The Customer Evidence and Segment Priority Standard: Choosing Whom the Organisation Must Understand, Serve and Retain

Different customers have different needs, constraints, buying behaviour and costs to serve. Use research and commercial evidence to decide whose unmet needs deserve attention and where to invest in the offer, service or customer relationship.

This standard turns customer evidence into a bounded segment-priority decision. It does not manufacture a universal taxonomy or declare that a statistically visible group is automatically attractive, reachable, ethical to target or feasible to serve.

1. The Market Connection Context

Segment choice affects where the organisation spends scarce attention. A weak choice can place acquisition money behind low-fit demand, ask sales to pursue work the operating model cannot deliver, under-serve high-value or high-need customers, conceal exclusion, and confuse customer count with durable value exchange.

The governed path is:

Market or stakeholder condition → evidenced job and context → distinguishable behaviour and need → service and economic reality → priority decision → value exchange and learning

The method keeps six lenses separate:

• need and job: what the person or organisation is trying to achieve and in which context;
• behaviour and journey: what they do, choose, delay, combine, abandon or work around;
• value exchange: what value they seek, contribute or enable and on what defensible basis;
• service and operability: what the organisation must reliably do to serve them;
• access, inclusion and risk: who is excluded, harmed, overlooked or burdened by the proposed boundary; and
• strategic relevance: why understanding or serving this group matters to the chartered decision now.

No lens automatically overrides the others. Revenue history cannot erase future need or exclusion. Stated preference cannot prove behaviour. A large group is not necessarily reachable or serviceable. A high-cost group is not automatically low value, especially where public, social, regulated or strategic obligations apply.

The decision to make

For each bounded segment hypothesis, the authorised forum selects one primary disposition:

• priority now for the named decision and resource boundary;
• priority for research because the evidence gap is material;
• serve through a differentiated route where needs or operating conditions materially differ;
• retain as an observed context without current priority status;
• combine with another hypothesis when the proposed distinction does not change action;
• split when one label hides materially different needs, behaviours, economics or safeguards;
• defer until a named dependency or evidence condition changes; or
• stop or reject when the boundary is unsupported, harmful, unusable or irrelevant to the decision.

The disposition expires. It does not authorise discriminatory treatment, a campaign, product build, pricing action, customer exclusion, service withdrawal or investment beyond the stated next gate.

Scope and exclusions

Use the standard when leadership can name the decision that segment priority must change: research allocation, offer focus, service design, retention attention, channel coverage, proposition testing or a later growth investment gate.

It can reconcile primary research, observed behaviour, journey and service evidence, sales and support records, commercial or funding evidence, accessibility and inclusion conditions, external context and existing segment models. It excludes automated individual decisions, sensitive-attribute profiling without qualified approval, final market sizing, price setting, media targeting, product roadmap approval, statutory equality or privacy opinion, and unsupported forecasts of demand, retention, revenue, margin or lifetime value.

Expected result

The result is an evidence-linked decision pack: a charter and research plan; customer evidence register; segment hypothesis and comparison ledger; priority decision and learning register; six separate measurement contracts in the Segment Measurement Contract Register; explicit source and proof limits; downstream handoff; and review or retirement conditions.

2. The Leadership vs. Delegation Boundary

Leadership owns why a segment decision is needed, the opportunity cost of attention, obligations to affected people, acceptable evidence, resource limits and the final priority. Researchers, analysts, marketers, sales leaders, service designers, customer teams, finance and data specialists can gather and challenge evidence. They cannot turn a cluster, persona, account tier or model output into organisational priority without authorised human judgement.

Internal leadership can lead when

• the decision owner can access behavioural, research, service and commercial evidence across functions;
• research participation and data use have an approved route;
• the team can distinguish observed patterns from inferred explanations;
• finance, service and customer owners can reconcile the value exchange;
• affected and excluded groups can be represented credibly; and
• no preferred segment, campaign or supplier controls the conclusion.

