Operational Evidence

I Used This Before I Built Anything

This page shows that the system I'm building in software already worked — without software.

I didn't start with a product. I started by using Drucker's feedback analysis through Ultraworking's monthly planning system at a paper mill.

The progression happened across three roles over two years. Each role added a layer of proof that the discipline works under real operational pressure.

What I'm Claiming

This is not a new idea.

It's a discipline that:

  • I used across three different roles
  • held up under real operational pressure
  • helped teams make better decisions over time
  • scaled from individual practice to enterprise-level impact

The software just makes it easier for others to do the same.

The Three-Role Progression

Phase 1: Performance Development Leader (Feb 2023 - May 2024)

What I Did:

Started using Ultraworking's monthly planning system on February 6, 2023. Every day, I captured:

  • what happened
  • what decision was made
  • what it suggested for next month

Not after the fact. Not cleaned up later. In the moment.

Example from the daily log (Feb 15, 2023):

What happened: "Decision point: Went with ten updating current SOP's for Paper Machine. Shared LSW with Derek Davis."

What it suggested: "What are the unarticulated fears and beliefs of the Paper machine key leaders and team members?"

This wasn't storytelling. It was capturing the decision and the uncertainty that came with it, before we knew how it would turn out.

What I Built:

From February to April 2023, I cross-referenced what I captured daily with conversations with leaders and operators. By April 10, 2023, I completed the Learning and Development Strategy FY23 Workup — the Strategy A3 tool.

The Strategy A3 wasn't a framework I applied. It was synthesized from operational data — what I captured daily + what I learned from conversations.

It had one mission at the top ("Stabilize Paper Mill Fundamental Operations"), and then every initiative was tied back to that mission with owners and tactics. Every week, I'd check: Are we still aligned? What changed? What do we need to adjust?

The Outcome:

Structured workforce training, problem-solving processes, and baseline procedures that aligned the organization to a single mission.

Phase 2: Shift Leader (May 2024 - Aug 2024)

What Changed:

I embedded into frontline operations. Instead of observing from a training role, I was on the floor making decisions in real time.

I started using Copilot for voice-to-text capture. While walking the floor, I'd capture:

  • what I was seeing
  • what the dashboard said
  • what the operators said
  • where the gap was

Then I'd validate with operators and document it using Situation-Complication-Question-Answer (SCQA) format for shift reports.

I did this for seven months on every shift. I collected 3,000 pages of documents.

What I Found:

The same patterns showed up that I'd seen as Performance Development Leader, but now I had real-time operational data to prove it.

The experience gap told the real story. The senior operators had 33, 29, 26, 26 years of experience. Then it dropped to 13, 10 years, and then fell off sharply.

That wasn't in any report. I found it by watching who was actually doing the work and how long they'd been there.

The Outcome:

Real-time documentation of inefficiencies. The delta between what happened on shift (especially night and weekend shifts) and what was communicated in meetings became visible.

Phase 3: Shift Leader / Koch GenAI Champion (Aug 2024 - Present)

What Changed:

I led the development of AI-powered shift knowledge capture systems using the Microsoft stack (Copilot, OneNote, Teams).

The method stayed the same: capture decisions in the moment, validate with operators, document patterns. But now I had GenAI to accelerate the capture and analysis.

What I Analyzed:

  • Total downtime analyzed: 1,408.7 hours
  • Root causes identified: Equipment issues (deferred maintenance) + process issues (installation procedures)
  • Operator workarounds documented: Masking systemic problems the dashboard couldn't see

The Key Insight:

The operators weren't the problem. The system was. The dashboard measured "uptime" (machine running), not "effective uptime" (machine running at full capacity). The operators knew this. Management didn't.

The Outcome:

  • Annual opportunity identified: $64,095,850 (Monticello facility)
  • Recovery target: 30 hours/year (2 shifts) = $1.365M/year per facility
  • Enterprise scale (400 sites): $546,000,000/year recoverable value

Strategic Framework:

  • Zero capex required: Uses existing Microsoft stack (Copilot, OneNote, Teams)
  • Scaffolds operator memory: Embeds judgment capture in daily flow
  • Turns downtime history into financial recovery: Measurable ROI
  • Demonstrated success: Live field data, not simulations or abstractions

The Loop That Stayed Constant

Across all three roles, the loop was the same:

  1. Capture: What did I actually see? (daily)
  2. Validate: Does this match what operators experience? (real-time)
  3. Pattern: What keeps showing up? (weekly/monthly)
  4. Adjust: What should we try next? (based on reality, not assumptions)

The tools changed (Ultraworking template → Copilot voice-to-text → GenAI integration), but the discipline stayed the same.

Why This Matters

A lot of tools try to help after things go wrong.

This way of working helps before stories harden and blame sets in.

It keeps decisions close to the people making them. It keeps learning tied to real work. It keeps momentum without pretending everything is under control.

What the Software Does (And Doesn't)

The software:

  • makes it faster to capture things
  • keeps context from getting lost
  • makes patterns easier to see over time

It does not:

  • tell people what to think
  • replace judgment
  • automate leadership

If the discipline isn't there, the software won't help.

The Documents That Prove It

This isn't theory. These are the actual documents from the work:

  1. Initial Training Analysis (Feb-March 2023: Ultraworking daily capture + monthly planning)
  2. Learning and Development Strategy FY23 Workup (April 10, 2023: Strategy A3 tool)
  3. Shift Reports (May-Aug 2024: 3,000 pages of SCQA-formatted documentation)
  4. Downtime Analysis (Aug 2024-Present: 1,408.7 hours analyzed, $64M opportunity identified)

These documents show the system working across three roles in a real industrial environment under real operational pressure.

Bottom Line

I'm not asking anyone to believe in a theory.

I'm showing the work I already did — and building software to make it easier for others.

TL;DR

I used Drucker's feedback analysis through Ultraworking's monthly planning system across three roles at a paper mill (Feb 2023 - Present). Started as Performance Development Leader, built Strategy A3 from daily capture + conversations. Became Shift Leader, embedded into operations, used Copilot for real-time capture (7 months, 3,000 pages). Became Koch GenAI Champion, analyzed 1,408.7 hours of downtime, identified $64M annual opportunity (Monticello) scaling to $546M enterprise-wide (400 sites). Zero capex, Microsoft stack only. The documents prove it worked. Now I'm building software to make it easier for others.

Next steps: Try the method yourself, then let me know what you find.

Ready to start? Get the Ultraworking spreadsheet and begin your 21-day capture [blocked]

Have questions? Share your experience or ask for guidance [blocked]