How to Turn an Account Universe into an ICP-Qualified Account List

Who this guide is for

This guide is for account-research, revenue-operations, and GTM teams that already have a defined account universe and need to decide which records pass an ICP. If the starting population has not yet been constructed, deduplicated, and bounded, begin with How to Build a Target-Account Universe from Public and Commercial Data.

Contents

1 Freeze the starting universe 2 Define the three output states 3 Apply the five qualification gates 4 Set evidence and unknown rules 5 Preserve traceability and reason codes 6 Control refreshes and list changes 7 Use the qualification ledger 8 Illustrative example 9 Common mistakes 10 What this does not mean 11 Common questions 12 Sources and methodology

1. Freeze the starting universe

Do not rebuild the market universe inside the qualification task. Freeze a version with:

• universe ID and extraction date;

• source systems and permitted use;

• unit: firm, entity, establishment, facility, or another explicit record;

• geography and industry scope;

• parent and site deduplication rules;

• coverage exclusions and suppressed records;

• original record count;

• owner and change log. This protects the denominator. If researchers add attractive accounts during screening without versioning the universe, pass rates and exclusions can no longer be reproduced. Official datasets use specific units. County Business Patterns publishes annual statistics for employer establishments by geography and industry and describes measures such as establishments, employment, and payroll. U.S. Census Bureau - County Business Patterns API The BLS QCEW concepts distinguish an establishment from a firm and define establishment-level employment and industry concepts. BLS - QCEW Concepts If the universe originated from establishments, preserve that fact. Parent resolution may combine records for account governance, but it does not retroactively turn the source measure into firm counts. Commercial data can add names, domains, roles, and relationships. Store vendor, license, field definition, retrieval date, and confidence. A populated field is not automatically verified.

2. Define the three output states

Every source record must end in one mutually exclusive list state for the current ICP version: Output state Meaning Permitted next step ICP-qualified All critical gates pass with sufficient evidence Enter governed prioritization and buying-group research Output state Meaning Permitted next step Research At least one critical field is unknown or contradicted, with no hard-gate confirmed_non_fit Complete a named evidence or source-reconciliation task; no automatic qualification Excluded At least one hard gate is confirmed_non_fit or an exclusion rule applies Suppress for this ICP version; preserve reason and review condition Do not create a hidden fourth state called “probably good enough.” An indirect clue can be stored as a confidence or research note, but a critical dependency remains in Research until acceptable evidence establishes its state. An output state is ICP-specific. An account excluded from an OEM ICP may still fit an end-user ICP. Store the ICP ID and version with the result.

3. Apply the five qualification gates

Use a sequential flow so each denominator is visible. Gate 1: account structure Confirm that the record represents the right entity, establishment, facility, division, or project. Resolve parent relationships, inactive units, duplicates, and sites outside scope. Do not qualify a parent on evidence that belongs to one unidentified subsidiary. Gate 2: application fit Confirm the product function, operating process, conditions, installed alternative, and consequence. Industry relevance is not enough. The exact offer must plausibly fit the named application. Gate 3: commercial relationship Confirm whether the account integrates, resells, operates, specifies, installs, or otherwise participates through the intended relationship. A manufacturer is not automatically an OEM, and a wholesaler is not automatically a relevant distributor. Gate 4: qualification and serviceability Check mandatory product, vendor, regulatory, delivery, warranty, language, inventory, installation, and support requirements. Any critical condition that cannot be met is a failure, regardless of company attractiveness. Gate 5: evidence sufficiency Verify that critical claims have current, source-matched evidence. First-party public-company filings available through EDGAR may support reported operations, facilities, products, or organizational structure. U.S. Securities and Exchange Commission - Search Filings SAM.gov opportunity notices may support project-specific roles and requirements. SAM.gov - Contract Opportunities Neither proves buying intent. Only records passing all five gates enter the ICP-qualified list.

4. Set evidence and unknown rules

Create a field-level evidence contract: Evidence state Definition Gate treatment confirmed_fit Accepted, current, scope-matched evidence satisfies the criterion May pass if all hard requirements are met confirmed_non_fit Accepted evidence fails the criterion Fail the applicable gate and route to Excluded where the criterion is a hard gate contradicted Accepted sources materially conflict Route to Research and reconcile scope, date, identity, or source precedence unknown The required claim has not been established Route to Research for critical fields not_applicable The criterion does not apply by design and a reason is recorded Remove from eligible denominators; never use as a substitute for unknown Define source precedence. A current facility document should normally outweigh an undated aggregator description for that facility. A parent filing should not override direct evidence that a particular site performs a different activity. For every unknown or contradicted state, write one next-evidence or reconciliation action:

• identify the operating site;

• find the relevant process or equipment statement;

• verify whether the account resells or operates the product category;

• confirm a qualification requirement;

• cost a service dependency;

• resolve channel ownership. Research queues without specific questions become permanent ambiguity.

5. Preserve traceability and reason codes

Each decision should retain:

• source record ID and universe version;

• normalized account and site identifiers;

• ICP ID and version;

• gate result and reason code;

• evidence URL or document reference;

• observation and access date;

• researcher and review owner;

• next action or exclusion review condition;

• state-change timestamp. Use controlled reason codes, such as:

• STRUCTURE_WRONG_UNIT

• APPLICATION_CONFIRMED_NON_FIT

• EVIDENCE_CONTRADICTED

• RELATIONSHIP_NOT_SUPPORTED

• QUALIFICATION_FAIL

• SERVICEABILITY_UNKNOWN

• CHANNEL_CONFLICT

• EVIDENCE_STALE Reason codes support counts, but the underlying evidence remains necessary. Do not let the code become an unsupported conclusion. Report a reconciliation: Starting universe = ICP-qualified + Research + Excluded If the equation fails, records were duplicated, dropped, or assigned overlapping final states.

