Session LeadersD. Scott Sink, Ph.D., P.E., Director, Integrated LeanSigma Certification Program, ISE at OSU

Jared Frederici, MBB and Senior Consultant, The Poirier Group and

Great Lakes Region Vice President for IISE

Team Member Chapter #1

Council on Industrial

and Systems

Engineering

The New Industrial (and Systems) Engineering:

Operational Analytics to Support Continuous

Improvement Part II

Foundational Data Role

Chapter # 1

Agenda

12:00 pm Scott Tee-up

Quick Review from Part I—the Framework for the Series

Foundational data role, where and how to begin — Jared

12:50 pm Q&A from webinar

1:00 pm Adjourn

ISE and IISE for Life—how IISE supports

you for your entire Career…..

Council on

Industrial and

Systems

Engineering—

Senior

Leadership of

ISE function

Young

Professionals

—Early

Career 1-15

years out

Industry

Advisory

Board—

Mid Career , Mgs

in ISE related

functions

Career Path and Timeline

You can get involved in Societies, Divisions and also ‘Affinity Groups’ like Young Professionals, Industry

Advisory Board and the Council on Industrial and Systems Engineering

Questions?How We’ll

Handle

Please write your question in the webinar

question web form. We will address as

many as we can at the end of the webinar

and send and email with follow up’s to

attendees for those not able to be

responded to.

Chapter #1 Highlights—

IISE’s First Chapter (1949)

and also the first Virtual IISE Professional Chapter (2016)

1. 186 Professional Members in Eastern Ohio but also from around the Country

2. Support, partner with the Youngstown State, Ohio University and Ohio State University

Student Chapters.

3. Partner with the Industry Advisory Board, CISE, and the Young Professionals Group

4. Partner with our Dayton/Cincinnati Professional Chapter on our Annual IISE All Ohio Event

and other things

5. 6 Timely, Valuable Webinars each year; topics developed from Voice of Member

6. 12 Monthly Memo’s help Members get to know each other and keep members aware of

upcoming opportunities AND also provide Self-Help Features on personal and professional

mastery

7. quarterly GoToMeeting small group calls with members that focus on topics of interest

from ‘affinity groups’/segments of our members.

Design for the Series of

Operational Analytics Webinars

(series of 5 at this point)

Webinar #1: Foundations 7 Dec 2017 (and GLR Conference)

Share the Framework, the Models, the Abstractions, the Principles

Management Systems Model

Intel “Triangle” Model

Webinar #2: Foundational Data Role--Measurement and Analysis

Planning March 2018Measurement Planning using Value Stream Maps, Data Models derive from refining the

Management System Model, The Data Management Role of ISE’s in Process Improvement

Projects

Webinar #3: Best in Class ILSS Project Final TG’s April 2018Showcase best in class projects, shine spotlight on Op Analytics

Webinar #4: Decision Support Role—M&A Execution July 2018Feature and Knowledge Extraction, Creating Chartbooks and VSM’s, supporting the

evaluation phase of DMAIC projects and then also the Control Stage.

Webinar #5: Putting it all together Aug 2018

Revisiting the Management Systems Model with Case Examples

Key Points—Story Line

▪ Building a Measurement and Analysis Plan requires a strong point of view regarding the role of

Measurement and Analysis in Process Improvement

▪ With a solid M&A Strategy and Plan, you can increase the chance for lasting improvement in key

processes.

▪ Most professionals aren’t systematic and disciplined enough about the M&A planning, the data model

development and as a result all the data needed to solve the process problems and keep them solved isn’t

available.

▪ The Data Model is reliant on a well done Value Stream Map, have to identify all the ‘control points’

AND have to map upstream and downstream. This then tees up being able to build you ‘causal model’.

▪ Once this is done, then the three ways to get data (ASK, OBSERVE, SYSTEM DATA) are employed to

build the data base (over time) and this then tees up being able to execute your analytics plan.

