Analytics choice

SyncAI uses optional analytics and advertising measurement only after you allow it. Necessary site functions work without these trackers. Privacy details

Back to Insights
Decision Case

Proxy Is Not Outcome

Orville Davis·Author

Proxy is not outcome. A KPI, a leading indicator, a model score, a green tile, a closed work-order count, an MTBF estimate, a sensor health %, or any other stand-in measure can inform a recommendation. It is not the verified operational outcome on the Decision Case. Treating a proxy as the outcome invents false closure, skips verification, and teaches the wrong lesson into Learning. Sync may surface proxies. A named human still decides. Verification still records the actual check.

Sync recommends. A named human decides. On the Decision Case, a proxy is a stand-in measure. It is not the verified operational outcome. A proxy can inform. It does not authorize, clear a gate, or replace a verified outcome.

The KPI moved. The leading indicator moved with it. The model score is high. The tile is green. The closed work-order count is up. The MTBF estimate looks better than last quarter. Sensor health % is in the band the meeting likes. Someone says the outcome is in, and the room treats the sentence as the result. Nobody wrote the check that the authorized action did what it was supposed to do. Nobody stored the measurement the question actually named. The Decision Case still has to show the inspection, the reading, or the verification that was actually recorded. A stand-in that moved is not that record.

This essay follows Trend Is Not Truth and stands beside Correlation Is Not Causation and Confidence Is Not Evidence. Trend keeps a rising or falling line, a co-moving tag cluster, a week-over-week delta, and a trend narrative from being stored as a recorded causal check or a verified outcome. Correlation keeps co-moving tags, coincident alerts, and dashboard trends from being stored as causation, and it keeps them from authorizing action. Confidence keeps a strong gut feel, a high model score, years of tribal certainty, or a manager's certainty from being stored as a recorded check. This essay is the stand-in itself: the number the room uses when it wants the case to look closed. A proxy can inform a recommendation. It does not authorize, clear a gate, or replace a verified outcome. Being sure the proxy will keep moving does not record the check.

Facts come before assumptions. The same gap has neighboring forms. Assumption Is Not Evidence keeps a belief, a habit, a prior shift's story, or an unverified model output from being stored as a recorded fact. "The proxy moved, so the outcome is in" is that story when the story is a stand-in. Unknown Is Not Clear records missing evidence as unknown, not as clear or ready. Blank Is Not Zero keeps an empty field from being stored as a measured zero. Silence Is Not Clearance keeps no reply, no alert, and an empty inbox from being read as approval. A proxy is what the room uses to skip those records. A green tile does not record the missing evidence as clear. A count drawn across a blank does not measure none. A narrative about the KPI does not write the decision the operator never recorded.

Orville Davis states the Decision Case order in Field Manual v0: Question, Evidence, Recommendation, Human decision, Action, Verification, and Learning. The manuals index lives at /manuals. Evidence is step 02. The Evidence chapter keeps a visible split between observed fact, assumption, and what is still missing. A proxy is not a fourth column that overrides the split. A proposal whose only basis is that a stand-in moved has no outcome a later reader can inspect as verified. A proposal with no verified outcome is not ready to be treated as closed. Honesty boundaries sit beside the spine: recommend is not authorize, evidence is required, and Stage-1 does not treat a KPI, a leading indicator, a model score, a green tile, a closed work-order count, an MTBF estimate, or a sensor health % as the live result.

A proxy is not a verified outcome

A verified operational outcome is something a later reader can inspect on the case: the condition that was present, the change a named person authorized, the measurement taken against that change, and the verification that the authorized action did what it was supposed to do. A proxy is a stand-in. It names a number that is easier to display than the outcome the question asked for. It does not name whether that outcome occurred. Those are different records. The tile does not write the check.

The meeting can note the proxy. Noting it records that someone saw the stand-in. It does not record the outcome, and it does not record a verified check. The outcome stays in the verification column, or it stays missing. The proxy stays in the assumption column, or it stays off the case as a display. A cleaner number is still not a verified outcome. Treating the proxy as the outcome invents false closure. The case looks finished. The check was never written. That skip is the failure this essay names.

Sync may surface proxies. Holding a proxy beside a recommendation can inform that recommendation. Informing a recommendation does not authorize the work. The recommendation is a proposal. The person who can accept the operational consequence records accept, reject, escalate, or return. If the only support is that a stand-in moved, the proposal is not a closed case. The system does not auto-accept because the proxy agrees with the draft. Verification still records the actual check.