Independent challenge is justified when

• a high-value label is based mainly on historical revenue or internal account status;
• commercial teams compete for the same customer definition or budget;
• a proposed boundary uses sensitive, proxy or potentially exclusionary attributes;
• non-users, lost customers, assisted routes or underserved groups are absent from internal data;
• sample, survivorship or channel bias could materially change priority; or
• the organisation needs neutral facilitation across customer, finance, operations and data authority.

The reader recognises that segment choice requires primary research, behavioural evidence, commercial value, service reality and executive trade-offs, not a marketing workshop.

Human authority and safeguards

Named people approve the research question, participant route, data purpose, use of attributes, economic method, inclusion boundary, priority decision and any downstream action. A segmentation model may reveal a pattern; it cannot decide who deserves service, investment or exclusion. Qualified privacy, legal, accessibility, equality, ethics, sector and research review remains with the organisation's authorised specialists.

Use a capable internal team where it can test contrary evidence and represent affected groups. Use an independent or narrower specialist where the team has conflicts, lacks inclusive research capability or needs a bounded privacy, accessibility, ethics, statistical or serviceability review. Stop or escalate when no proportionate route can protect participants or preserve the decision boundary. Ask any provider: who owns participant authority and complaints; how are contrary and excluded cases retained; what burden moves to customers or staff; which artefacts and definitions remain with the organisation; and exactly where does the provider return the decision rather than make it?

Stop or narrow the work when participation is coercive, evidence use exceeds permission, a proxy exposes protected or sensitive information, a priority decision could withdraw essential access without qualified authority, a supplier cannot reveal method or limitations, or the sample excludes groups material to the decision.

3. The Step-by-Step Execution Standard

Step 01: Charter the segment-priority decision

Purpose and decision served

Turn a broad request for “customer segments” into one decision about where evidence and organisational attention should go.

Accountable owner

The CEO, CMO or customer executive holding the resource decision.

Delegated lead and contributors

Segment lead with customer research, service, sales, marketing, finance, data, accessibility and information authorities.

Preconditions, inputs and access

A important decision, affected market or stakeholder boundary, initial evidence inventory and willingness to stop a favoured hypothesis.

Actions

1. State the resource or operating decision segment priority must support.
2. Define market, geography, offer, relationship, population and period boundaries.
3. Name included non-users, former users, intermediaries and assisted routes where relevant.
4. Record assumptions, prior segment models, conflicts and no-change consequence.
5. Approve dispositions, exclusions, stop rules and expiry conditions.

Evidence to retain

Charter, scope map, prior-model inventory, conflict register and sponsor approval.

Decision or scoring rule

Proceed only when different priority outcomes would cause different authorised action.

Output and template

Segment Decision Charter and Research Plan.

Acceptance gate

The decision owner, research authority and receiving business owners accept the same boundary and use.

Leadership checkpoint

Confirm the work is not merely renaming an existing sales list or campaign audience.

Human control point

The decision owner approves purpose and can stop a harmful or irrelevant boundary.

Failure modes and safeguards

Failure modes include “everyone,” product-led labels and a predetermined priority. Safeguard with a decision-use test and explicit alternatives.

Change, stop or escalation condition

Re-charter when the resource decision, population, offer or obligation materially changes.

Handoff to the next step

Transfer the approved charter and unresolved authority questions to Step 02.

Step 02: Assign research, data, commercial and inclusion authority

Purpose and decision served

Make collection, interpretation, sensitive use and priority accountable before evidence is combined.

Accountable owner

The executive decision owner.

Delegated lead and contributors

Governance lead with research, information, privacy, accessibility, finance, service, sales and customer authorities.

Preconditions, inputs and access

Approved charter, data inventory, participant routes, known obligations and conflict declarations.

Actions

1. Assign one named priority owner and decision forum.
2. Assign participant, research-method, information, attribute-use, finance and service authorities.
3. Define approval, challenge, recusal, containment and stop rights.
4. Review sensitive, protected, proxy, inferred and third-party attributes before use.
5. Approve retention, access, anonymisation, contact and complaint routes.

Evidence to retain

Authority matrix, permissions, data-purpose decisions, attribute register, conflicts and recusals.

