11.1 Requirements Verification & Quality Characteristics

Key Takeaways

  • ECO Domain 3 Task 5 requires the business analyst to verify requirements by evaluating them against defined quality characteristics to ensure they are well-formed, technically sound, and fit for development.
  • Requirements Verification answers the foundational question: 'Are we building the requirements right?', focusing on internal consistency, structural correctness, and syntactic clarity rather than business outcome.
  • The seven essential quality characteristics for verified requirements are Unambiguous, Complete, Consistent, Correct, Feasible, Modifiable, and Testable (Verifiable).
  • Formal verification techniques range in rigor from informal desk checks and peer reviews to structured walkthroughs and formal Fagan inspections with discrete participant roles (Moderator, Author, Reader, Recorder, Inspector).
  • Defect detection during verification systematically targets omission, contradiction, ambiguity (such as weasel words and passive voice), and implementation bias (prescribing the technical 'how' instead of the business 'what').
Last updated: September 2026

11.1 Requirements Verification & Quality Characteristics

[!NOTE] PMI-PBA Examination Alignment: Domain 3 (Analysis) represents the single largest domain on the PMI-PBA examination, accounting for 35% of all scored items (~61–62 questions). Within Domain 3, Task 5 mandates that the business analyst "Verify requirements by evaluating requirements against defined quality characteristics to ensure they are well-formed, complete, and fit for development." Verification is the primary quality-assurance gate for business analysis deliverables. Questions frequently challenge your ability to distinguish between verification and validation, identify specific requirement defects (such as ambiguity or implementation bias), apply formal inspection rules, and assess requirements against the seven core quality characteristics.


The Verification Imperative: Building the Requirements Right

In requirements engineering, an enormous proportion of software failures, budget overruns, and schedule collapses originate not in faulty code, but in defective requirements. When ambiguous, contradictory, or incomplete requirements flow unchecked into technical architecture and development, the cost to remediate those defects multiplies exponentially as the solution advances through the lifecycle.

According to The PMI Guide to Business Analysis and Business Analysis for Practitioners: A Practice Guide, Requirements Verification is the deliberate process of reviewing requirements and models to ensure they meet rigorous quality standards, adhere to organizational templates and notations, and are sufficiently detailed, unambiguous, and technically sound to serve as an uncompromised foundation for design, construction, and testing.

The Foundational Dichotomy: Verification versus Validation

The PMI-PBA examination strictly differentiates between Verification and Validation:

  • Verification answers: "Are we building the requirements right?" (Syntactic, structural, and compositional quality). It evaluates the internal quality of the documentation, models, and specifications. It asks whether the requirement statement is clear, testable, modular, and compliant with standards.
  • Validation answers: "Are we building the right requirements / right product?" (Semantic, business, and strategic value). It evaluates whether the requirement satisfies the business need, achieves the business case KPIs, and delivers stakeholder value.

A requirement can be perfectly verified (flawlessly written, unambiguous, highly testable, grammatically pristine) yet completely invalid (it delivers a capability that no user wants, does not align with corporate strategy, or solves the wrong problem). Conversely, an idea can be valid (a brilliant concept that would save millions) yet unverified (expressed as a vague, incomplete, untestable paragraph that cannot be coded).

+-----------------------------------------------------------------------------------+
|                         The Quality Verification Gateway                          |
+-----------------------------------------------------------------------------------+
| Raw Elicitation Findings / Draft User Stories / Initial Models                    |
|                                    ↓                                              |
| [ VERIFICATION GATE: Evaluated against 7 Quality Characteristics & Syntax Rules ]  |
|                                    ↓                                              |
| Verified Requirements: Well-Formed, Complete, Unambiguous, Testable, Feasible    |
|                                    ↓                                              |
| [ VALIDATION GATE: Evaluated against Business Case, KPIs & Stakeholder Value ]    |
|                                    ↓                                              |
| Approved Requirements Baseline (Ready for Architecture, Sizing & Delivery)       |
+-----------------------------------------------------------------------------------+

The Seven Essential Requirements Quality Characteristics

To pass verification, every functional requirement, non-functional requirement, user story, and model must satisfy seven internationally recognized quality criteria. Certified business analysts deconstruct draft statements against these benchmarks:

1. Unambiguous

  • Definition: The requirement is written such that it can be interpreted in only one way by every reader, regardless of their background (developer, tester, business sponsor, or compliance officer).
  • The Flaw: Reliance on subjective adjectives, vague qualifiers, and "weasel words" (e.g., rapid, robust, user-friendly, seamless, optimized, as appropriate, if possible, state-of-the-art).
  • Bad Example: "The system shall provide rapid search response times for customer records and display results in a user-friendly format."
  • Good Example: "The system shall display the search results matching the customer query within 1.5 seconds of user submission for a database containing up to 10,000,000 active customer records under a concurrent load of 500 active search queries."