▪ Illustrative Case Studies will be utilized to bring key points home and demonstrate how the principles

and our method plays out in practice.

The ‘Use Case’

What we mean by use case is what are the types of projects, the types of operational analytics situations we are focused

on. So here are some examples of typical types of process improvement projects we are focused on, they are typical ILSS,

DMAIC type improvement projects:

▪ a “Connectivity Solutions” company serving the automotive industry. Change Over times on extrusion machines is

too long and too variable. The ISE struggled to build a comprehensive data model and also ran into MSA issues

with existing data.

▪ a large bakery, cookie/cracker line and then a food processing plant, a chicken nugget/strip line have yield loss

problems. There are between 11 and 13 ‘control points’ in the value stream and only 3 have data stamps. How do

we optimize process performance without more data stamps?

▪ Transactional Processes: variety of types of transactional processes where data stamps necessary to solve the problem

simply do not exist. Lead Times known but cycle times not known hence can’t zoom in on steps

where waste is occurring from a data fact standpoint.

▪ The challenge and issue and common theme in all of these is that no one has really every really built a rational

data model that would support continuous improvement of core, key value streams. Even in org’s with ERP

systems, the configuration of the data stamps, where control points were conceived and created is almost

always in adequate for Process Maturity Levels 3-5.

Internal usage only

Where are we in the process?

9

Joey M. Greg V.Mike D.

Amber O.

Joe K.Pat M.

MY PROJECT

Floor Scale

Vehicle Scale Pick, pack and ship operations

begin at the end of assembly,

and build the order to

completion with picking and

packing of kits, then send to

the shipping area.

Management Engineering and Process Improvement • The Ohio State University Wexner Medical Center • © 2013 Analyze TollgateManagement Engineering and Process Improvement • The Ohio State University Wexner Medical Center • © 2013 Analyze Tollgate

Data was captured throughout the entire process

Doctor

orders

surface

echo for

patient

Scan is scheduled

through Non-

Invasive Lab at

Ross Heart

Hospital

Inpatient receives

scan in their room

(portable scan)

Results are

read by

physician in

lab

Sonographer

travels with

machine to

patient’s room

Sonographer

preps room

and patient

Sonographer

conducts

scan

Sonographer

finds computer

and enters

manual data

If contrast needed,

nurse will travel

from lab to room

and administer

Sonographer

travels to next

scan or

returns to lab

Process Variables

• Echo Orders Placed

• Quantity: inpatient and outpatient

• Location: In-lab, portable

• Priority: STAT, Priority, Routine

• Scan Type: TAVR, limited, full echo

• Staffing

• Type: Sonographers, Nurses, Physicians,, etc.

• Availability: Sick, Vacation, FMLA, etc.

• Equipment

• Available machines for in-lab and portable

What information are we interested in?

• Turnaround Time

• Goal: 24 hours or less

• Scan Cycle Times

• How long different parts of the process take

• # Scans not Completed Every Day

• How is the process capable of performing?

• Staffing levels

• Priority

• Order Time

Appointment Time

Time Sonographer

arrive in room

• Time call fellow

• Time call nurse

• Time nurse arrive

Results Time

• Staffing Mix

and Levels

• Equipment

Availability

• # of Scans not

completed

every day

• Scan start time

• Scan end Time

• Time sonographer

leaves room

• Time start data entry

• Time finish data entry

Details from scan

process

IHIS Heartlab Data Collection Other

Pt Location

Scan Type

03/20/2018

Management Engineering and Process Improvement • The Ohio State University Wexner Medical Center • © 2013 Analyze TollgateManagement Engineering and Process Improvement • The Ohio State University Wexner Medical Center • © 2013 Analyze Tollgate

Value Stream with graphical

distributions of control point metrics

03/20/2018

Bron-Shoe Projects

5SChris and Simon

Shipments

Open Up

Data Entry

Plating Strip

RepairPolish

Plate

Check Out

Shipping/Packaging

Color out/Inspection

Color up/Inspection

Amelia

VictorSilver Process

Bron-Shoe Baby-Shoe Growth

Repair Last Year

Silver Restoration Project Integration

Lead Time

Create Standard Work

Create a disciplined improvement for critical stages in the process.