Decision or scoring rule

No attribute or derived group may be used beyond its authorised purpose and decision boundary.

Output and template

Authority and attribute rows in the Segment Decision Charter and Research Plan.

Acceptance gate

Every decision-critical evidence and safeguard decision has a named human authority.

Leadership checkpoint

Review whether the proposed priority could affect access, price, service or treatment.

Human control point

Qualified organisational authorities approve or reject sensitive and consequential uses.

Failure modes and safeguards

Failure modes are blanket consent, proxy blindness and analyst self-approval. Safeguard with purpose-specific decisions and auditability.

Change, stop or escalation condition

Pause unauthorised data use, participant contact or high-consequence classification immediately.

Handoff to the next step

Transfer authorised attributes, purposes and access limits to Step 03.

Step 03: Define the evidence grammar and value exchange

Purpose and decision served

Prevent customers, accounts, users, transactions, households and needs from being treated as interchangeable units.

Accountable owner

The customer decision owner.

Delegated lead and contributors

Research and data leads with finance, service, sales and metric owners.

Preconditions, inputs and access

Charter, source definitions, relationship identifiers, economic methods and service outcome definitions.

Actions

1. Define person, organisation, account, user, beneficiary, payer and intermediary as relevant.
2. Define need, job, behaviour, outcome, retention, contribution, service burden and relationship period.
3. Separate observed fact, participant account, modelled value and inferred explanation.
4. Reconcile identifiers, populations, periods, exclusions and source lineage.
5. Mark conflicts and missing baselines REQUIRES VALIDATION.

Evidence to retain

Definition register, identity and unit decisions, metric contracts and unresolved conflicts.

Decision or scoring rule

A definition is fit only for its stated population, period and decision use; one account record is not automatically one customer.

Output and template

Definition rows in the Customer Evidence and Need Register.

Acceptance gate

Research, finance, service and data owners accept decision-critical definitions or a validation gate.

Leadership checkpoint

Confirm the value exchange includes what the organisation must do, not only what it receives.

Human control point

Finance and customer authorities approve economic and service interpretations.

Failure modes and safeguards

Failure modes are lifetime-value shorthand, account-person confusion and transaction-only value. Safeguard with explicit units and obligations.

Change, stop or escalation condition

Split the decision where incompatible units or populations cannot be reconciled.

Handoff to the next step

Transfer the grammar and source gaps to Step 04.

Step 04: Inventory existing evidence and blind spots

Purpose and decision served

Understand what each source can reveal and which people or behaviours it systematically misses.

Accountable owner

The evidence authority.

Delegated lead and contributors

Evidence coordinator with CRM, analytics, sales, service, finance, research, partner and external-data owners.

Preconditions, inputs and access

Approved evidence grammar, source permissions and a list of material customer contexts.

Actions

1. Register each source's native unit, population, period, collection method and owner.
2. Record known missingness, channel, survivorship, non-response and historical-policy bias.
3. Separate current users from lost, declined, non-consuming and unsupported people.
4. Identify duplicated or dependent sources that cannot corroborate one another.
5. Prioritise evidence gaps that could reverse the decision.

Evidence to retain

Source inventory, lineage, access decision, coverage map and evidence-gap ranking.

Decision or scoring rule

Source volume does not equal fitness; accept each source only for claims its population and method support.

Output and template

Customer Evidence and Need Register.

Acceptance gate

Material blind spots and dependent sources are explicit before research questions are finalised.

Leadership checkpoint

Review whose absence from internal data could distort priority.

Human control point

The information authority approves source use and the decision owner accepts remaining blind spots.

Failure modes and safeguards

Failure modes are CRM truth, digital-only visibility and treating churned customers as unimportant. Safeguard with a source-coverage matrix.

Change, stop or escalation condition

Narrow claims when evidence cannot represent a material population.

Handoff to the next step

Transfer accepted sources and decision-critical gaps to Step 05.

Step 05: Design inclusive primary research

Purpose and decision served

Gather evidence about jobs, contexts, behaviour, barriers and value that internal records cannot establish.

Accountable owner

The research authority.