A KPI and a leading indicator are not the outcome

A KPI is a chosen number: a count, a rate, a ratio, a score the plant already reports. A leading indicator is a number chosen because it is expected to move before the outcome the question is about. Both can inform a recommendation. Neither is the verified operational outcome. The KPI can rise while the condition the question named is unchanged. The leading indicator can fall because the window changed, the definition changed, or a third condition moved it. The operator reads the number as a report about the stand-in. The operator does not read it as the result.

A recommendation drafted from the KPI is still a recommendation. Sync recommends. Humans decide. The named person accepts, rejects, escalates, or returns against evidence a later reader can inspect. If the only support is that the indicator moved the way the room hoped, the proposal is not ready to be stored as closed. A hoped-for proxy move is not the check.

This essay states no OEM limit and no operating threshold. A KPI that crosses a mark the meeting invented is not a verified outcome, and this page does not supply the mark. The case holds the measurement that was stored, or it holds unknown. It does not hold a limit this essay made up so the tile would look decisive.

A model score and a green tile are not the check

A model score is a number a model produced. A high score can inform a recommendation. It is not a recorded check, and it is not a verified outcome. Confidence already keeps that score out of the observed-fact column. Calling the same score a proxy does not put it back. The score can be steep. The room can be sure. Neither the score nor the surety accepts the operational consequence, and neither one writes the verification.

A green tile is a display. Dashboard Is Not Decision keeps that display off the decision. Color does not put it back as an outcome. Green is not a verified outcome of improvement. Red is not a verified outcome of failure. Amber is not a partial check. A green tile, an improved trend, or a cleared alert tile must not be read as authorization, verification of outcome, work-order execution, or plant control.

An alert that fired while the tile was green is still an alert. Alert Is Not Decision keeps seeing, acknowledging, or silencing that alert off the decision. A proxy beside the alert does not put it back. Clearing the queue because the score "looks healthy" must not be read as authorization, verification of outcome, work-order execution, or plant control. Sync may emit alerts, coverage gaps, and recommendations. Sync may surface proxies. A named human still opens a Decision Case on a question, cites approved evidence, and accepts, rejects, escalates, or returns.

A closed work-order count is not the result

A closed work-order count is a count of records marked closed in the system that holds work. The count can inform a recommendation. It is not the verified operational outcome on the Decision Case. A record can be closed because the clock ran out, because the status was changed, or because the queue was cleared. Closure of a work record is not the check that the authorized action did what it was supposed to do. More closed records are not, by themselves, a better plant.

Sync does not write the work order. An ACTION disposition on the Decision Case is work intent. It is not plant execution. The systems that already hold execution authority write the work order or the isolation. A count of those records, however it is surfaced, stays a count. It does not move execution onto Sync, and it does not replace verification. The case stays open until the check is on the case, even when the closed work-order count has already moved.

An MTBF estimate and a sensor health % are not the plant check

An MTBF estimate is an estimate of mean time between failures. It is a calculation on a history, a window, and a definition of failure. The estimate can inform a recommendation. It is not the verified operational outcome of the action this case authorized. A better estimate can appear because the window dropped an old failure, because the failure definition changed, or because the history is incomplete. The estimate does not record what this authorized change did.

A sensor health % is a coverage display: how much of a watched set is reporting, or how a health figure was scored. Sensors, rows, and patrol routes are coverage. They are not control of the plant, and a percentage of that coverage is not the outcome the question asked for. A high sensor health % does not mean the failure mode was removed. A low one does not, by itself, authorize the work. Stage-1 evidence is the record held on the case. A live connector that pulls historian or control-system tags sits outside this edition. Simulated or seeded telemetry is a practice record. A practice health percentage is still a practice number. It is not a live plant result, and it is not a substitute for the verification the operator still has to write.

If the case cannot show the asset configuration, the failure and maintenance history, and the human judgment that produced the estimate or the percentage, the figure is a number without a trail. Evidence Lineage Is Not Optional already requires that trail. A proxy does not shorten it. The figure can sit on the case as a display, labeled as a display. The label does not promote the stand-in into a verified outcome. Honesty Boundary Is Not Optional says Sync states what is simulated, seeded, or out of scope. Naming that limit does not promote the proxy into the result. The Honesty boundaries chapter keeps Stage-1 from claiming a live result the edition does not hold.