Maintain process flow while reducing waste and reduce anticipated turnover risk from obstructing flow.

– Parts waiting at Check Out (Addition of Denna)

– Underutilization of Labor (Incentive)

– Anticipated Repair Turnover (Training/Cross-Training)

– Anticipated Data Entry Turnover (SWP/Training/Cross-Training Plan)

*4 areas of concern

5S ProjectsMy ProjectLast Year’s Project

Focus of Project:Mixing Dumping, Oven and Waste

INPUTS

OUTPUT

Clarifying the Process

Drain Utilization “Hot Spot” Chart Where do we attack Waste Generation?

16

Legend

• 70% of Waste Collected from:

• 25% of Waste Collected from:

• 5% of Waste Collected from:

• 5 Worst Drains:37, 9, 20, 21, 2

• 3 “Areas” to Target: Raw

Process, Bagger Area, DishwashData Collection Defense

Improvements and Financial Impact

$57,660

$103,562

$49,746

• Currently improvements of $107,406 in increased run capacity

• Through SMED and optimization software compliance there is $103,500 left in potential increased run capacity

X = 16.03 min/unit X = 7.62 min/unit

Improvement Implementation Plan

Improve Recap – Staging Solution Elements

Current StateStandardization

Maintenance and Staging

SMED

Verify/Audit

Forecasted

Actual

Click event on timeline for in-depth explanation

So far we have seen a reduction in:

• Average by 3 minutes

• Deviation by 4.5 minutes

Current State Future State (Forecasted)

Management System

Model

Who’s Managing

and Improving

(run it, improve it,

cater to crises, do

the dumb)

Upstream

System(s)

Customer,

Suppliers,

upstream

processes

Downstream

Systems (SC,

Customer)

Bottom Half of

Intel Analytics

Triangle (data

management

role)

Top Half of Intel Analytics Triangle—

well engineered “Dashboards”

Provoke timely and effective decisions

and actions (shorten ‘latency’)

▪ “Above the line” analyst role

• Extract features based on questions you have to answer by

‘torturing’ the data until it speaks to you and others. Pick right

metrics of interest!!

• Apply curiosity & business acumen to data & analyses – create new

knowledge, insights, ‘aha’s’

• Apply data visualization techniques to aid in telling the right story –

as in life, so in business: the best story wins …Develop the Art

of Powerful Visualizations and stay focused on driving the

‘end game’

Goal!!!

▪ Then it’s a recursive process of top

down and bottom up in this model.

▪ what do you need to provide to

decision makers so they can make the

right adjustments to processes so they

perform better

▪ And, what data do you need in order

to provide that information?

▪ “Above the line” analyst role

• Extract features from data through

statistical analyses

• Apply business acumen to data &

analyses – create new knowledge

• Apply data visualization techniques to aid

in telling the right story – as in life, so in

business: the best story wins …

▪ Foundational data role

• Select and gather data from many

sources, preferably through automated

extract, transfer, & load (ET&L) process

• Assure data are cleaned & ready for

analysts to use – data quality monitors

• Assure data are integrated & can be

joined with other data – think LEGOs

• Assure data storage is high reliability &

user-friendly – SSAS cubes, databases

Adapted from S. Cunningham;

Intel Corporation; 2013

There is Estimated to be $371B Left

on the Table for Manufacturers

through OA

Source: Capgemini, Why companies need to focus on operational analytics, 2017

▪ The $371B opportunity for manufactures specifically, comes from a 2014 study from technet

▪ A 3X multiplier is huge comparing customer facing improvements vs. operational

▪ But don’t disregard the $162B in productivity, where OA isn’t out of scope

▪ In a recent study, 75% of US organizations are prioritizing operational analytics over customer

facing / front end or back office analytics for improvements…

▪ Canadian Minimum Wage up to $15 / hr in Ontario by 2019, $15 / hr in California in 2022

Agenda

12:00 pm Scott Tee-up

Quick Review from Part I—the Framework for the Series

Foundational data role, where and how to begin — Jared

12:50 pm Q&A from webinar

1:00 pm Adjourn

The Intel Analytics

Triangle Framework

▪ Starts with the Questions!!!!