Delegated lead and contributors

Qualified research lead with recruiters, accessibility support, service staff, analysts and affected-person representatives.

Preconditions, inputs and access

Research questions, authorised participant groups, recruitment criteria, consent route and evidence gaps.

Actions

1. Convert internal assumptions into answerable research questions.
2. Recruit actual or likely users across relevant contexts, including underserved and assisted routes.
3. Choose methods according to the claim: observation, interview, diary, survey, journey or service evidence.
4. Pilot instruments for leading language, accessibility and burden.
5. Predefine analysis, limitations, withdrawal, retention and escalation routes.

Evidence to retain

Research plan, recruitment matrix, materials, consent or authority, field notes, analysis record and limitations.

Decision or scoring rule

Qualitative research can establish patterns and mechanisms in the observed context; it cannot establish prevalence without an appropriate design.

Output and template

Research-plan and participant-coverage rows in the Segment Decision Charter and Research Plan.

Acceptance gate

The research authority accepts method-to-question fit, inclusion and participant protections.

Leadership checkpoint

Approve exclusions that could materially affect service or priority.

Human control point

Participants can decline or withdraw through the approved route; the research authority can stop fieldwork.

Failure modes and safeguards

Failure modes are convenient-customer recruitment, preference polling and persona confirmation. Safeguard with contrary groups and question-method fit.

Change, stop or escalation condition

Change the design when emerging groups, barriers or harms invalidate recruitment assumptions.

Handoff to the next step

Transfer authorised research evidence and limitations to Step 06.

Step 06: Reconcile needs, behaviours, contexts and outcomes

Purpose and decision served

Build evidence-backed patterns without erasing meaningful variation or converting quotes into segment size.

Accountable owner

The research owner.

Delegated lead and contributors

Research analysts with service, sales, customer, data and independent challenge contributors.

Preconditions, inputs and access

Primary research, existing evidence, approved definitions and participant protection rules.

Actions

1. Code observed jobs, contexts, behaviours, barriers, workarounds, outcomes and contradictions.
2. Link every finding to source, participant or case scope and proof status.
3. Distinguish common pattern, material minority, extreme case and unresolved variation.
4. Compare stated need with observed behaviour and service evidence.
5. Record alternative explanations and gaps requiring validation.

Evidence to retain

Analysis framework, coded evidence, finding cards, contrary cases, proof status and reviewer challenge.

Decision or scoring rule

A finding is accepted only for its evidence boundary and cannot imply population prevalence without appropriate quantitative support.

Output and template

Finding rows in the Customer Evidence and Need Register.

Acceptance gate

Material findings are traceable, bounded and challenged by evidence from more than one relevant perspective where feasible.

Leadership checkpoint

Review findings that challenge current strategy, product or account classifications.

Human control point

The research authority approves interpretation and disclosure safeguards.

Failure modes and safeguards

Failure modes are quote theatre, majority bias and forced consensus. Safeguard with source traceability and material minority records.

Change, stop or escalation condition

Return to research when the decision depends on an unresolved group or mechanism.

Handoff to the next step

Transfer accepted patterns and limitations to Step 07.

Step 07: Form testable segment hypotheses

Purpose and decision served

Propose groups whose distinctions matter to the chartered action, rather than decorative personas.

Accountable owner

The segment decision owner.

Delegated lead and contributors

Segment lead with research, service, commercial, operations, data and inclusion contributors.

Preconditions, inputs and access

Accepted findings, evidence grammar, strategic context and candidate action differences.

Actions

1. State each hypothesis as distinguishable need, behaviour, context or service condition.
2. Explain why the distinction would change research, offer, service or retention action.
3. Define inclusion, exclusion and ambiguous cases without using stereotypes.
4. Record evidence, alternative boundary, expected service response and validation need.
5. Test whether combining or splitting changes the decision materially.

Evidence to retain

Hypothesis cards, boundary rules, ambiguous cases, action link, alternatives and proof status.

Decision or scoring rule

A segment hypothesis is eligible only if members share a decision-relevant condition and the boundary supports a different responsible action.

Output and template

Segment Hypothesis and Priority Ledger.

