwriter, a new investor, or a fraud investigation.

This intent shapes which questions are asked. Questions hunt for the gaps an external party would find, not the gaps that merely inconvenience staff.

2Assessment design principles

Fixed format of 12 questions

Every ODX assessment contains exactly 12 questions across three sections of four questions each.

This is a deliberate constraint with three consequences:

Comparability. Because every organization answers the same 12 questions, results can be compared directly. Assessments that allow question customization cannot produce valid cross organization benchmarks.

Completion. Twelve questions take five to ten minutes. Executive completion rates fall sharply beyond that.

Discipline. A fixed budget of 12 questions forces each one to earn its place. Questions that do not discriminate between a strong and a weak organization are removed.

The limitation is that 12 questions cannot cover a domain exhaustively. ODX assessments are diagnostic instruments that identify where to look more closely, not comprehensive audits.

Executive answerable

Every question must be answerable by a senior executive from what they personally know, without consulting a subordinate or retrieving a document.

A question such as "our average approval cycle time is under three days" fails this test, because most executives cannot answer it from memory. A question such as "we can state, from records, how long a routine approval takes" passes, because the executive knows whether that capability exists.

Behavioural anchoring

Where possible, questions ask about observable behaviour rather than general states.

"Authority levels match each role" invites an aspirational answer, because disagreeing is an admission of mismanagement.

"We review and adjust authority limits when the organization's size or structure changes materially" is harder to overstate, because it refers to something that either happened or did not.

Behavioural anchoring improves discrimination: it widens the distribution of scores and reduces the tendency of all respondents to cluster near the top.

Bilingual design

Every question, hint, section, maturity description, and report element exists in English and Arabic. Arabic is not a translation layer applied after the fact. Questions are written and reviewed in both languages, with attention to register: formal enough for board level use, direct enough to be understood without ambiguity.

Right to left layout is applied throughout the platform, including generated reports and presentations.

Framework alignment

ODX assessment design is informed by the OECD Principles of Corporate Governance and the COSO Internal Control framework.

These are used as design references, not compliance targets. ODX does not certify alignment with either framework, nor with any national code derived from them. The reason for anchoring to OECD and COSO rather than a single jurisdiction is that most national governance codes derive from these sources, which keeps ODX applicable across markets without claiming jurisdiction specific compliance.

3Answer scales

Every question uses a five point scale. The scale type is chosen to match the nature of the question, so that the middle points carry real meaning.

Agree

Used for questions about perception and shared understanding.

Frequency

Used for questions about recurring behaviour.

Readiness

Used for questions about whether a capability or artefact exists.

Compliance

Used for questions about adherence to a defined control or rule.

Capability

Used for questions about the level of a developed skill or function.

Mixing scale types within an assessment is intentional. A control question answered on an agreement scale loses precision, and a perception question answered on a compliance scale implies a formality that does not exist.

4Scoring methodology

Step 1: convert the answer to a 0 to 100 value

Each answer value from 1 to 5 is converted linearly:

score = ((value - 1) / 4) x 100

The scale label does not affect the arithmetic. Only the numeric value matters.

Step 2: calculate the section score

Each section score is the weighted average of its four question scores, using question weights:

section score = sum(question score x question weight) / sum(question weight)

Section weight is not applied here. It applies only at the overall level.

Step 3: calculate the overall score

Each question carries an effective weight combining its own weight and its section weight:

effective weight = question weight x section weight
overall score = sum(question score x effective weight) / sum(effective weight)

Both section and overall scores are rounded to whole numbers.

Worked example

Delegation of Authority, with an organization that answers as follows.

Section scores, using question weights only:

Clarity of authority: (75x1.50 + 25x1.50 + 75x1.25 + 75x1.00) / 5.25 = 61

Approval flow: (50x1.25 + 25x1.00 + 75x1.00 + 75x1.25) / 4.50 = 57

Control and oversight: (75x1.50 + 25x1.50 + 50x1.25 + 50x1.50) / 5.75 = 50

Overall score, using effective weights:

Total weighted score / total effective weight = 894.375 / 16.20 = 55

Maturity level 3, Defined.

Note that the overall score of 55 is lower than the simple average of the three section scores, which is 56, because Control and oversight carries the highest weight and scored lowest. This is the weighting system working as designed.

This example lands on 55, the first score in Level 3. That is not contrived, it is what these answers produce, and it is worth leaving in place: a single answer moved one step on Q9 or Q12 would drop the organization into Level 2, which is exactly why the boundary is published rather than inferred from a colour.

5The weighting system

Principle

Weight represents consequence of failure, not topic importance.

This distinction matters. Every domain in an assessment feels important to the people who own it. Consequence of failure is an objective test that can be defended to a reviewer.