A proxy does not clear a gate

A gate on the Decision Case is a recorded step that is either satisfied or not: evidence a later reader can inspect, a named human decision, a verification written down. A KPI does not satisfy the gate. A leading indicator does not open it. A model score does not stand in for the missing record. A green tile does not close it. A closed work-order count does not initial it. An MTBF estimate does not replace it. A sensor health % does not clear it. A proxy can inform. It does not authorize, clear a gate, or replace a verified outcome.

If the gate is evidence, the case shows the observed fact, the labeled assumption, and what is still missing. A proxy in the assumption column does not fill the observed-fact column. If the gate is the human decision, the case shows accept, reject, escalate, or return by a named person. The tile does not sign that act. If the gate is verification, the case shows the check against the authorized action. A later movement in the proxy is not that check. Skipping the check because the stand-in already moved is false closure.

Unknown, blank, and silence stay what they are

Unknown stays unknown. Filling the unknown line with a KPI, a model score, or a sensor health % does not change the state to clear. Bad or not-ready evidence stays blocked. A proxy that looks related to the blocked item does not relabel blocked as ready. The operator must not treat unknown as go because the stand-in has moved. Go would mean the evidence is sufficient for the next recorded step. A proxy is not sufficiency.

Blank stays blank. Zero is a verified measurement of none. A proxy written into an empty field collects a number. It does not collect the value. Recovery figures that were left blank stay blank. They are not inferred from a KPI that looked better. A blank verification field is not a verified outcome of none, and a closed work-order count does not close it.

Silence stays silence. No reply is not approval. An empty inbox is not clearance. A quiet dashboard is not controlled. A proxy on a quiet screen is still a proxy. It does not write accept, reject, escalate, or return. It does not record a check that nobody performed. A narrative about the quiet green tile does not turn the missing response into the named decision.

Sync recommends. Humans decide.

Authorization is the named act, and the act still needs a basis the case can show. A drafted next action is a proposal. Recommend Is Not Authorize keeps that proposal off the decision. A proposal that cites a proxy is still a proposal. The KPI can be favorable. The model score can be high. Neither the number nor the score accepts the operational consequence. Human Decision Is Not Optional says a named person accepts, rejects, escalates, or returns. Until that act is recorded, the case is still a proposal. A proxy does not authorize action. The system does not auto-accept because the stand-in agrees. Sync recommends. A named human decides.

Learning is what a later question is allowed to inherit. The closed case shows what was believed, what was authorized, and what was checked. Verification Is Not Optional says authorization records who decided, and the case stays open until the check is written down. Verification still records the actual check. A proxy move is not that check. Learning Requires a Verified Outcome says the later question inherits that verified outcome, not a hope. Learning requires a verified outcome, not a hoped-for proxy move. What was believed may include a labeled note that a KPI, a score, or a count moved. Learning does not inherit the proxy as if the check had confirmed the outcome. A later crew that repeats the work because the last stand-in "already moved," without the verification on the prior case, is inheriting a display. That is the wrong lesson. The Learning chapter publishes the closed record. It does not publish a proxy as a result.

A KPI, a leading indicator, a model score, a green tile, a closed work-order count, an MTBF estimate, a sensor health %, and any other stand-in measure are not the verified operational outcome. They can inform a recommendation. They do not authorize, clear a gate, or replace a verified outcome. Treating a proxy as the outcome invents false closure, skips verification, and teaches the wrong lesson into Learning. Sync recommends. A named human decides. Verification still records the actual check.

The neighboring boundaries still hold

This boundary is not the neighboring ones. Each of those essays still holds. None of them promotes a proxy to a verified outcome.

Trend Is Not Truth is the direction, the delta, and the narrative that the direction is already the result. Correlation Is Not Causation is the chart of co-movement. Confidence Is Not Evidence is the surety attached to a statement, including a high model score. Proxy Is Not Outcome is the stand-in the room stores in place of the check: the KPI, the leading indicator, the score, the green tile, the closed work-order count, the MTBF estimate, the sensor health %. A slope does not become an outcome because someone calls it a KPI. Tags that travel together do not become the result because a count moved with them. Being sure does not close the case. The label stays. The verified-outcome column stays empty until the check is on the case.

Unknown, blank, and silence are still the missing record. A proxy does not interpret them into clearance. A case that is unknown, blank, or quiet is not made ready by a stand-in that has moved. A moving stand-in is not a measurement.