▪ “Above the line” analyst role

• Extract features from data through

statistical analyses

• Apply business acumen to data &

analyses – create new knowledge

• Apply data visualization techniques to aid

in telling the right story – as in life, so in

business: the best story wins …

▪ Foundational data role

• Select and gather data from many

sources, preferably through automated

extract, transfer, & load (ET&L) process

• Assure data are cleaned & ready for

analysts to use – data quality monitors

• Assure data are integrated & can be

joined with other data – think LEGOs

• Assure data storage is high reliability &

user-friendly – SSAS cubes, databases

Adapted from S. Cunningham;

Intel Corporation; 2013

Case Study: Why starting w/ the top

of the pyramid can be deceiving…

Cleansed over

200 metrics

and created a

operational

metric

database

In many cases,

different

departments

were

calculating

metrics

differently,

wrong

decisions were

being made

1. Data Gathering – Challenges,

Advice, Examples

▪ My Data Doesn’t Exist

• Go get it!

• Determine data type needed

• Determine how much of it is needed

and work backwards:

• http://www.raosoft.com/samplesize.html

• http://www.nss.gov.au/nss/home.nsf/pages/S

ample+size+calculator

• https://www.surveymonkey.com/mp/sample-

size-calculator/

• Also think about what future tools

you may need it for (Minitab min and

max for normality example)

• Remember Power value targets (.8)

• Remember practical vs. statistical

significance

• Remember ROI and cost of data!

1. Data Gathering – Challenges,

Advice, Examples

▪ Sampling Plans

Random Sample:

• Equal chance of any data point being retrieved

• May select those not in target group or interest,

indiscriminate

• Sample sizes need to be larger typically to be

representative

• Can be expensive – remember ROI of data….

Stratified or Segment Random Sample:

• Samples based on the basis of a representative

strata or segment

• Still random, but more focused

• May give more relevant information

• May be more cost effective

Quota Sampling:

• Still sampling by segment but not random

• Specific number in each segment sampled

• Cheaper, but may not be full representative

(remember MSA bias, autocorrelation)

1. Data Gathering – Challenges,

Advice, Examples

▪ Data Gathering

• How much of it do you need to solve

your unique challenge?

• Remember to work backwards

• What’s the cost of getting to .8?

• Pair statistical power with

pragmatic, common sense

What Tests Will Be Ran?

Remember, .8 is usually the minimum acceptable level for statistical power

1. Data Gathering – Challenges,

Advice, Examples

▪ Measurement Plan is Everything!

• Ensure operational definitions are clear not only

with you, but your stakeholders

• Think with the end in mind, holistically, not just

about measurements you need right now

• What other measurements ‘could’ you get while

undergoing manual sampling or installing

measurement system? Could you reduce rework

later by adding elements?

What Tests Will Be Ran?

1. Data Gathering – Challenges,

Advice, Examples

▪ Quality of your Measurement

System

• It won’t be feasible or practical to complete

Gage R&R’s on every metric in your

measurement system

• Although doing so where appropriate will

allow you to check the quality of your

sampling plans

• Target <30% R&R

1. Data Gathering –

Challenges, Advice, Examples

▪ My Data Does Exist But I Need IT

• This applies to both student and professional projects

• We often go through an extremely inefficient process of

obtaining data from IT. 10.3% VA Ratio, 20 Day Lead Time!