Question weights

Weights are capped at 1.50. Above that, a single question would dominate a four question section and the section score would stop reflecting the domain.

Examples from Delegation of Authority:

Weight 1.00 applies to "Authority passes to a named alternate when a decision maker is unavailable." Failure here creates delay.

Weight 1.25 applies to "We review and adjust authority limits when the organization's size or structure changes materially." Failure here means limits set at an earlier scale become either an obstacle or an open door.

Weight 1.50 applies to "Related party transactions follow a defined approval route." Failure here is the most common source of governance failure in owner led businesses and carries direct legal exposure.

Section weights

Sections are weighted by risk concentration, not by question count.

The test that decides which is which is time to remedy, not importance. Every section of every assessment is important to the people who own it, so importance cannot separate them. How long a failure takes to fix, and whether it can be fixed at all once discovered, can.

The clearest worked case is IPO Readiness. Its governance section carries 0.90 and its financial section carries 1.20, which reads backwards if the test is importance: board composition is not less important than reporting. It is not backwards under time to remedy. Directors can be appointed and committee charters drafted in weeks. Three years of audited financial history cannot be manufactured at any price, so a gap there is absolute and time bound in a way a governance gap is not.

The same test explains the others. Customer understanding sits at 0.90 because an organization that misreads its customers this year can correct it next year; customer retention sits at 1.20 because a customer already gone cannot be corrected. AI governance sits at 1.20 because company data that has entered a public model cannot be withdrawn.

Effect on the score

For Delegation of Authority, the weighting produces these section shares of the overall score:

Control and oversight carries approximately 43% of the score. This is deliberate. Clarity and flow failures cost efficiency. Oversight failures cost the organization. For an assessment intended to indicate whether a framework would survive external scrutiny, weighting toward oversight is the honest choice.

All weights are published in section 9 of this document. An organization can reproduce its own score from its answers.

This system is applied in all nineteen assessments as of 29 July 2026. It previously existed in this document and in one assessment, which meant a reader who accepted the invitation above and tried it on any other assessment would have found uniform weights and no system to reproduce. That is corrected: 228 question weights and 57 section weights are set, and none is a default.

One known and accepted disagreement, recorded so it is not read later as an oversight. A consistency check compares, within each assessment, the section carrying the top section weight against the section whose questions carry the most weight. Both express risk concentration, so they should point at the same section, and in most assessments they do. In Governance they do not: Board and committees carries the top section weight of 1.20, while Transparency and risk holds two questions at 1.50 and sits at 1.00. Both were set with that assessment's content in view, one reading that a regulator checks board structure first and the other that the failures with legal exposure sit in transparency. A disagreement between two considered judgements is not an error, and it is left standing rather than resolved by arithmetic.

6The maturity model

The five levels

Level names in Arabic: أولي، قيد التطوير، محدّد، مُدار، استثنائي

Name changes, and names declined

On 6 August 2026 Level 5 was renamed from Optimized to Exceptional, and in Arabic from متميّز to استثنائي. On 7 August 2026 Level 1's Arabic name became أولي, with the English unchanged. No boundary moved in either change and no score changed band: both are labels on the same measurement.

Renaming Levels 3 and 4 has been proposed twice, in two different pairs of words, and declined both times. The reasoning is recorded here because it is about what the bands measure rather than about vocabulary.

Level 3 has been proposed as "advanced" and as "Mature". The behavioural table above defines the band as practices that are documented but not embedded, and Level 3 is the band this methodology records as being overstated most often. Any name asserting that development is complete contradicts the description directly above it. "Defined" is also the precise word for "documented", which is what the band measures.

Level 4 has been proposed as "distinguished" and as "Leading". Both are comparative: they describe a position relative to other organizations. This band is defined absolutely, against the answer frequency anchors, and ODX withholds peer comparison entirely until a peer group reaches the published minimum. A comparative name would therefore assert a comparison in reports that state, a page later, that no comparison is available. The band also spans 70 to 84, below the target standard of 85.

Either name becomes available only if the behavioural table above changes to support it, in the same revision. That is a decision about what the band measures, not a translation of what it is called.

How the boundaries are derived

The boundaries are not chosen for convenience. They are derived from the answer scale.

Because answers convert to 0, 25, 50, 75, and 100, the meaningful behavioural anchors are:

50, corresponding to "Sometimes", "Partially ready", "Partially compliant", or "Somewhat". This is the point at which a practice exists but is inconsistent.

75, corresponding to "Often", "Mostly ready", "Largely compliant", or "Agree". This is the point at which a practice is reliable.

The bands are positioned relative to these anchors:

The published colour key

Where ODX renders the bands as colour, the mapping is published here so that a colour on a screen or in a report can be checked against this document rather than inferred.