6. Control refreshes and list changes

An ICP-qualified list is versioned, not permanent. Refresh when:

• the source universe changes materially;

• a company is acquired, closed, renamed, or reorganized;

• a facility opens, closes, or changes production;

• the application or product scope changes;

• a certification, service, or channel condition changes;

• evidence passes its freshness rule;

• the ICP itself is revised. Do not overwrite old states. Append a new decision version so the team can explain why an account entered or left the list. Use incremental refreshes between full reviews. Recheck high-impact fields and explicit triggers rather than re-researching every descriptive attribute. If a record changes from Research to ICP-qualified, preserve the evidence that resolved the unknown. Qualified-list changes should not silently trigger contact actions. Account prioritization, buying-group mapping, suppression, and campaign governance remain separate stages.

7. Use the qualification ledger

Field Required value Universe ID/version Frozen source population Source record/account unit Original and normalized identity ICP ID/version Qualification definition applied Structure gate Pass / Hold / Fail + evidence state and reason Application gate Pass / Hold / Fail + evidence state and evidence Relationship gate Pass / Hold / Fail + evidence state and evidence Qualification/service gate Pass / Hold / Fail + evidence state and dependency Evidence-sufficiency gate Pass / Research required / Fail + evidence state Final state ICP-qualified / Research / Excluded Decision date and owner Audit fields Next review Trigger, date, or not applicable The ledger should support both record-level review and aggregate gate reconciliation.

8. Illustrative example

Illustrative example - explicitly fictional. ForgeLine Components begins with a frozen, fictional universe of 320 facility-level records. Each record exits at its first unresolved or failed critical gate so final states remain mutually exclusive. Gate Entered Passed Research Excluded Account structure 320 276 24 20 Application 276 118 96 62 Relationship 118 84 18 16 Qualification/serviceabil ity 84 61 12 11 Evidence sufficiency 61 45 16 0 Final reconciliation:

• ICP-qualified: 45

• Research: 24 + 96 + 18 + 12 + 16 = 166

• Excluded: 20 + 62 + 16 + 11 = 109

• Check: 45 + 166 + 109 = 320 End-to-end qualified share is 45 / 320 = 14.1%. The application-pass rate is 118 / 276 = 42.8%, using only records that passed structure. These calculations describe the fictional reviewed universe; they do not estimate a market, predict response, or represent qualified leads. The Research list is not blended into the score. ForgeLine assigns each of its 166 records a specific evidence action and prioritizes only those whose unresolved field could change the decision materially.

9. Common mistakes

• Treating the full account universe as outreach-ready.

• Rebuilding or expanding the denominator during qualification without a new version.

• Counting establishments as firms or parents.

• Passing unknown critical fields because the account looks attractive.

• Using one broad industry code as application evidence.

• Deleting excluded records and losing suppression or reason history.

• Calling ICP-qualified accounts qualified leads.

• Allowing list refreshes to trigger ungoverned outreach automatically.

10. What this does not mean

This guide does not explain how to construct the original target-account universe; that is a prior Market Intelligence decision. It does not say every record requires exhaustive research. Evidence effort should be proportional to the criterion’s importance and the reversibility of the next action. An ICP-qualified account has passed a defined account-fit process. It is not proof of an individual contact, buying committee, signal, intent, response, meeting, opportunity, revenue, or customer outcome. Applicable contact, privacy, channel, and suppression rules still apply.

11. Common questions

Can indirect evidence ever establish confirmed fit? Only when the field contract accepts that source type and the evidence is current, scope-matched, and sufficient for the criterion. A critical product, qualification, or service dependency remains Research while its state is unknown or contradicted. Should research records be scored? They may be ordered by evidence value or research priority, but do not combine an incomplete fit score with qualified accounts. Show missing critical fields explicitly. What if the universe contains parents but the ICP needs facilities? Create a versioned structural enrichment step that expands parents into relevant units. Preserve the original parent universe and do not mix parent and facility counts. When can an excluded account return? When its exclusion review condition occurs or the ICP changes. Create a new decision record; do not erase the earlier exclusion. How is this different from lead qualification? This process qualifies account fit before contact and engagement. Lead qualification evaluates a defined lead event and criteria under a separate process.

12. Sources and methodology

The output states, gates, evidence contract, reason codes, ledger, and reconciliation method are SellersGPT editorial tools. Official sources support their stated units and public-document functions. No public statistic or document is treated as purchase intent. 1 ICP07-CENSUS-CBP - U.S. Census Bureau, “County Business Patterns APIs.” https://www.census.gov/data/developers/data-sets/cbp-zbp/cbp-api.html Accessed August 24, 2026. 2 ICP07-BLS-QCEW - U.S. Bureau of Labor Statistics, “Quarterly Census of Employment and Wages: Concepts.” https://www.bls.gov/opub/hom/cew/concepts.htm Accessed August 24, 2026. 3 ICP07-SEC-EDGAR - U.S. Securities and Exchange Commission, “Search Filings.” https://www.sec.gov/search-filings Accessed August 24, 2026. 4 ICP07-SAM-OPPORTUNITIES - SAM.gov, “Contract Opportunities.” https://sam.gov/opportunities Accessed August 24, 2026.