2. Complete

  • Definition: The requirement encapsulates all information necessary for designers and developers to build the capability and for quality engineers to test it. It defines all inputs, expected outputs, operational boundary conditions, data constraints, and explicit exception-handling behaviors. It contains no missing scenarios, unresolved questions, or "To Be Determined" (TBD) tags.
  • The Flaw: Stating the "happy path" while completely ignoring alternative paths, boundary limits, and error scenarios.
  • Bad Example: "The system shall calculate and disburse the annual bonus for all eligible employees based on their performance tier."
  • Good Example: "The system shall calculate the annual bonus for all active, full-time employees employed as of December 31 using the formula: Base Salary × Performance Tier Percentage (Tier 1 = 15%, Tier 2 = 10%, Tier 3 = 5%, Tier 4 = 0%). If an employee has been employed for less than 12 months, the calculated bonus shall be pro-rated based on the number of full calendar months worked. Inactive or terminated employees shall receive $0.00."

3. Consistent

  • Definition: The requirement does not contradict, conflict with, or duplicate any other requirement, business rule, organizational policy, or standard domain terminology.
  • The Flaw: Internal contradictions between modules, or conflicting rules defined by disparate stakeholder groups that the analyst failed to harmonize.
  • Bad Example: Requirement REQ-INV-104 states: "The system shall allow warehouse managers to edit and update line-item quantities on posted inventory delivery manifests at any time." while Requirement REQ-AUD-302 states: "All posted inventory delivery manifests are immutable legal audit records and shall not be edited or deleted under any circumstance."
  • Good Example: "Posted inventory delivery manifests shall be immutable and read-only. Corrections to posted delivery manifests shall be executed exclusively by creating a linked Supplementary Discrepancy Note authorized by a warehouse manager."

4. Correct

  • Definition: The requirement is factually accurate, technically sound, and accurately conveys the true business reality, legal statute, or operational capability needed. Both business and technical authorities verify its truthfulness.
  • The Flaw: Relying on outdated policies, misinterpreting statutory regulations, or documenting false assumptions as business rules.
  • Bad Example: "The portal shall automatically archive and permanently delete all patient medical records exactly three years following the patient's last clinical encounter to minimize cloud storage expenses."
  • Good Example: "The portal shall retain adult patient medical records in an active, encrypted archive for a minimum of seven (7) years following the patient's last clinical encounter, in accordance with HIPAA § 164.316 and State Medical Board record retention mandates."

5. Feasible (Achievable)

  • Definition: The requirement is technically, operationally, and legally capable of being implemented within the existing constraints of the organization's technology stack, architecture, project budget, and delivery schedule.
  • The Flaw: Demanding impossible capabilities, violating the laws of physics, or specifying features requiring non-existent technologies or budget-breaking infrastructure.
  • Bad Example: "The mobile application shall perform local biometric retinal scans using standard low-resolution smartphone front cameras with 100.00% verification accuracy under complete darkness without requiring an active internet connection."
  • Good Example: "The mobile application shall authenticate registered users using the mobile operating system's native biometric APIs (iOS FaceID/TouchID or Android BiometricPrompt), falling back to a 6-digit cryptographic PIN if biometric hardware is unavailable or unconfigured."

6. Modifiable (Maintainable)

  • Definition: Requirements are structured, formatted, and modularized such that changes can be made easily, consistently, and without unmanaged ripple effects. This requires unique identifier tags, hierarchical decomposition, atomic statements, and explicit cross-referencing.
  • The Flaw: Huge compound paragraphs combining multiple business rules, database actions, and UI designs into an indivisible wall of text.
  • Bad Example: A single unnumbered 500-word block of narrative text that describes user authentication, password resets, session timeouts, database lockout tables, and audit reporting all in one tangled paragraph.
  • Good Example: Distinct, modular, atomic requirements:
    • REQ-AUTH-01: User Authentication Protocol
    • REQ-AUTH-02: Account Lockout Parameters
    • REQ-AUTH-03: Password Reset Workflow
    • REQ-AUTH-04: Session Inactivity Timeout

7. Testable (Verifiable)

  • Definition: There exists a finite, cost-effective, objective process (inspection, analysis, demonstration, or automated/manual test) by which a person or machine can verify that the delivered solution satisfies the requirement. A requirement that cannot be tested cannot be verified.
  • The Flaw: Subjective criteria, open-ended aspirations, or metrics that cannot be empirically measured.
  • Bad Example: "The claims processing workflow shall be intuitive, aesthetically pleasing, and minimize user frustration for newly onboarded adjusters."
  • Good Example: "A newly onboarded claims adjuster who has completed standard 4-hour onboarding shall be able to process a standard auto property claim from intake to adjudication in under 8 minutes with zero data entry validation errors, achieving a first-time success rate of at least 90% across 20 usability testing trials."

Requirements Quality Characteristics & Verification Checklist

The following checklist provides an authoritative operational reference for conducting requirements verification audits:

Quality CharacteristicCore Evaluation Criteria / Inspection QuestionDefect Indicators & Red FlagsRemediation & Best Practice Pattern
UnambiguousCan this requirement be interpreted in more than one way by different readers?Use of weasel words (rapid, robust, seamless, flexible, intuitive, user-friendly, optimized, as needed); passive voice omitting the actor.Replace subjective qualifiers with precise, quantitative engineering metrics, SLAs, and explicit actor-action-target syntax.
CompleteAre all operational conditions, edge cases, error states, and data ranges fully specified?Missing "else" conditions; absent error handling; unstated default values; presence of "TBD" or "etc."; unstated data boundaries.Apply decision tables to account for all input permutations; define explicit system behavior for exceptions, timeouts, and boundary limits.
ConsistentDoes this requirement contradict, duplicate, or conflict with any other statement or standard?Conflicting business rules across departments; contradictory non-functional constraints; inconsistent field naming or domain taxonomy.Cross-reference requirements against the project glossary and data dictionary; resolve conflicts across stakeholders before baselining.
CorrectDoes this requirement represent factually true business rules, operational realities, and legal statutes?Documenting outdated policies; misinterpreting legal/regulatory requirements; author guessing operational workflow without SME confirmation.Validate factual statements directly with legal, compliance, and authoritative domain SMEs; cite formal regulatory policy source documents.
FeasibleCan this requirement be built within the known technical, architectural, schedule, and budgetary limits?Architecture team flags technical impossibility; requires unproven third-party algorithms; exceeds allocated infrastructure budgets.Conduct technical feasibility spikes; consult solution architects early; decompose requirement into achievable iterative increments.
ModifiableCan this requirement be modified in the future without corrupting adjacent specifications?Massive compound paragraphs containing 5+ distinct functional rules; unnumbered statements; deeply tangled dependencies.Enforce atomicity (one requirement per statement); assign unique alphanumeric identifiers (e.g., REQ-FIN-042); maintain traceability links.
TestableIs there a finite, objective test or demonstration that yields a definitive pass/fail result?Inability to write a binary acceptance test; absence of measurable criteria; subjective satisfaction metrics ("satisfies executive management").Structure acceptance criteria using Given-When-Then (BDD) syntax; specify exact numerical thresholds, tolerances, and verification methods.

Detecting Common Requirements Defects

During verification, the business analyst actively hunts for four pervasive categories of defects:

+===================================================================================+
|                      Common Requirement Defect Categories                         |
+===================================================================================+
| 1. OMISSION: The missing condition, unstated exception, or unhandled failure state |
| 2. CONTRADICTION: Opposing rules or conflicting constraints across documents      |
| 3. AMBIGUITY: Multiple interpretations, vague adjectives, passive voice actors    |
| 4. IMPLEMENTATION BIAS: Prescribing the technical "how" instead of business "what"|
+===================================================================================+

1. Defect of Omission

Omission is the most frequent and costly defect in requirements analysis. It occurs when the author documents what happens when things go right, but completely omits what must happen when things go wrong.

  • Example: A requirement specifies that when an applicant enters a valid Social Security Number (SSN), the system pulls an external credit report. The requirement omits: What happens if the credit bureau API times out after 10 seconds? What happens if the credit bureau returns an 'SSN Not Found' error? What happens if the applicant is a foreign national with an Individual Taxpayer Identification Number (ITIN)?
  • Remediation: The business analyst utilizes structured techniques such as boundary value analysis, decision trees, and state-transition models to systematically uncover every unhandled branch.

2. Defect of Contradiction

Contradiction occurs when two or more statements within the requirements baseline cannot be satisfied simultaneously.