The colour is a key to the bands. It is not a severity scale. ODX does not claim that a lower band requires more urgent attention than a higher one. The bands are derived from the answer scale, which measures how consistently a practice is performed. Urgency depends additionally on what the practice protects, and that varies by domain: Level 2 in a compliance assessment and Level 2 in an innovation assessment sit in the same band and carry different urgency. ODX publishes no evidence that urgency is domain independent, and the uniform boundary argument above declines a weaker version of the same move.

For that reason the colours are not a red to green ramp. A hazard ramp places a neutral colour at Level 3, where it reads as arrival, and Level 3 means documented but not yet embedded. It is the band overstated most often, as the pass recorded below found in seven of nineteen assessments.

The sequence runs violet to yellow green, ordered and monotonic in lightness, so it survives greyscale reproduction and the common forms of colour vision deficiency. Red and green are not the poles, so no reading depends on distinguishing those two. Every cell carries the band name in words and the score in numerals: the colour supports the label and never replaces it.

Why the boundaries are uniform

The same boundaries apply to every ODX assessment. This follows from the derivation: because every assessment uses the same five point conversion, the behavioural anchors sit at the same scores regardless of domain.

Varying boundaries by assessment would require evidence that the same score means something different in different domains. No such evidence exists, and inventing one would make the model less defensible, not more.

Writing a level description, and the two tests that check one

Each level carries a characteristics line describing what an organization at that level looks like, and a risks line describing what is still exposed. The band that goes wrong most often is Level 3, because its published meaning is narrow and easy to overstate: practices are documented but not yet embedded.

A Level 3 pass over all nineteen assessments on 29 July 2026 rewrote seven of them. The two tests it used are recorded here in their corrected form, because both were first attempted as mechanical proxies and both proxies were wrong.

Test one. The risks line must carry a counterweight that qualifies the achievements the characteristics just stated. Not any risk: a risk that says the practice is not yet embedded. Board Effectiveness previously claimed committees function and papers are timely, and then listed a risk about board expertise and diversity, which is a real risk on a different axis and left the achievements standing unqualified.

Test one is not hedge word detection. The first attempt searched for "may", "still", "not yet" and similar, and it produced a false positive on IPO Readiness, whose risks line reads "Key-person dependency and undocumented processes raise diligence red flags". That is a counterweight. It names two things that are still absent and it qualifies the characteristics directly; it simply does so declaratively. A pass built on the proxy would have rewritten a correct Level 3.

Test two. The characteristics must not claim anything the next band claims. The first attempt used a fixed list of band 4 words: embedded, managed, proactive, integrated, continuous. That list is useful as a first sweep and it is not the test, because the boundary is per assessment.

The working form of test two is to read Level 3 against that assessment's own Level 4. That correction is what found Financial Governance, which claimed "controls are documented and tested". No word on the list appears in it. A control that has been tested is a control someone has proved operates, which is not a documented but not embedded state, and testing now sits in the Level 3 risks line as the thing that has not happened yet.

The model to write toward

Project Management Maturity is the model, and the reason is specific enough to copy.

Its Level 3 risks line says "Benefits realization tracking may be absent". Its own Level 4 characteristics say "Benefits realization is tracked and reported". The counterweight names the precise thing that must become true for the organization to leave the band, and the next band says that thing. A reader can see the boundary from either side.

Its four achievements are also structural existence facts rather than embedding claims: a methodology is used, a PMO exists, projects are tracked against plan, steering committees meet. Four simultaneous achievements look like an overclaim and are not, because none of them asserts that the practice has been internalised. That is the distinction the count of achievements cannot see.

7The ODX target standard

What it is

The ODX target standard is 85, the threshold at which an organization enters Level 5 Exceptional.

It is applied uniformly at both assessment and section level.

What it is not

The target standard is not an observation of what other organizations achieve. It is not an industry average, a peer benchmark, or a measurement of top performing companies.

It is a definitional threshold within the ODX maturity model: the score at which practices are embedded and self correcting rather than merely consistent.

This distinction is important and ODX states it explicitly wherever the standard appears. A claim about other organizations requires data about other organizations. A claim about a maturity threshold requires only a published model, which this document provides.

Why 85 and not another number

85 is the Level 5 boundary, and Level 5 is the level at which a practice no longer depends on individual diligence to hold. Setting the target lower would mean declaring an organization at target while its practices are still merely consistent. Setting it higher would place the target above the model's own top band.

Prior versions

Earlier versions of ODX displayed an "industry benchmark" and a "top performer level" derived from constants that had no measured basis. These have been removed. Any ODX report generated before this document's version date may contain those figures and should not be relied on for comparative claims.

8Peer benchmarking

How peer groups are formed

A peer group is built from other organizations that have completed the same assessment and whose anonymized results contribute to benchmarks.

Matching is applied in this order:

1. 1.Industry sector, employee count range, organization type, and headquarters country
2. 2.Industry sector, employee count range, and headquarters country
3. 3.Industry sector and employee count range
4. 4.Industry sector

The engine uses the narrowest group that meets the minimum size. The group description shown to the user always reflects the level actually used. A user is never shown a description implying a narrower peer set than the one their comparison was drawn from.

The requesting organization is always excluded from its own peer group.

Minimum group size

No peer comparison is displayed unless the peer group contains at least 10 distinct organizations.

Distinct organizations, not distinct results. Ten assessments completed by three companies is not a peer group of ten.

Below this threshold, ODX displays no percentile, no peer average, and no comparative claim of any kind. The report shows the score, the maturity level, and a statement that the peer benchmark is still forming.

This threshold is set to avoid two failures. First, a percentile calculated from a small group is statistically meaningless and would mislead. Second, a small group creates a re identification risk: in a group of two, the "peer average" is one other organization's exact score.

What is disclosed

When a benchmark is displayed, ODX shows aggregates only:

• Peer group average
• Top quartile
• Top decile
• The number of organizations in the group
• A description of how the group was defined

Progress toward the minimum group size

Below the minimum of 10 distinct organizations, no benchmark is displayed, per the threshold stated in this section. ODX discloses progress toward that minimum on a graduated basis, to keep the participant informed without exposing a peer group narrow enough to risk re-identification.

The floor of three exists because a count below it, combined with a narrow peer group definition (industry, size band, country), can approach identifying a specific other organization. This is not the re-identification risk named for a displayed benchmark, since no individual score is ever implied, but the same caution applies to the count itself.

What is never disclosed

• Any organization name
• Any individual organization's score
• The minimum or maximum score in the group
• Any distribution detail that would allow a single organization's score to be inferred

Contribution, and why this section is not called consent

Contribution is on by default. An organization's anonymized scores contribute to peer benchmarks from the moment it has results, and its owner or administrator can switch that off at any time in company settings. Organizations that switch it off are excluded from all peer calculations. They can still complete assessments and receive their own results.

How the choice is actually offered. Step 4 of registration carries a labelled control, "Contribute to benchmarking", with a bilingual explanation stating that results are included anonymously, that other executives benefit from more accurate benchmarks, and that the company name and identity are never revealed. It is pre-selected. The person registering can switch it off before submitting, and whatever they choose is what the company is created with.

Only the first person to register a company is asked. Colleagues who join later through an invitation code inherit the setting the company already has. The database default of on applies to companies created by any other path.

Correction, 29 July 2026. An earlier revision of this section asserted that there was no step at registration and that nobody was asked. That was wrong, and it was written into this document as a correction of a sentence that had been accurate. The original wording, that organizations declining during registration are excluded, was right. The error came from changing a published claim without opening the registration flow to check it, which is the same failure this document's own rules exist to prevent.

What is still true, and why the section is not called consent. A pre-selected control is a weaker thing than an opt in, and the database default of on covers every path that does not pass through that screen. So the honest description is an informed default: the choice is presented and explained, it is pre-selected, and it can be changed at any time by the company owner or administrator.

Disclosure at the point it matters, added 29 July 2026: before the first question of every assessment, the runner states in both languages who can see the answers, who can see the score, that ODX support staff can access results, and that anonymized scores contribute to benchmarks unless the company switches it off.

What is contributed. Overall score and section scores, and the organization identifier used only to count distinct organizations and remove duplicates. No names, no email addresses, no individual answers. No comparison is shown until the group reaches the minimum organization count described above.

9Assessment catalogue

This section documents each assessment: its purpose, the domains it covers, and the full question set with scales and weights.

All nineteen are documented here as of 29 July 2026, with every weight set by the rubric in section 5 and a share of overall score stated for each. No assessment carries a default or a placeholder weight.

An earlier version of this section marked six assessments as provisional and stated no share for them, because their weights were the positional defaults they shipped with rather than decisions. That is resolved: the weighting pass set 61 question weights and 15 section weights on 29 July 2026.

Delegation of Authority

Purpose. To indicate whether an organization's delegation framework would withstand outside scrutiny from an auditor, an underwriter, a new investor, or a fraud investigation.

Intended respondents. Chief executive, chief financial officer, board member, company secretary.

Design references. OECD Principles of Corporate Governance, COSO Internal Control framework.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Clarity of authority

Section 2: Approval flow

Section 3: Control and oversight

Organizational Maturity

Purpose. To indicate whether the operating model can carry the organization to its next stage of scale without the founder in every decision.

Intended respondents. Chief executive, chief operating officer, general manager, head of strategy.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Structure and roles

Section 2: Processes and standards

Section 3: Coordination

Governance

Purpose. To indicate whether the governance framework would satisfy a regulator, an incoming investor, or a listing adviser reading it cold.

Intended respondents. Chairman, board member, company secretary, chief executive.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Board and committees

Section 2: Policies and controls

Section 3: Transparency and risk

KPI Health Check

Purpose. To indicate whether the measures reaching leadership are trustworthy, few enough to act on, and actually changing decisions.

Intended respondents. Chief executive, chief financial officer, head of strategy, head of performance.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Measure quality

Section 2: Targets and accountability

Section 3: Review and use

Strategy Execution

Purpose. To indicate whether a stated strategy is actually being executed, or whether it exists as a document while the organization runs on last year.

Intended respondents. Chief executive, chief operating officer, head of strategy, board member.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Clarity and alignment

Section 2: Execution and ownership

Section 3: Tracking and learning

Leadership Systems

Purpose. To indicate whether leadership development and accountability run on systems, or depend on the individuals currently holding the roles.

Intended respondents. Chief executive, chief human resources officer, board member, owner.

Repositioned 29 July 2026. The title was Leadership Effectiveness Assessment, which claimed more than the twelve questions support. They measure whether leadership runs on systems, not how effective any leader is.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Direction and decisions

Section 2: People and teams

Section 3: Accountability and trust

HR Maturity

Purpose. To indicate whether the people function is running the disciplines a growing or listing organization is judged on, from workforce planning to nationalization compliance.

Intended respondents. Chief human resources officer, chief executive, head of talent.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Talent and resourcing

Section 2: Performance and reward

Section 3: Development and culture

Board Effectiveness

Purpose. To indicate whether the board functions as a board, rather than whether it is constituted like one. Governance asks whether the structure exists; this asks whether it acts.

Intended respondents. Chairman, board member, company secretary, chief executive.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Board Composition and Independence

Section 2: Board Oversight and Accountability

Section 3: Board Dynamics and Effectiveness

Risk Management Maturity

Purpose. To indicate whether the risk management system underneath the board actually works: identification, ownership, treatment and monitoring.

Intended respondents. Chief risk officer, chief financial officer, chief executive, audit committee chair.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Risk Identification and Assessment

Section 2: Risk Governance and Ownership

Section 3: Risk Response and Monitoring

Compliance Maturity

Purpose. To indicate whether the organization knows which obligations apply to it, can evidence them, and acts when testing finds a gap.

Intended respondents. Chief compliance officer, general counsel, chief executive, audit committee chair.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Compliance Framework and Governance

Section 2: Compliance Culture and Awareness

Section 3: Compliance Monitoring and Response

Financial Governance

Purpose. To indicate whether financial controls, reporting and oversight would withstand an external auditor, a lender, or a tax authority looking closely.

Intended respondents. Chief financial officer, audit committee chair, chief executive, board member.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Financial Controls and Compliance

Section 2: Financial Reporting and Transparency

Section 3: Financial Oversight and Decision-Making

Project Management Maturity

Purpose. To indicate whether projects are governed as a portfolio with evidence, or run individually on the competence of whoever is assigned.

Intended respondents. Chief operating officer, head of PMO, programme director, chief executive.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Project Methodology and Standards

Section 2: Project Governance and Oversight

Section 3: Project Delivery and Performance

Change Readiness

Purpose. To indicate whether the organization can absorb and sustain change, as distinct from whether it can launch it.

Intended respondents. Chief executive, chief human resources officer, transformation director, chief operating officer.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Change Leadership and Sponsorship

Section 2: Change Culture and Mindset

Section 3: Change Execution and Sustainability

Innovation Capability

Purpose. To indicate whether innovation is a funded discipline with a portfolio and measured returns, or a word used for whatever is new.

Intended respondents. Chief executive, chief innovation or strategy officer, head of product, board member.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Innovation Strategy and Leadership

Section 2: Innovation Culture and Environment

Section 3: Innovation Execution and Commercialization

Customer Centricity

Purpose. To indicate whether the customer is at the centre of how the organization operates, as distinct from how it describes itself.

Intended respondents. Chief executive, chief customer officer, chief marketing officer, customer experience director.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Customer Understanding and Insight

Section 2: Customer Experience and Service

Section 3: Customer Loyalty and Retention

IPO Readiness

Purpose. To indicate whether the artifacts a regulator, an underwriter or a diligence lawyer will demand can actually be produced today.