Acceptance gate

Every hypothesis has a usable boundary, evidence, action implication and documented ambiguity.

Leadership checkpoint

Reject labels that describe identity without explaining need, behaviour, context or service consequence.

Human control point

The decision owner and relevant inclusion authority approve hypotheses for comparison only.

Failure modes and safeguards

Failure modes are catchy names, demographic determinism and clusters with no action difference. Safeguard with boundary and decision tests.

Change, stop or escalation condition

Reject a hypothesis when its responsible use cannot be explained or safeguarded.

Handoff to the next step

Transfer eligible hypotheses to Step 08.

Step 08: Establish service and economic reality

Purpose and decision served

Understand the value exchange and operating commitment required for each hypothesis without reducing people to revenue.

Accountable owner

The economic buyer.

Delegated lead and contributors

Finance, service, sales, customer, operations and strategy owners with research challenge.

Preconditions, inputs and access

Eligible hypotheses, reconciled units, service evidence, commercial or funding records and obligation context.

Actions

1. Define value sought, value delivered, contribution or funding logic and service obligation.
2. Estimate only with accepted baselines, ranges, source lineage and uncertainty.
3. Include acquisition, fulfilment, support, retention, failure and capability burden where applicable.
4. Record strategic, public, social, regulated or inclusion value that money alone does not represent.
5. Separate historical performance from future opportunity and causal claim.

Evidence to retain

Value-exchange records, finance models, service conditions, assumptions, ranges and accepted limitations.

Decision or scoring rule

No segment is prioritised by revenue alone; use the chartered lenses and preserve obligations and uncertainty.

Output and template

Value and service rows in the Segment Hypothesis and Priority Ledger.

Acceptance gate

Finance and service owners accept what is known, modelled, excluded and still unvalidated.

Leadership checkpoint

Review who bears the cost and whether past under-service depresses apparent value.

Human control point

The authorised forum accepts non-financial obligations and economic limits.

Failure modes and safeguards

Failure modes are false lifetime value, cost-to-serve exclusion and historical self-fulfilment. Safeguard with ranges and service obligations.

Change, stop or escalation condition

Stop quantified comparisons when compatible definitions or finance authority are absent.

Handoff to the next step

Transfer bounded value-exchange evidence to Step 09.

Step 09: Test reachability, serviceability, inclusion and risk

Purpose and decision served

Determine whether the organisation can responsibly identify, reach, understand and serve each hypothesis.

Accountable owner

The customer executive.

Delegated lead and contributors

Marketing, sales, service, operations, accessibility, privacy, risk and data authorities.

Preconditions, inputs and access

Hypotheses, value exchange, channel evidence, service capabilities, attribute permissions and risk conditions.

Actions

1. Test reach through relevant channels without equating platform availability with permission or fit.
2. Test operational ability to deliver the differentiated promise and support route.
3. Identify exclusion, accessibility, proxy, discrimination, manipulation and dignity risks.
4. Record capability and evidence required before differentiated action.
5. Define prohibited uses, monitoring, complaint, override and stop routes.

Evidence to retain

Reach and service evidence, capability gaps, inclusion review, risk decisions and prohibited-use register.

Decision or scoring rule

A group cannot be priority now if the organisation cannot responsibly reach or serve it within accepted constraints; it may remain priority for research or capability.

Output and template

Reach, service and safeguard rows in the Segment Hypothesis and Priority Ledger.

Acceptance gate

Every eligible hypothesis has an accepted service route or named enabling gate and qualified safeguard decision.

Leadership checkpoint

Review whether targeting logic would create unequal or harmful treatment.

Human control point

Qualified authorities approve consequential attribute use and service differentiation.

Failure modes and safeguards

Failure modes are addressable-audience logic, digital exclusion and proxy discrimination. Safeguard with permitted-purpose and affected-person review.

Change, stop or escalation condition

Contain any harmful or unauthorised use; defer priority where service promises cannot be supported.

Handoff to the next step

Transfer eligible hypotheses and conditions to Step 10.