  • Example: A functional requirement states that all customer billing addresses must be verified against the United States Postal Service (USPS) address database. Simultaneously, an international expansion requirement states that customers residing in Canada and the United Kingdom must be able to register and purchase products using their local postal codes.
  • Remediation: Traceability matrices and requirements management tools must be used to perform cross-dependency queries, ensuring that rules governing the same entity (e.g., Customer Address) are evaluated collectively.

3. Defect of Ambiguity & Linguistic Traps

Ambiguity enters requirements through imprecise natural language. The business analyst must purge requirement documents of common linguistic traps:

  • Passive Voice: Hides the actor responsible for the action. "The loan file shall be approved within 24 hours." (Who approves it? The loan officer? The branch manager? An automated underwriting engine?)
  • Weasel Words: Words that appear to state a requirement but provide no measurable commitment. "The system should provide adequate security features as appropriate."
  • Open-Ended Clauses: Clauses that invite scope creep or leave requirements undefined. "The export module shall generate reports in CSV, PDF, Excel, etc." (What does 'etc.' include? XML? JSON? Fixed-width text? Never permit 'etc.' or 'and so forth' in verified requirements).

4. Implementation Bias: The "How" versus the "What"

One of the most heavily tested topics on the PMI-PBA examination is Implementation Bias (premature design). Requirements must state what the business or user needs to achieve, never how the technical solution will be engineered.

When a business analyst prescribes implementation details:

  1. It artificially constrains the technical architecture team, preventing them from selecting the most efficient, scalable, or modern technology.
  2. It makes the requirements fragile; if the underlying technical framework changes, the business requirements must be rewritten.
  3. It often confuses a single design option with the underlying business necessity.
Flawed Requirement (Implementation Bias - The "How")Verified Requirement (Business Need - The "What")
"The system shall store customer registration data in an Oracle 19c relational database table using a PL/SQL stored procedure that encrypts the password with SHA-256.""The system shall securely persist customer registration credentials such that passwords are stored using an industry-standard irreversible cryptographic hashing mechanism."
"The user shall click a blue dropdown menu created with React.js to select their state of residence.""The system shall require the user to specify their state of residence from a standardized list of US jurisdictions."
"The platform shall send an Apache Kafka message event to topic 'payments_queue' containing an XML payload.""The platform shall publish payment transaction notifications asynchronously to downstream financial accounting services within 500 milliseconds of payment capture."

Formal Verification Review Techniques

Requirements verification is not an informal, casual proofreading exercise. It is a formal quality control gate executed using proven review techniques that vary in their structure, expense, and defect-detection rigor:

+-----------------------------------------------------------------------------------+
|                         Verification Rigor Spectrum                               |
+-----------------------------------------------------------------------------------+
| [LOW RIGOR]                                                          [HIGH RIGOR] |
| Desk Check  ───>  Peer Review  ───>  Structured Walkthrough  ───>  Fagan Inspection|
| (Author only)    (1 Colleague)        (Cross-functional team)      (Formal roles, |
|                                                                    metrics & data)|
+-----------------------------------------------------------------------------------+

1. Desk Check (Author Self-Check)

  • Format: Informal review performed exclusively by the author of the requirements before sharing the document with anyone else.
  • Process: The analyst reviews their own specifications against a standard organizational checklist (e.g., verifying that all IDs are unique, no TBDs exist, and formatting adheres to corporate templates).
  • Limitation: Lowest defect detection rate. Authors suffer from cognitive blind spots and tend to read what they intended to write rather than what is actually written on the page.

2. Peer Review (Buddy Check)

  • Format: Informal or semi-formal review conducted by one or two business analyst colleagues.
  • Process: A peer analyst who did not write the specification reviews the artifact to detect ambiguity, inconsistencies, or structural flaws. They provide constructive markup or margin notes.
  • Advantage: Fast, low-overhead method to catch obvious syntactic errors and terminology inconsistencies before exposing the document to technical and business stakeholders.

3. Structured Walkthrough

  • Format: A formal or semi-formal review meeting organized and led by the author of the requirements, attended by a cross-functional group of stakeholders (business SMEs, developers, architects, and QA engineers).
  • Process: The author walks the audience through the requirements document or model step-by-step, explaining the rationale, business rules, and flow. Attendees ask clarifying questions, challenge assumptions, and highlight potential defects.
  • Primary Purpose: Knowledge sharing, early validation of technical feasibility, and defect discovery.
  • Exam Trap: A walkthrough is not an approval or sign-off meeting. Its primary objective is educational alignment and identifying defects that must be resolved prior to formal baseline reviews.

4. Formal Software Inspection (Fagan Inspection)

Developed by Michael Fagan at IBM, the Fagan Inspection is the most formal, rigorous, and mathematically documented verification technique in software engineering and business analysis. Studies demonstrate that formal inspections catch between 60% and 90% of all requirements defects prior to software construction.

A formal inspection is strictly governed by six discrete procedural phases and enforces rigid participant roles:

The Six Phases of Formal Inspection

  1. Planning: The inspection coordinator (Moderator) evaluates the document against entry criteria (e.g., document completed, spell-checked, formatting compliant). If entry criteria are met, the Moderator forms the inspection team and schedules meetings.
  2. Overview: (Optional). The Author presents a brief high-level overview of the project context and architectural scope to the inspection team.
  3. Individual Preparation: Critical Phase. Each inspector independently and meticulously reads the requirements document line-by-line prior to the meeting, comparing every requirement against a formal defect checklist. Inspectors log every suspected defect on their preparation log. Rule: If inspectors have not completed individual preparation, the Moderator must cancel the inspection meeting.
  4. Inspection Meeting: The team convenes for a strictly time-boxed meeting (typically 90 to 120 minutes max to prevent fatigue). The Reader paraphrases the requirements line by line. As defects are raised, the Scribe records them. Solutions are never debated during the meeting; only defects are logged.
  5. Rework: The Author resolves all documented defects, correcting ambiguous language, clarifying business rules, and updating models.
  6. Follow-Up: The Moderator independently inspects the reworked document to ensure every logged defect has been properly remediated without introducing new secondary defects. If exit criteria are met, the document is officially certified as verified.

The Five Discrete Inspection Roles

On the PMI-PBA examination, questions frequently test your knowledge of who does what in a formal inspection:

  • Moderator (Inspection Leader): The neutral facilitator who chairs the process. Must not be the author. Manages the meeting, enforces inspection rules, maintains pace, prevents team members from arguing or designing solutions, and determines whether the document passes or requires re-inspection.
  • Author: The person who wrote the requirements artifact. Attends the meeting to answer factual clarifying questions and observe defects. Cardinal Rule: The Author must never act as Moderator, Reader, or Scribe, and must never defend their writing or debate whether a defect is valid.
  • Reader: A team member (often a developer or tester) who reads or paraphrases the requirements document aloud to the group, sentence by sentence, in their own words. Paraphrasing immediately exposes whether the requirement is unambiguous; if the Reader's interpretation differs from the Author's intent, an ambiguity defect is instantly exposed!
  • Recorder (Scribe): The participant responsible for logging each identified defect, its classification (omission, ambiguity, contradiction), and its severity on the formal Defect Log. Must be accurate and objective.
  • Inspector (Reviewer): Cross-functional participants (QA test leads, solution architects, domain SMEs) who evaluate the artifact strictly from their professional perspective to identify defects, risks, and non-compliance.
Loading diagram...
Formal Fagan Inspection Workflow and Participant Roles
Test Your Knowledge

A lead business analyst is auditing a draft Software Requirements Specification (SRS) for a digital wealth-management system. The analyst encounters the following requirement statement: 'The transaction engine shall securely record client portfolio changes in a MongoDB NoSQL cluster using JSON documents transmitted via an encrypted REST API, responding within a reasonable timeframe.' What two distinct requirements quality defects are present in this statement, and how should they be remediated?

A
B
C
D
Test Your Knowledge

During a formal Fagan inspection of a complex regulatory reporting requirements document, the meeting devolves into an intense argument between the lead software architect and the requirements author. The architect contends that a particular data aggregation requirement cannot be satisfied using the current enterprise service bus, and spends 25 minutes drawing alternative architectural microservice topologies on the whiteboard while the author defends their writing. According to standard Fagan inspection rules, what procedural breakdown occurred, and what action should have been taken?

A
B
C
D
Test Your Knowledge

A business analyst receives a functional specification for an automated insurance claims intake portal. Requirement REQ-CLM-204 states: 'The portal shall validate the claimant's policy number upon submission. If the policy number is valid, the claim shall be routed to an adjudicator.' During verification, the quality assurance lead points out that the requirement fails to define what happens if the policy number is expired, canceled, formatted incorrectly, or not found in the database. Which requirements quality defect has been identified, and what verification technique is most effective at preventing this type of defect?

A
B
C
D