Intended respondents. Chief executive, chief financial officer, board member, investment director.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Governance and board

Section 2: Financial and reporting

Section 3: Organization and controls

AI Readiness

Purpose. To indicate whether the organization could adopt AI safely today: whether the data is reachable, the staff are prepared, and the exposure is being watched.

Intended respondents. Chief executive, chief technology officer, chief data officer, chief information security officer.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Data and infrastructure

Section 2: People and skills

Section 3: Governance and responsible AI

Digital Transformation Readiness

Purpose. To indicate whether digital investment is changing how the business works, as distinct from how much technology it has bought.

Intended respondents. Chief executive, chief technology officer, chief digital officer, transformation director.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Digital Strategy and Vision

Section 2: Digital Capability and Talent

Section 3: Digital Execution and Culture

ESG Readiness

Purpose. To indicate whether the organization can produce the environmental, social and governance figures an investor, a lender or a regulator will ask for.

Intended respondents. Chief executive, chief sustainability officer, board member, investor relations director.

Section weights.

Target standard. 85 at assessment and section level.

Section 1: Environmental Stewardship

Section 2: Social Responsibility

Section 3: ESG Governance and Reporting

Cross assessment rules

Rules established during content review that bind every assessment reviewed after them. They exist so that two assessments cannot measure the same construct and disagree, and so a respondent who takes several is not asked the same thing twice.

Owner dependency is measured once, in Organizational Maturity. Whether the organization can decide and transact without the owner or founder personally approving is the defining operating model question for a family business and a listing blocker for an IPO candidate, so it sits where the operating model is assessed. Governance, Board Effectiveness, IPO Readiness and Financial Governance may touch adjacent ground, for example the reserved powers of the board or the approval of related party transactions, but must not restate the dependency question itself. Established 28 July 2026.

Standalone completeness over minimal overlap. A respondent who takes only one assessment must still get a complete picture of the domain that assessment claims. Where two assessments legitimately reach the same territory, they differ by the question they ask of it rather than by one of them staying silent. Delegation of Authority asks whether authority is defined and controlled; Organizational Maturity asks whether decisions move. Established 28 July 2026.

Risk is split by altitude. Governance asks whether the board sees risk: does a written register reach it, on a cadence, and are issues surfaced rather than absorbed. Risk Management Maturity asks whether the management system underneath works: identification, ownership, treatment, and monitoring. Neither restates the other, and a respondent taking only one still gets a complete answer at that altitude. Established 28 July 2026.

Related party transactions are split by act. Delegation of Authority owns the approval route, meaning which authority level may approve such a transaction and under what limit. Governance owns board approval and disclosure, meaning whether the board itself sanctions the transaction and whether it is disclosed. These are different questions about the same transaction rather than one restated, so both may be asked. Established 28 July 2026.

A freed slot goes to the largest unclosed gap. When a duplicate is found and removed, the slot it frees goes to the highest priority unclosed gap in that assessment, not to a near variant of what was removed. The set stays at twelve, so a deletion is an opportunity to close something rather than an opportunity to rephrase.

Two worked instances. Governance carried a control question that duplicated Financial Governance; the freed slot went to the general assembly question, closing the shareholder governance gap that had been carried since the Governance review. Financial Governance carried an authorization limits question that duplicated Delegation of Authority; the freed slot went to digital tax reporting, which is the sharpest market specific question in the catalogue after nationalization.

If there is no gap worth filling, say so and leave the set at eleven for the author to rule on, rather than inventing a twelfth. Established 28 July 2026.

Structure versus function. The general form of several splits above, stated once so a future reviewer can apply it to a case not yet met. When two assessments reach the same object, one may own whether the structure exists and is configured correctly, and the other whether it actually functions. The structural question asks for a fact a regulator or a listing adviser would check. The functional question asks whether the thing acted, and it is answered by an event that either happened or did not.

The worked instance is Governance and Board Effectiveness. Governance asks whether the chairman holds no executive position, whether a third of the board is independent, and whether committees have written charters. Board Effectiveness asks whether the chairman sets the agenda, whether independent directors meet without executives present, and whether the audit committee met without management. Neither is left with a hole: a respondent taking only one still learns about chairmanship, independence and committees, at that altitude.

The test for whether the split is real: the functional question cannot be answered high unless something happened. If it can, it is a restatement of the structural question in different words. Established 28 July 2026.

Candidates to check. Compliance Maturity and Financial Governance both reach objects Governance already asks about structurally. Whether the same split applies is to be flagged when those assessments are reviewed and resolved across them, not decided inside a single review.

Corrective action is split by trigger. Organizational Maturity asks whether any accountability mechanism exists at all: when a target is missed, is an owner and a corrective action recorded. KPI Health Check asks whether a target miss specifically carries one. An organization can have the first without the second, and that difference is worth measuring. Established 28 July 2026.

Questions are single barrelled, without exception. A question that asks two things cannot be answered honestly by a respondent for whom the two differ, and the score it produces is uninterpretable. Where two clauses both matter and only one slot exists, the second clause goes to the deferred gap list rather than being carried inside another question. A rule with an exception is a rule that erodes. Established 28 July 2026.

The test for whether two clauses are one question: two clauses are one question when neither is meaningful alone. "A review concluded an initiative should stop, and it stopped" is one question, because concluding without stopping is not a state worth scoring above zero. "Processes are documented and each has a named owner" is two, because a respondent can hold either without the other and both are worth knowing. Established 28 July 2026.

Ownership is asked in three assessments and this is not drift. Organizational Maturity asks whether any accountability mechanism exists, KPI Health Check asks who owns a target, and Strategy Execution asks who owns a strategic goal. The objects differ, and under the standalone completeness rule a respondent taking any one of the three must still get a complete answer at that altitude. A future reviewer should not consolidate them. Established 28 July 2026.

When two assessments hold the same construct, it goes to the better anchored instrument. The ownership rules above decide which assessment owns a construct. This decides the harder case, where the rule points one way and the wording points the other, and it is the rule that actually resolved it.

The worked instance is Delegation of Authority Q5, "decisions within delegated limits are made without escalation to the chairman or chief executive". Naming the chairman gives the question real force in a family business, and there is a defensible reading on which it belongs to Delegation of Authority: it asks whether the delegation instrument is honoured, which is control rather than movement.

It went to Organizational Maturity anyway, and not because the owner dependency rule said so. It went because Organizational Maturity's version is anchored to a dated window while this one is a frequency judgement, so the same construct is measured better there. Delegation of Authority keeps the question in a role neutral form, which asks whether the delegation is honoured without restating who is depended upon.

The rule. Where two assessments reach the same construct and the ownership rules do not settle it cleanly, compare the instruments: an artifact, a dated event or a figure beats a judgement, and the construct goes to whichever assessment holds the stronger one. A rule that assigns a construct to an assessment measuring it worse has assigned the label and lost the measurement. Established 29 July 2026.

When the triggering event has not occurred, answer within the question's own terms. Every question anchored on an event carries this case, so it is ruled once here rather than per hint. It arose when three hints written in the same pass gave two different instructions for the same situation, which would have told a respondent taking Organizational Maturity and KPI Health Check to score an absence in one and to look further back in the other.

Where the question states a window and the event did not occur inside it, the absence is the finding and the answer is the floor. An organization that missed no target in a year is either not setting real targets or not tracking them, and that is more informative than a full score. A hint may not direct a respondent outside the window the question sets, because that silently redefines the question for that respondent only and makes the answer incomparable with everyone else's.

Where the question states no window, it asks about the most recent occurrence whenever it was, and the respondent answers on that. Only where the event has never occurred at all is the absence the finding. "The last structural change we made was completed within one quarter of the decision to make it" is unbounded by construction, so a floor answer for an organization that restructured two years ago would be wrong.

Three shapes, not two, and the difference decides the answer. Applying the rule across four passes surfaced a distinction that was being made implicitly and is recorded here so it is not rediscovered.

An existence question asks whether something happened at least once: "a review concluded an initiative should stop, and it stopped". Window stated and nothing happened means the floor.

A coverage question asks what proportion of a population met a standard: "every project started in the last 12 months had an approved charter". If the population is empty, the floor still applies, because an organization that started no projects in a year has no evidence of chartering. If the population is merely smaller than the question assumes, answer against the population that existed: a board that met twice answers Q5 against those two meetings, and the missing meetings are measured by the governance question that counts them, not twice here.

This is not a second inverted question. The ruling against using the zero reports inversion twice governs question *design*, meaning whether to write a new question whose high score depends on a bad event having happened. This governs how an absence is *read* in questions that are already anchored on an event, which is a different act and has to be consistent across the catalogue. Established 29 July 2026.

Question design: anchor on an artifact, not on a judgement

This was tested against six assessments and it overturned the hypothesis it started as, so it is stated as a finding rather than a preference.

The initial expectation was that score compression came from scale choice: that agree invites one sweeping judgement while compliance and frequency give a respondent somewhere honest to land short of the top. Six assessments do not support it. Eight agree questions produced wider distributions than seven did, and the relationship is not monotonic.

What predicted compression far better was how socially costly it is to answer low. The two assessments whose subject is personally embarrassing, the operating model the chief executive designed and the quality of leadership itself, produced identical 14 point bands. Assessments on subjects with a culturally available excuse produced wider ones.

