Test Research Handoff State Transitions

A handoff label is useful only if the events that change it have defined meanings. A viewed-document notification, an expired timer and a recipient's acceptance are different inputs. Test the next state produced by each input before allowing that state to release downstream work.

Use a small transition fixture with explicit owners and blocked effects. The exercise below follows one fictional assignment through valid and invalid events. Its output is a completed test record, not a general packet template or a claim that an existing CRM implements these controls.

Fix the assignment and two independent state dimensions

Researcher Leena owns assignment H44 for account L44. Data steward Omar asks whether site S9 lies inside the service boundary and whether L44 is its operating entity. The agreed completion rule permits a bounded unresolved answer after the named official sources and authorized internal records are checked. It does not authorize outreach, merging entities or identifying a buying group.

Packet P1 supports the site's location but leaves the operator unresolved. Content state describes draft, submitted, correction requested or accepted-unresolved. Transfer state separately describes not transferred, destination held or verified at the intended account. An accepted packet may therefore remain unusable in the destination system.

The W3C Data Quality Vocabulary note provides a framework for describing quality so users can judge fitness for purpose. It does not prescribe these states. Here, acceptance means the specified evidence product is adequate for an internal coverage review, not that every claim is confirmed.

Run the initial acceptance tests

• T1, notification input: P1 is submitted to Omar, and the viewer reports that he opened it. Expected state: content remains submitted; transfer remains not transferred. Disallowed transition: accepted-unresolved. Leena still owns the unaccepted packet. Blocked effect: no task may describe the operator as verified or initiate contact because of the viewing event.

• T2, timer input: the review checkpoint expires without a decision. Expected state: submitted with an overdue-review flag, routed to the designated cover steward. Disallowed transition: automatic acceptance or automatic rejection. Leena retains packet responsibility; the team lead owns reviewer routing. Blocked effect: no downstream release is generated merely to clear an aging queue.

• T3, explicit acceptance input: Omar records that P1 satisfies the bounded question, including its unresolved operator conclusion. Expected state: accepted-unresolved for P1 and not transferred. Disallowed transition: operator confirmed. Omar owns any decision to commission further research. The known location may inform the agreed internal review, but entity merging and operator-directed outreach remain blocked.

These are synthetic test outcomes. The same unchanged packet receives different treatment because the input authority and meaning differ. A system that converts all three inputs into one green completion label fails the fixture.

Test new scope and a corrected packet

T4 arrives after acceptance: a representative asks for the site's buying group. Expected result: H44 retains its accepted-unresolved history; a separate proposed assignment H45 is created for scoping by the requester. It is not accepted into research automatically. The disallowed transition is reopening H44 as defective simply because someone wants additional work. No buying-group task or external communication is released until the new assignment is appropriately agreed.

T5 is a genuine correction. Omar identifies a copied site reference in P1's evidence label. Leena owns the correction request and produces P2 with S9 consistently identified. The operator conclusion remains unresolved. Omar explicitly accepts P2, which supersedes P1 for current use while preserving P1 and the correction reason.

The content transition is correction requested to accepted-unresolved for P2. However, the transfer preview targets account L45 rather than the approved L44. Expected combined state: accepted content, destination held. The integration owner must resolve the mismatch. Disallowed transition: usable-in-CRM, even though the text is accepted. Blocked effect: writing P2's evidence into L45 or treating a successful upload as a successful handoff.

T6 supplies the corrected destination reference. In the mock destination, the integration owner verifies account L44, packet revision P2 and the unresolved operator label, then records the check. Expected state: accepted-unresolved plus verified transfer. Omar owns limited internal use; Leena no longer owns an unaccepted deliverable. Operator confirmation, account merging and outreach still do not follow from this transition.

Reopen without rewriting the earlier result

T7 introduces a new authoritative statement naming the site operator. It is relevant to H44's unresolved claim, unlike T4's buying-group expansion. Expected result: a linked reopening record H44-R1 enters evidence review, with Leena assigned by the steward. P2 remains the historical accepted-unresolved result at its recorded cutoff.

The new statement has not yet passed entity and effective-date checks. The disallowed transition is changing P2 retrospectively to operator-confirmed. The blocked effect is propagation of the new name as a verified current operator before that review. Omar owns the affected internal decision and marks its operator-dependent use pending.

Keep the fixture results with the event reference, prior state, allowed next state, rejected transition, accountable owner and blocked effect. The test passes only when all seven inputs preserve their distinct consequences. Closing an assignment, accepting corrected content, verifying a destination and reopening an evidence question must remain separately understandable.

Continue with Create a Useful Sales Research Service-Level Agreement or Move Researched Leads into a CRM with a Clear Handoff.