Step 10: Compare priorities without an automatic score

Purpose and decision served

Expose trade-offs across need, behaviour, value, service, inclusion, evidence and strategy.

Accountable owner

The authorised decision owner.

Delegated lead and contributors

Segment lead with economic, research, service, operations, customer and risk owners.

Preconditions, inputs and access

Eligible hypotheses, accepted evidence, value-exchange boundaries, safeguards and resource constraints.

Actions

1. Approve lenses, veto conditions and evidence threshold before comparison.
2. Compare hypotheses on separate evidence-linked rows.
3. Record interactions, overlaps, ambiguous members and competing resource needs.
4. Test combine, split, priority-for-research and differentiated-route alternatives.
5. Retain uncertainty, dissent and the consequence of not prioritising.

Evidence to retain

Comparison record, evidence links, veto decisions, overlap analysis, dissent and option notes.

Decision or scoring rule

A score may organise discussion but cannot offset a prohibited use, material exclusion, unserviceable promise or missing decision-critical evidence.

Output and template

Comparison rows in the Segment Hypothesis and Priority Ledger.

Acceptance gate

The forum can explain why each hypothesis is prioritised, researched, differentiated, retained, combined, split, deferred or stopped.

Leadership checkpoint

Review opportunity cost across segments, not only attractiveness within each one.

Human control point

The forum records the judgement; no model issues the disposition.

Failure modes and safeguards

Failure modes are weighted-score certainty, winner-takes-all allocation and hidden moral judgement. Safeguard with vetoes and reasons.

Change, stop or escalation condition

Reopen comparison if a material group, obligation, dependency or resource constraint was omitted.

Handoff to the next step

Transfer the recommended dispositions to Step 11.

Step 11: Make the executive segment-priority decision

Purpose and decision served

Authorise a bounded priority and the correct next action without turning evidence into permanent classification.

Accountable owner

The executive forum holding the resource decision.

Delegated lead and contributors

Decision owner with research, finance, service, customer, information and receiving owners.

Preconditions, inputs and access

Accepted decision pack, source limitations, dissent, safeguards and owned next gate.

Actions

1. Restate the chartered decision, resource boundary and expiry.
2. Review evidence, blind spots, overlaps, rejected alternatives and no-action consequence.
3. Choose priority now, priority for research, differentiated route, context, combine, split, defer or stop.
4. State authorised and prohibited uses of the segment boundary.
5. Assign owner, resources, learning gate, safeguards and review trigger.

Evidence to retain

Decision minutes, disposition, rationale, dissent, use boundary, resource decision and handover.

Decision or scoring rule

Each hypothesis receives one primary disposition for this decision; different uses require separate approval.

Output and template

Executive Segment Decision and Learning Register.

Acceptance gate

The decision has evidence limits, responsible-use rules, owner, resource boundary and receiving-owner acceptance.

Leadership checkpoint

Confirm priority does not silently mean exclusion or reduced service for others.

Human control point

Only the authorised forum allocates resources and accepts residual exposure.

Failure modes and safeguards

Failure modes are permanent persona publication, vague “focus” and implied targeting approval. Safeguard with use-specific authority and expiry.

Change, stop or escalation condition

Do not issue priority now when evidence, serviceability or safeguards are decision-critical gaps.

Handoff to the next step

Transfer the controlled decision and evidence pack to Step 12.

Step 12: Operate learning, change and retirement

Purpose and decision served

Test whether the priority improves the intended value exchange and remains responsible as markets and evidence change.

Accountable owner

The receiving customer or growth owner.

Delegated lead and contributors

Review owner with research, service, finance, data, accessibility and risk contributors.

Preconditions, inputs and access

Approved disposition, measure contracts, use boundary, review trigger and complaint or stop route.

Actions

1. Monitor research learning, service delivery, value exchange, inclusion and guardrails separately.
2. Reconcile new evidence against original definitions, populations and decision use.
3. Review drift in needs, behaviours, channels, economics, service capability and obligations.
4. Renew, change, split, combine, validate, defer, retire or stop the priority.
5. Close obsolete data uses, labels, access and downstream instructions.

Evidence to retain

Learning log, outcome evidence, complaints, drift review, changed decisions, retirement and access closure.

Decision or scoring rule

A priority remains active only while its evidence, use, service conditions, safeguards and owner remain valid.

Output and template

Review rows in the Executive Segment Decision and Learning Register.

Acceptance gate

The review owner confirms validity or issues a changed, retired or stopped disposition.

Leadership checkpoint

Review whether prioritisation improved value exchange or reinforced historical under-service.

Human control point

A named authority decides continuation after harm, complaint or material drift.

Failure modes and safeguards

Failure modes are stale segment labels, self-fulfilling allocation and model drift. Safeguard with expiry, affected-person evidence and closure.

Change, stop or escalation condition

Contain harmful use immediately and re-charter materially different market or customer decisions.

Handoff to the next step

Close the standard or transfer a newly approved decision to the relevant A1 operating guide.

4. The Evidence & Measurement Checklist

Preflight and authority

• One resource or operating decision explains why segment priority is needed now. Evidence: approved charter. Owner: executive sponsor.
• Market, population, offer, geography, relationship and period boundaries are explicit. Evidence: scope map. Owner: decision owner.
• Non-users, former users, intermediaries and assisted routes are considered where material. Evidence: population inventory. Owner: customer owner.
• Research, data, attribute, finance, service, inclusion and priority authorities are named. Evidence: authority matrix. Owner: executive sponsor.
• Existing segment commitments and supplier or performance conflicts are declared. Evidence: conflict register. Owner: governance owner.

Evidence access and quality

• Every source states its native unit, population, period, method, owner and permitted use. Evidence: source inventory. Owner: evidence authority.
• Channel, survivorship, non-response, missingness and historical-policy bias are visible. Evidence: source limitation rows. Owner: data owner.
• Research participants are actual or likely affected people recruited through an authorised route. Evidence: recruitment and participation record. Owner: research authority.
• Qualitative patterns are not presented as population prevalence without an appropriate design. Evidence: analysis boundary. Owner: research lead.
• Dependent datasets and repeated accounts are not treated as independent corroboration. Evidence: lineage and dependence record. Owner: evidence coordinator.

Method execution

• Person, organisation, account, user, beneficiary, payer and intermediary units are distinguished where relevant. Evidence: definition register. Owner: data owner.
• Jobs, needs, contexts, behaviours, barriers, outcomes and contradictions link to retained evidence. Evidence: Customer Evidence and Need Register. Owner: research owner.
• Each hypothesis explains a distinguishable condition and the responsible action it changes. Evidence: segment hypothesis card. Owner: segment lead.
• Inclusion, exclusion, ambiguous, overlap, combine and split rules are explicit. Evidence: boundary and overlap rows. Owner: segment lead.
• Historical behaviour, modelled value and future opportunity remain separate claims. Evidence: proof-status and value-exchange records. Owner: finance owner.

Governance, privacy, security and risk

• Sensitive, protected, proxy, inferred and third-party attributes have purpose-specific review. Evidence: attribute-use decision. Owner: information authority.
• Participant contact, consent or authority, withdrawal, retention and disposal routes are approved. Evidence: research governance pack. Owner: research authority.
• Accessibility and inclusion review covers disabled, assisted, offline and underserved people where relevant. Evidence: inclusion review. Owner: accessibility authority.
• Prohibited targeting, exclusion, manipulation and automated-decision uses are recorded. Evidence: prohibited-use register. Owner: relevant risk authority.
• Complaints, correction, override, containment and stop routes have named human owners. Evidence: safeguard and escalation record. Owner: customer executive.

Deliverables and acceptance

• Charter, authority, attribute, conflict and research-plan records are versioned and accepted. Evidence: Segment Decision Charter and Research Plan. Owner: executive sponsor.
• Sources, definitions, findings, contrary cases and blind spots are traceable. Evidence: Customer Evidence and Need Register. Owner: evidence authority.
• Every segment hypothesis has evidence, action implication, value exchange and service conditions. Evidence: Segment Hypothesis and Priority Ledger. Owner: segment lead.
• Comparison retains overlap, ambiguity, rejected alternatives, dissent and no-priority consequence. Evidence: comparison decision rows. Owner: decision owner.
• The final record names disposition, use boundary, resources, safeguards, owner and review. Evidence: Executive Segment Decision and Learning Register. Owner: forum chair.