The design rule that follows. Changing the scale does not rescue a question an executive cannot afford to answer low, because the respondent will find the top of whatever scale is offered. What works is removing the judgement from the question altogether:

• Ask for an artifact. A written delegation matrix, a documented grade, a risk

register, a register of interests.

• Ask for a dated event. Something that happened, or did not, in the last 12 or 24

months.

• Ask for a figure the respondent either holds or does not. A turnover rate, a

training spend, a quota status.

"Our controls prevent errors and misuse" cannot be answered low by anyone. "We can produce evidence that our financial controls operated in each of the last 12 months" can, and it fails honestly.

Scale choice remains hygiene: agree on a recurring behaviour is the wrong tool and should be frequency, and a factual state should be compliance. It is worth fixing. It is not the lever. Established 28 July 2026.

*Remaining assessments are documented as their content review is completed. Each will follow the same structure.*

10The business case computation

The results page carries one business case section: an estimate of what closing the gap to the ODX target standard is worth. This section states every constant behind it and what kind of claim it is.

What kind of claim this is

These constants are ODX judgements, not derived results. That distinguishes them from the maturity bands in section 6, whose boundaries follow from the answer scale and are therefore defensible without further argument. Nothing in the answer scale implies that a governance gap converts to a particular percentage of payroll.

They are published here so that a reader can see exactly what was assumed, substitute their own figures, and judge the output accordingly. They are not a forecast, and ODX does not represent them as one.

The inputs

Where no employee count is on file, every currency figure is withheld rather than estimated. The percentage and duration figures still render, because they do not depend on headcount.

The constants

All are stored in system_settings.roi_model and are adjustable by an administrator without a deploy.

workHoursPerYear was, until 4 August 2026, a literal inside the results page rather than a published constant: an undisclosed assumption inside a figure labelled Estimated. It is recorded here for that reason.

Ambition, and why there are three levels

Every figure scales linearly with the chosen level, so the full gap figure is what the reader sees at Ambitious. ODX does not recommend a level. It has no basis on which to say that a given organization should target half its gap rather than all of it, so the control belongs to the reader and the page shows the consequence of each choice rather than picking one.

This is also why the leave behind formats show all three rows rather than one: a document cannot carry a control, and picking a level silently would assert a recommendation ODX does not make.

One computation, not two

Until 4 August 2026 the results page carried two sections, showing two currency figures for the same organization on the same screen. They read as two independent estimates corroborating each other. They were one formula with two knobs, and they differed in two ways rather than one: the second applied the ambition factor, and it used whatever cost figure sat in its own input while the first used the 180,000 constant. Nothing on the page disclosed the second difference.

They are now one section with one set of inputs.

The design duration

Every assessment is written to twelve questions answerable without preparation and shaped to take about 10 minutes. That figure is a design target describing how the content is built, not a measured completion time, and the interface says "designed for" rather than "takes".

A measured median was computed on 4 August 2026 and not published: nine of the nineteen assessments had no completed sessions at all, a third of the sessions belonged to seeded demonstration users, and the remainder came from a single participant. A median drawn from one person is a population statistic in form only. It will be published when a sample supports it, with its size stated.

11Limitations

ODX states its limitations plainly because a diagnostic that overstates its own reliability is worse than no diagnostic at all.

Self assessment bias

Responses are self reported. Executives may answer aspirationally, particularly on questions where a low answer implies personal failure. ODX mitigates this through behavioural anchoring and by favouring questions with an observable referent, but it cannot eliminate it.

Single respondent limitation

A standard assessment reflects one person's view of the organization. Where the respondent's view diverges from operational reality, the score will reflect the view.

Team assessments, where multiple executives complete the same assessment independently, address this by revealing where leadership perceptions diverge. The divergence is often more informative than the score.

No verification

ODX collects no evidence. A respondent who states that a written delegation of authority document exists is not asked to produce it. ODX is not an audit and should not be presented as one.

Twelve question coverage

Twelve questions cannot cover a domain exhaustively. Each assessment identifies where attention is needed. It does not replace a detailed review of the areas it flags.

Section score volatility

With four questions per section, a single answer moving one point shifts the section score materially. Section level findings should be read as directional. Overall scores, drawing on all 12 questions, are more stable.

Benchmarks depend on participation

Peer benchmarks improve as more organizations participate. Until a peer group reaches the minimum of 10 organizations, no comparison is shown. Early users will see their score and maturity level without a peer comparison, and will be notified when their benchmark becomes available.

Modelled figures

Some report sections present modelled figures, such as estimated cost of the maturity gap. These are directional models built on stated assumptions, not measurements. Every such figure discloses its formula and its assumptions in the report itself. Where an assumption uses a value the organization has not provided, ODX states which value was assumed.

12Version history

Contact

Questions about this methodology can be directed to info@odxos.com.