Total Lead Time = 19.5 Days

VA Days/% = 2 Days = 10.26%

NVA Days/% 17.5 Days = 89.74%

Find the appropriate IT resource to ask, send an email with rough estimation of what

is needed

Determine IT Resource, Ask IT for

Data

Wait for Response Wait for Response

Cycle Time ~ 1-2 Days

VA Ratio = 25%

NVA Ratio = 75%

Cycle Time ~ 1-2 Days

VA Ratio = 0%

NVA Ratio = 100%

Measure Phase Warrants Data Request, Fulfillment and Analysis

Process

DMAIC / DFSS Project

Typically, at this stage, IT will ask for more granularity, what specific fields are needed and

rework will occur

IT Request for More Information

Rough estimation of what data could be needed from system

Determine Data Needs

Cycle Time ~ 1-2 Days

VA Ratio = 100%

NVA Ratio = 0%

Wait to Ask

1-2 Days 3-5 Days 1-2 Days

Once you receive the data, you normally need a few minor

changes (different time series, additional field, etc)

Receive Data and Request Minor Modifications

Cycle Time ~ .25-.5 Days

VA Ratio = 0%

NVA Ratio = 100%

Wait for Response

1-2 Days

Receive final data from original request and begin analysis

Receive Final Data from Original

Request

Cycle Time ~ .25-.5 Days

VA Ratio = 80%

NVA Ratio = 20%

Analyze Data

4-5 Days

Normally about 25% through analyze, you realize you need

additional data and make another request.

Realize You Need More Data and

Request

Cycle Time ~ .25-.5 Days

VA Ratio = 0%

NVA Ratio = 100%

Wait for Response

1-2 Days

Receive data from 2nd Request and finish analysis

Receive Final Data from 2nd Request

Cycle Time ~ 1-2 Days

VA Ratio = 100%

NVA Ratio = 0%

Customer: Student Receives Data Needed

to Finish Project

Data Request and Fulfillment Process

1. Data Gathering –

Challenges, Advice, Examples

▪ My Data Does Exist But I Need IT

• Think about the rework loop created

• Really nailing your measurement plan will help

• Huge reduction in lead time and VA Ratio

Data Request and Fulfillment Process

Complete Data request form, send to appropriate IT resource

for formal, detailed request

Complete Data Request Form, Request Data

Wait for Response

Measure Phase Warrants Data Request, Fulfillment and Analysis

Process

DMAIC / DFSS Project

Full estimation of all data needs based on proper measurement

plan

Determine Data Needs

Cycle Time ~ 4-6 Days

VA Ratio = 100%

NVA Ratio = 0%

Wait to Ask

.25 Days 3-5 Days

Receive final data request

Receive Final Data Request

Cycle Time ~ 1-2 Days

VA Ratio = 100%

NVA Ratio = 0%

Customer: Student Receives Data Needed

to Finish Project

Total Lead Time = 12.25 Days

VA Days/% = 7 Days = 57%

NVA Days/% 5.25 Days = 43%

2. Data Selection – Challenges,

Advice, Examples

• So now you’ve got data, either from

your own sampling or from IT, etc.

• Look for systems in your data!

• Are you observing variation within

processes, or separate systems

(with perhaps different

measurement systems….?)

• Leverage rational subgroups!

2. Data Selection – Challenges,

Advice, Examples

• Use rational subgroups to do a variety of things – minimize variation within

subgroups to isolate signal, minimize noise.

• You may have too much data, need pragmatic subgroups – or isolate a

potential “X”

3. Data Storage – Challenges,

Advice, Examples

▪ Data Storage

• How much did you obtain?

• Of what type is the data?

• Relational database or object

oriented? (SQL, noSQL)

• How secure does it need to be? Is

there any confidential information?

• How fast do you need to recall it?

• Will you have multiple users in the

system to pull data?

• Will you be doing analysis

alongside of the data?

• What if you get so much that you

exhaust excel and Access's

limitations?

