Duty Manager Briefing — Entegris Penang Plant 2

CMP slurry mixing line showing anomalies at 2:15 AM.
You have 3 hours 45 minutes to decide.

The Chemical Mechanical Planarization (CMP) slurry mixing line — supplying four semiconductor fabs with time-critical batches — is recording SPC breaches. The next scheduled delivery is at 6:00 AM. Senior leadership is unreachable until 6:30 AM. One process technician is on site. You must decide: continue, partial hold, or halt for inspection.

⏱ Decision deadline: 5:45 AM latest
🏭 Line: CMP-MX-07
📦 Delivery: 6:00 AM, 4 fabs
013-Question Filter
→
02Analyze Information
→
03Two Tests & Decision
→
04Pre-Mortem Check
→
05SBAR Brief
Stage 1 of 5

3-Question Filter

You have 10 data sources in front of you. Not all of them belong in your decision. Use three questions to separate signal from noise — fast.

The three filter questions

  • Q1 — Relevant? Is this data relevant to the decision I need to make right now?
  • Q2 — Reliable? Is this data reliable enough to act on?
  • Q3 — Current? Is this the most current data available?

If a data source fails any one of these — it goes to the side. Not discarded. Noted. But it does not drive your decision.

How to use these cards

Click a card to select it, then click one of the three tagging buttons below to classify it. Tag all 10 before continuing. Your goal: narrow to 3–4 critical signals.

Tag selected card as:
10
data sources
›
0
filtered out
›
0
uncertain
›
0
critical signals
Reflection
Which data source surprised you most — and why does it pass or fail the filter?
Stage 2 of 5

Analyze with two tools

More analysis does not mean a better decision. You need two tools: Pareto (which signals drive 80% of the risk?) and trend reading (where is this heading, and how fast?).

Tool 1 — Pareto (80/20)

The vital few: which 2–3 alarm types account for 80% or more of the total alarm count? Click each bar to mark it as vital few or trivial many. Everything after the 80% line is noise.

Alarm frequency — CMP-MX-07 (10:00 PM – 2:15 AM) Click to mark vital / trivial
Vital few (above 80% line) Trivial many

Tool 2 — Trend reading

A reading tells you where you are. A trend tells you where you are heading. For each parameter, identify the direction and whether the rate of change is accelerating, constant, or slowing.

SPC process data — CMP-MX-07 (10:00 PM – 2:15 AM)
Time Particle count (/mL) Status Slurry pH Status Agitator vibration
Spec: particle count 400–750/mL | UCL 800/mL | Slurry pH 10.0–10.4 | LCL 9.96 | Agitator vibration normal < 2.8 mm/s
Trend question 1
Particle count is rising while slurry pH is dropping. What single equipment fault could cause both simultaneously?
Trend question 2
The rate of change between 1:00 AM and 2:00 AM is steeper than between 10:00 PM and 11:00 PM. What does an accelerating trend tell you about your decision window?
1-sentence analysis summary
Write your summary using the formula: Signal + Trend + Urgency
Example format: "[Signal] has [trend direction] [rate] while [second signal] has [trend] — the combined pattern indicates [interpretation] requiring a decision by [time]."
Stage 3 of 5

Two tests before you decide

Before committing to any call, run two tests. Together they tell you how cautious to be and what specific conditions must trigger a change in course.

Test 1 — Reversibility

The core question

Can I recover if I am wrong? High stakes + irreversible outcome = always choose the cautious path, regardless of probability. High stakes + reversible outcome = act, but set a tight trigger to reverse course.

Option A — Continue production
If you continue and the particle count causes contaminated slurry to reach the fab and be used in CMP processes — is this recoverable?
Option B — Halt for inspection
If you halt the line and the inspection finds no fault — you were wrong. Is this recoverable?
Test 2 — Decision thresholds

Set your IF / THEN lines now — before the pressure hits

A threshold you have not written down does not exist. Set the specific, measurable conditions that will automatically trigger action — while you are calm, not reactive.

Your halt triggers for CMP-MX-07
Fill in threshold values. Be specific — a number, a count, or a time duration.
/mL on 2 consecutive readings → halt
for more than 20 minutes → halt
mm/s at any point → halt + escalate immediately
Your decision
Based on the reversibility test and your thresholds — what is your call right now for line CMP-MX-07?
Stage 4 of 5

Pre-mortem check

Assume your decision is wrong. Project forward to 72 hours from now and work backwards from the failure. Strong decisions survive a pre-mortem. Weak ones get improved by it.

Why this matters — confirmation bias

Once you have made a decision, your brain amplifies evidence that confirms it and filters out evidence that challenges it. The pre-mortem is the deliberate antidote — it forces you to search for the disconfirming evidence before it is too late to act on it.

Step 1 — Imagine the worst (5 min, no filter)
It is 72 hours later. Things went badly wrong. Write every possible failure mode — no editing, no probability weighting. The ones that feel unlikely are often the most important.
Step 2 — Plot your top risks on the matrix
Click each cell to mark where your highest-concern risks sit. Red cells = critical zone — require immediate preventive action. Use the labels from your list above (R1, R2, etc.).
← Likelihood →
Rare
Unlikely
Likely
Almost certain
Critical impact
High impact
Medium impact
Low impact
Click cells to mark where your risks land — cells show R1, R2, R3 as you click
Step 3 — Prevention and contingency
For each of your top 2 risks: what do you do NOW to prevent it, and what is your backup if prevention fails?
Risk 1 — what you do NOW to prevent
Risk 1 — contingency if it happens anyway
Risk 2 — what you do NOW to prevent
Risk 2 — contingency if it happens anyway
Does this change your decision?
After the pre-mortem — does your call from Stage 3 hold, or do you want to revise it?
Stage 5 of 5

Communicate clearly — SBAR

You are calling the Plant Director at 2:30 AM. You have 60 seconds. Write your SBAR brief — then see how every element connects directly back to the work you did in stages 1–4.

The 3 communication traps

  • Data dumping: leading with all the numbers before stating the problem. Give the conclusion — not the spreadsheet.
  • Hedging without a recommendation: presenting options without committing to one. Be decisive AND appropriately cautious.
  • Burying the recommendation: saving the actual ask for the last 30 seconds. Lead with the decision. Reasoning second.
S
Situation
What is happening right now — 1–2 sentences, current state only. Draws from: your Stage 1 critical signals.
B
Background
Context the listener needs — parameters, maintenance history, staffing, deadline. Draws from: your Pareto vital few + WO CM-8812 from Stage 1 filter.
A
Assessment
Your professional judgement — data + trend + risks on both sides. Draws from: your 1-sentence summary from Stage 2 + reversibility test from Stage 3.
R
Recommendation
Your specific ask — what you need from the listener. Always close with next update time. Draws from: your decision + thresholds from Stage 3 + pre-mortem safeguards from Stage 4.
How all five stages connect into your SBAR
Your complete decision journey
0110 sources → 3–4 signals
→
02Pareto + trend reading
→
03Reversibility + thresholds
→
04Pre-mortem + safeguards
→
05SBAR to leadership
Each stage feeds directly into the next. Your SBAR is not a separate task — it is the output of the entire chain.
Entegris · DDD Capstone Decision Exercise
Participant Decision Summary
CMP Slurry Mixing Line CMP-MX-07 · Plant 2 · 2:15 AM Scenario

Capstone complete

Below is a full record of every decision you made and everything you wrote — from your first data filter to your final SBAR brief.

Your personal commitment
Name one specific decision you face in the next 30 days where you will apply this process — and which stage you will use first.