Skip to main content
RDMAIC

RDMAIC · Process improvement

There is room for improvement in every process.

I make the work visible, use data to separate signal from noise, and help the people who own the process decide what to improve.

Sometimes that means improving an end-to-end process. Sometimes it means redesigning one recurring workflow with AI. The starting point is the same: real work, useful evidence, and human ownership.

Let’s talk Explore the ways to work

Worked example · supplier delivery

The supplier is late. Where should the conversation start?

Generalized worked example of supplier deliveries. Its identifiers, values, chronological observation order, and four analysis-stage boundaries are site-created. Delivery margin is zero on the promised date and negative when late. Period 4 has the highest late share. Part families B and D concentrate the lateness but have different histories. The M and R coded characteristic combination concentrates late deliveries across both families. One final-focus delivery remains on time, providing a useful contrast for supplier questions. The example does not identify a company or prove a cause. Individuals chart of generalized delivery order and delivery margin. Limits are calculated separately within P1–P4.

A supplier-level KPI shows the concern. The delivery history helps decide where the supplier conversation should begin.

Site-created values · investigative pattern preserved

Delivery margin across four analysis stages

80 observations · 20 in each predefined stage
Delivery margin in days · 0 = promised date · negative = late

Scroll to follow the complete Individuals chart · P1 → P4

  • On time or early
  • Late
  • Control signal
  • Stage centre
  • Control limits
  • Promised date · 0

P4: 11 of 20 deliveries are late. No observations are beyond their stage control limits in this example.

The focus becomes specific

  1. Part families B + D B 11/24 · D 14/24
  2. Operating combination M / R 19/24 late
  3. The next question becomes specific
    What changed for these part families under this operating combination—and what was different in the deliveries that still arrived on time?

This identifies where to investigate; it does not prove that the supplier or coded characteristics caused the lateness.

See the full analysis and distribution plots
Generalized supplier-delivery evidence

Generalized case scenario: identifiers, values, chronological observation order, and analysis-stage boundaries are site-created while preserving the investigative pattern.

What the data supports. Lateness is concentrated in part families B and D under the M/R characteristic combination. Period 4 is weakest overall. Family D shows an older issue; family B deteriorates later.

What it does not prove. The coded characteristics or the supplier caused the lateness. That requires supplier process knowledge and additional evidence.

Individuals-chart calculations by stage
Stage Observations Centre Average MR LCL UCL Late
P1 20 +1.260 1.684 −3.220 +5.740 4/20
P2 20 +1.050 1.763 −3.640 +5.740 5/20
P3 20 +0.840 1.758 −3.836 +5.516 5/20
P4 20 −0.430 2.668 −7.528 +6.668 11/20

Questions to take to the supplier

  1. What changed at the beginning of period 4 for these part families and this characteristic combination?
  2. Do these deliveries share a production route, resource, planning rule, material flow, or dispatch process?
  3. At which milestone does delay first appear: confirmation, material readiness, production start, completion, or dispatch?
  4. Why did part family D show lateness earlier while family B deteriorated later?
  5. What was different in the delivery from the same focus group that still arrived on time?
  6. Which milestone timestamps or operating classifications should be examined next to test these explanations?
Site-created example values in a site-created observation order and four predefined analysis stages, grouped by period, part family, and coded characteristics. Negative delivery margin is late.
Period Part family Characteristic 1 Characteristic 2 Delivery margin in days · 0 = promised date · negative = late
4 D M R −3.9
4 B M R −2.8
4 B M R −3.4
4 D M R +0.4
4 B M R −4.1
4 D M R −3.2

Two primary paths

Choose the path that matches the work in front of you.

Both paths begin with the current work and end with your people able to make a better decision.

01

Project Coaching

Improve a process with the people who run it.

I coach one improvement professional through their own live project, or support several Green Belts and Black Belts in one organization as each advances a separate project. The practitioners perform and own the work; I bring experience, questions, and challenge across six sessions.

  1. Define + EDA
  2. Measure + EDA
  3. Analyze
  4. Analyze → Improve
  5. Improve
  6. Control
The outcome

From a visible current state toward a change worth testing and an evidence-based adoption or revision decision.

Explore Project Coaching
02

AI Workflow Sprint

Learn to redesign recurring improvement work with AI — inside real constraints.

I facilitate a learning-by-doing cohort for improvement and development professionals. Each person or pair brings one recurring workflow from real work, examines it step by step, decides what should stay human, what is better handled with deterministic automation, and where AI should contribute, then tests one bounded first experiment in the organization’s actual environment.

Per-workflow experiments · shared learning

A bounded first experiment for each workflow, plus a shared view of the controls and organizational enablers worth developing.

Explore the AI Workflow Sprint

A practical improvement rhythm

See the work. Decide with evidence. Choose the next useful action.

What happens next depends on the work. Coaching and the Sprint continue into real use; a bounded analysis may instead end with a clearer decision or the next evidence question.

  1. 01

    Make it visible

    Map the process, workflow, handoffs, constraints, and variation as they really are.

  2. 02

    Decide what matters

    Use data and working knowledge to find the question worth answering next.

  3. 03

    Act on what you learn

    Test a focused change, make a decision, or identify the evidence still needed — while the process owner keeps the decision.

Supporting paths

Bring in analysis or build the capability to continue.

Data Analysis

Use data to decide what deserves action.

In a bounded engagement, I turn an agreed question and suitable data into a focused analysis and a clear visual story. Scope and secure transfer are agreed before files are shared.

Question → analysis → decision
Explore Data Analysis

Training

Build improvement capability across a group.

For groups of four or more, I teach Lean Six Sigma and customized improvement programs through real processes, projects, data, and tools — so the capability stays inside the organization.

Method → practice → capability
Explore Training

Human ownership

The method serves the people doing the work.

Your team keeps the decisions. I make the work easier to see, draw on experience from coaching improvement projects, and bring analytical questions and practical challenge to the next decision or experiment.

Jukka-Matti Turtiainen, Lean Six Sigma Master Black Belt and founder of RDMAIC Oy
Jukka-Matti Turtiainen · RDMAIC Oy

I’m Jukka-Matti Turtiainen, a Lean Six Sigma Master Black Belt certified by Gregory H. Watson. My work connects process knowledge, data, and practical experimentation — across industry, healthcare, improvement education, and international capability building.

  • 25+ Lean Six Sigma courses delivered
  • 150+ improvement projects coached
  • ESTIEM Green Belt co-creator
Read about my background

Conversation

Let’s talk about what kind of support would help.

Project Coaching, the AI Workflow Sprint, Training, and Data Analysis fit different situations. We can discuss the work in front of you and which kind of support would be most useful.

Let’s talk jukkis@rdmaic.com