The check that closes the case is still verification. The lesson a later case may inherit is still the verified outcome, not a hoped-for proxy move. The trail behind a recommendation is still evidence lineage. An alert is still not a decision. A dashboard is still not a decision. Those records are either on the case or they are not. If they are not, a proxy does not supply them.

Where the public statement lives

Field Manual v0 is the public contents of the Decision Case. Start at the manuals index or open Sync Field Manual directly. Evidence is step 02. Observed fact, assumption, and what is still missing stay distinct. A proxy does not collapse that split. A proposal with no basis is not ready for the human decision that follows. Verification is step 06. A stand-in that moved after the work is not the check. Learning is step 07. A later case inherits the verified outcome, not the proxy the last shift treated as the result. Honesty boundaries keep the edition from treating a KPI, a leading indicator, a model score, a green tile, a closed work-order count, an MTBF estimate, or a sensor health % as a live fact. Later editions can deepen a chapter. The spine stays in this order.

Decision Case spine

  1. 01Question
  2. 02Evidence
  3. 03Recommendation
  4. 04Human decision
  5. 05Action
  6. 06Verification
  7. 07Learning

The standing rule sits beside the spine: Honesty boundaries.

What this article is not claiming

This is an essay about the record, not a customer case study. It names no plant, states no savings figure, and claims no prevented failure. It states no OEM limit and no operating threshold. It does not claim that Sync executes plant work, writes a work order, isolates equipment, or controls a plant. It does not claim that a proxy is the outcome, that a KPI is a verified operational outcome, that a leading indicator is the check, that a model score closes the case, that a green tile authorizes action, that a closed work-order count is the result, that an MTBF estimate replaces verification, or that a sensor health % clears a gate or replaces a named human decision. It does not claim that self-guided onboarding is a live product path.

Stage-1 readiness means a signed-in user can complete the Decision Case — question, evidence, recommendation, human decision, action, verification, and learning — and Field Manual v0 describes that journey. This edition does not describe plant execute, a live connector tag pull, SMTP invite delivery, or automatic revocation of access on expiry as live. It does not describe Sync writing work orders, isolating equipment, or controlling the plant. Simulated or seeded telemetry and assets are practice records. They are not live plant results, and a proxy on a practice screen is not a verified outcome. Self-guided onboarding is not claimed as a live product path.

The series continues with Threshold Is Not Diagnosis, on why crossing a numeric threshold, limit, alarm band, or set-point is not a diagnosis and not a Decision Case. Companion reading: Trend Is Not Truth on why a rising or falling line, a co-moving tag cluster, a week-over-week delta, or a trend narrative is not a recorded causal check and not a verified outcome, Correlation Is Not Causation on why co-moving tags, coincident alerts, and dashboard trends are not a recorded causal check and do not authorize action, Confidence Is Not Evidence on why a strong gut feel, a high model score, or a manager's certainty is not a recorded check, Assumption Is Not Evidence on why a belief, a habit, a prior shift's story, or an unverified model output is not a recorded fact, Unknown Is Not Clear on why missing evidence is recorded as unknown, not as clear or ready, Blank Is Not Zero on why an empty field is not a measured zero, Silence Is Not Clearance on why silence, no reply, no alert, and an empty inbox are not authorization, Alert Is Not Decision on why seeing or clearing an alert is not a decision, Dashboard Is Not Decision on why a display is not authorization, Verification Is Not Optional on why the case stays open until the check is recorded, Evidence Lineage Is Not Optional on why a recommendation you cannot reconstruct is not an industrial decision, and Learning Requires a Verified Outcome on why a later case inherits the verified outcome, not a hoped-for proxy move. A Reliability Assessment asks whether the records can support a conclusion. A Strategic Pilot is a governed proof around one operating decision. The Verification chapter is the public statement of the check this essay refuses to replace with a proxy.

Read the check, then bring a question

Field Manual v0 states the order and the verification step. Proxy is not outcome. A KPI, a leading indicator, a model score, a green tile, a closed work-order count, an MTBF estimate, a sensor health %, or any other stand-in measure can inform a recommendation. They are not the verified operational outcome. They do not authorize, clear a gate, or replace a verified outcome. Learning requires a verified outcome, not a hoped-for proxy move. Sync recommends. A named human decides. Verification still records the actual check. The Reliability Engineer workspace is where a signed-in Decision Case is completed. A Reliability Assessment is the bounded review when the question is whether the records can support a conclusion. None of those is a claim that Sync executes plant work, or that self-guided onboarding is a live product path.