• Think with the end in mind…!

3. Data Storage – Challenges,

Advice, Examples

▪ Note that

while you may

be thinking

here

▪ You should be

thinking here

▪ Because

there are

ramifications

here

Basic Storage and Analysis

▪ Microsoft Excel

▪ Microsoft Access

▪ Microsoft PowerPoint

▪ Visual Basic / VB.net

▪ Arena

Advanced Storage and Analysis

Existing Skills +

Data Analytics and Visualization

• SQL

• R

• SAS

Programming

• Java

• Python

Big Data Technologies

• Hadoop

• MapReduce

• Pig

• Tableau

• D3

• Ruby

• CPLEX

• Hive

• Hbase

• Aster

3. Data Storage – Challenges,

Advice, Examples

You may need to move from “Basic” to “Advanced”

4. Data Cleansing – Challenges,

Advice, Examples

▪ Data Cleansing

• Remember VersaCold example

• Sometime it’s easier when you collect

your own data…

• Tie source data interactions and

calculations to your final BI interface

• Often times issues are in calculations or

“between” interface work (example on

rounding issues)

5. Data Integration – Challenges,

Advice, Examples

▪ Data Integration

• To begin to tell the entire story, you’ll

likely need to bring in data from different

sources

• Some could be from your own

measurement system, some from IT,

some extracts or “pulls”, some

automated via FTP, etc.

Note that you

typically need to

leverage some

data warehouse

“type” of interface

to handle this,

either yourself, or

within your

infrastruture

5. Data Integration –

Challenges, Advice,

Examples

5. Data Integration –

Challenges, Advice,

Examples

5. Data Integration – Challenges,

Advice, Examples

Data CubeData Model

Organized Data in Pivot View

• Once you’ve isolated the sources, and

have brought them into a “data

warehouse” type of application, create a

data model

• Leverage “cubes” and “hypercubes” in

your data model for efficient processing

5. Data Integration –

Challenges, Advice,

Examples

Sales

Revenue $

Sales Person

Leveraging Cubes and Hypercubes

The Question?• How do we get the right stuff in the right place at the right time

in the right amount?Data Selection & Gathering• What data in AX is most important?• Connect AX to Access with ODBC linksCleansing• Pull tables into Access • Check data integrity issues/errors; analyze for “good” vs “bad”

outliers• Qty constraints can cause these issues (MOQ, pack qty)Integration & Storage• Create a relationship diagram of the tables within Access• Ensure the data is organized, connected through like SKU

numbersFeature & Knowledge Extraction• Pull data into Minitab and complete exploratory data analysis• Observe trends in variation, subgroups, and use information to

create formulas/heuristic models for key metrics• Standard deviation of demand is the biggest driver of SS levelVisualization• DONE goal: a visual database and user interface that is

sustainable and easy to use by the Sutphen team

• Act like a Business Intelligence interface• To update (edit) metric levels like order quantity,

reorder point, safety stock• To track volatility (standard deviation), perform

outlier analysis, generate reports• To observe, analyze, and create charts• Looking at reports to flag levels that need adjusted,

altered temporarily due to upcoming usage

Sutphen Example – Bottom of Pyramid Work

Where is the data coming from within the database?

AX, Purchase order history• Part #• Qty• Date

order, date receive

• LOC• Vendor

AX, Transfer order history (from WH)• Transfer

destination

• Trans. source

• Qty• Part #

AX, Kanban info• Min• Max• LT

transfer• LT

reorder• Vendor• Part #

AX, BOM • Initial• Final• Factor

into demand forecasting

• Part(s) #

AX, Count Journal• Time stamp

of date PO sent

• Part #• (will be in

purchase order history)

DATA SOURCES

DATA ELEMENTS

ACCESS• Pull in (6) tables

separately from AX• ODBC link

• One master table to refresh automatically

Minitab and Exploratory Data Analysis

Formulas and Heuristic Models (1 tool, for

sustainability and simplicity)

Use Detailed Process Map, mark all control points (where to collect X data ) – use case, what’s important to the user?

Where detailed design

begins*

AX, Item Master• Cost• Part #

(item ID)• Descriptio

n (item name)

• MOQ

Data Modelling and Whse Development

1. Start with the

Fundamental Questions

2. Figure out what data

elements exist, where,

etc.

3. Build data model (AIR

Database) and organize

into a ‘hypercube’

4. Sit an analytics app on

top

5. Torture the data to

support answering

questions and to

provoke timely

improvement decisions

and actions

The Data that

doesn’t exist that

you collect with

based on your

Measurement

Plan

Excel or Access Minitab

..\Data Modeling,

Sample Sizing and

Measurement

Planning.pptx

MATERIAL

JOB

Job #

Notes

A

Cust.

D

L

Seq. # BFA

OFA

Bill Date

Hours

Weight

Description

Est. Weight

Contract Value

Actual Cost

BilledEst. Cost

Name

Work

1 1

1

1

N

N

Material Note

UseN

M

Description

1

N

Contains1

N

Delivery Date

(0, N)

(0, N) (0, N)

Actual Hours

Est. Hr/Ton

Est. Hours

Contract Value ($)

Job Name

JQuote #

Actual Weight

Actual Hr/Ton

Fab

PEOPLE

Name

ROLEName

Association

SUBCONTRACTOR

Name

POEstimate

Quote #

CHANGE_ORDER

CO #

Contains

SEQ_STATUS

Seq_Status

Contains

Contains

In

N

1Hour_Proportion

JOB_FINANCIAL JOB_STATUSJStatus

In

1

N

Contains

Work(0, N)

(0, N)

(0, N)

(0, N)

Association

Phone

Email

QUOTE

JOB_SEQUENCE

CO_STATUS

CO_Status

InN

1

THE CURRENT SYSTEM’S FLOW OF INFORMATION:

Change Request Phase

Change Execution Phase

Resource Information

Job Beginning Phase

Job Execution Phase

Is this flow of information

design feasible?

Key points

1. Good analytics come from good problem statements, access to the right data, and applying the right techniques

• Good problem statements should have an associated decision support and business outcome linkage

• Analysts need an all-access pass to data & the wisdom to collect it—do not only search where the ‘light’ is best and don’t be shy

• Simple techniques trump complex techniques—modeling the problem properly, logically often is the biggest issue

2. Some people have every skill (business acumen, data, technique) to perform a good analysis – but it tends to result in a slow ‘craft’ process

• Analytics exist on the learning curve – what used to take six months now takes two weeks with the right data and analytics

• There is little time for ‘craft’ development in most businesses – speed wins, so rapidly building proficiency at Op Analytics is critical for Young ISE Professionals

Key points

3. Investment in the data foundation has a positive ROI, as analysts

move faster when they trust the data – results in faster results

• First lesson of AIR – if storage is cheap, store it – you might need it later, get

the data model and granularity decisions made right the first time.

• Second lesson of AIR – maintain the illusion of simplicity for your customers,

they appreciate clarity and logic and straight forward answers to complex

questions that make sense to them.

4. Good data visualizations can tell the right story quickly, because

people are predisposed to believe what they see in a chart …

• Be on guard! Some folks use How to Lie with Statistics as a field guide

• Kahneman’s lesson: W Y S I A T I – what you see is all there is

S. Cunningham; Intel Corporation; 2013

Kahneman’s Lessons

(Thinking Fast and Slow)

1. Intuitions are not as reliable as we think

2. Decision Making Outcomes are Unpredictable and not always ‘rational’

3. People seek data that is compatible with Beliefs they already have

4. WYSIATI—what you see is all there is: human beings are predisposed to believe

a story that best fits the facts presented, even if the data and facts are one-sided or

incomplete.

5. Beware of the Anchoring Effect (whatever number is ