Define research watch states without turning them into ratings
A research watch state should describe the next work needed, not an investment opinion. “Waiting for source” and “ready for review” are operational labels. They do not mean avoid or buy. Define each state with entry evidence, a next action, and an exit condition.
W3C provenance supplies background for linking records to activities. SEC search is one example of a source route used during a state transition. The state table below is a proposed manual system, not a representation of current product features or an investment-rating framework.
Choose states that change behavior
Begin with a small set: collecting sources, checking evidence, waiting for a defined input, ready for review, and archived. A state is useful only if it tells someone what to do next. If “interesting” and “watching” require the same undefined action, they add little operational clarity.
Store the reason beside the state. “Waiting” without a named missing input can hide an abandoned task. “Waiting for the agreement exhibit to become accessible” identifies a specific barrier and allows another researcher to help. Add a review date or observable trigger where appropriate, without treating that date as evidence.
Hypothetical state table
Assume fictional Westfield Pumps is being added to a source archive. This example is hypothetical.
| State | Entry condition | Next action | Exit condition |
|---|---|---|---|
| Collecting | Identity established, key reports missing | Locate defined reports | Sources inventoried |
| Checking | Sources available, passages uninspected | Read and log selected evidence | Material checks recorded |
| Waiting | Named external input unavailable | Check trigger or retrieval route | Input found or task retired |
| Ready for review | Evidence packet and limits prepared | Reviewer replays selected claims | Review outcome recorded |
| Archived | No current research task | Preserve retrieval path | New authorized task reopens |
An issuer in “ready for review” is not necessarily attractive. It simply has a research packet ready for inspection. A company in “waiting” may have an important unanswered question, but the label does not determine the answer.
Make transitions accountable
Record the prior state, new state, date, reason, and supporting source or action. Do not update the state merely because someone opened the dossier. A transition should correspond to work performed or a changed barrier. If a key source becomes invalid, returning from ready-for-review to checking is sensible and should remain visible.
Separate task state from document state. A dossier can be waiting for one agreement while several other documents are fully inspected. Source-level statuses preserve detail; the dossier state summarizes the next blocking action. Forcing every file to share the same status loses useful information.
Verify with a handoff test
Give a reader the state, reason, and next-action fields. Can they take a concrete step without asking what you meant? If not, revise the definition. Then test an edge case: a new amendment arrives after the packet was marked ready. The state system should allow rechecking rather than implying permanent completion.
Keep archived records searchable. Archiving ends active work; it should not erase sources or the reasons a task ended. A later reopening should retain the historical transition and create a new action context.
Limits and common mistakes
States compress complexity and cannot replace the underlying evidence log. A single label cannot describe every uncertainty. Keep specific limits attached to claims and sources. Nor do operational states establish confidence, expected returns, or portfolio priority.
Common mistakes include using green and red labels that imply recommendations, leaving waiting without a trigger, and marking a task complete when only files were downloaded. Another is treating archived as deleted. Define state meaning in plain language and ensure it describes actual research work.
A final record check
Audit dormant waiting items separately. Ask whether the missing input is still relevant, whether the source route remains usable, and whether the next action has an owner. A dormant item can be retired with a reason. Leaving it labeled active indefinitely creates an illusion of ongoing monitoring, especially in a manual workflow where nobody is actually performing the checks.
FAQ
Should confidence be a watch state?
Prefer a separate field with a clearly defined purpose. Workflow progress and strength of a conclusion answer different questions.
Can a task move backward?
Yes. New evidence, access failures, or source corrections can require more checking. Record the reason so the transition is informative rather than appearing accidental.