Measurement

• Research coverage and decision-evidence fitness are measured separately from segment size. Evidence: measurement contracts. Owner: research owner.
• Value exchange includes customer or stakeholder outcome and organisational service commitment. Evidence: accepted value model. Owner: economic buyer.
• Baselines use compatible populations, periods and definitions or state baseline required. Evidence: baseline decision. Owner: metric owner.
• Thresholds for priority, research, differentiated route, defer and stop are approved before results. Evidence: dated gate record. Owner: decision owner.
• Inclusion and harm guardrails cannot be offset by commercial attractiveness. Evidence: veto conditions. Owner: relevant authority.

Decision and handover

• Each hypothesis receives one primary use-specific disposition. Evidence: executive decision row. Owner: authorised forum.
• Priority now states what attention or resource is authorised and what remains unapproved. Evidence: authority boundary. Owner: executive sponsor.
• Priority for research names the question, population, method, owner and evidence threshold. Evidence: research handoff. Owner: research authority.
• Rejected alternatives, dissent, blind spots and accepted uncertainty remain visible. Evidence: decision minutes. Owner: decision owner.
• Receiving owners accept definitions, safeguards, resources, measures and review triggers. Evidence: signed handover. Owner: receiving owner.

Review, change and stop controls

• Every priority has an expiry or material-change trigger and dated review. Evidence: lifecycle row. Owner: review owner.
• Need, behaviour, channel, economics, service, obligation and model drift trigger review. Evidence: change assessment. Owner: customer owner.
• New evidence is reconciled against original population, method, definition and decision use. Evidence: review workpaper. Owner: evidence owner.
• Harmful, unauthorised or discriminatory use triggers containment and human decision. Evidence: incident and stop record. Owner: relevant authority.
• Retired labels, attributes, data uses, audiences and instructions have closure actions. Evidence: retirement register. Owner: review owner.

Topic-specific measurement contract

Measure A: Decision-critical customer evidence coverage

Measure B: Material participant and context coverage

Measure C: Priority-to-action integrity

Measure D: Named customer or stakeholder outcome

Measure E: Value-exchange sustainability

Measure F: Material segment-use guardrail breaches

Evidence classifications used in this guide

• VERIFIED: accepted evidence for the stated population, period, method and decision use.
• INFERRED: a reasoned interpretation with alternatives and limitations visible.
• PROPOSED: a hypothesis, boundary, priority, measure or action awaiting organisational acceptance.
• REQUIRES VALIDATION: decision-critical evidence is missing, conflicting, stale or outside permitted use.

Primary and supporting sources

• GOV.UK Service Manual: User research for government services: VERIFIED public-service research guidance at the 25 August 2026 source cut-off. It supports end-to-end, inclusive research focused on what people do and need rather than popularity. Application to an organisation's segment decision remains PROPOSED.
• GOV.UK Service Manual: Finding participants for user research: VERIFIED guidance at source cut-off. It supports defining target groups, recruiting actual or likely users, including disabled people and avoiding recruitment bias. It does not prescribe a universal sample for this guide.
• GOV.UK Service Manual: User research in discovery: VERIFIED guidance at source cut-off. It supports understanding likely users, current behaviour, barriers, needs and the whole service before design. It does not establish market size or commercial priority.
• ICC/ESOMAR International Code 2025: VERIFIED international research and data-analytics code at source cut-off. It informs transparency, responsibility and confidence boundaries; organisational compliance requires qualified review.
• UK ICO: What is automated individual decision-making and profiling?: VERIFIED only for terminology and context at the 25 August 2026 source cut-off. The former URL redirects to this current route, and ICO guidance is subject to review following changes in UK law. It is not a current-law conclusion; applicable legal duties remain REQUIRES VALIDATION through jurisdiction-specific qualified authority at use and publication.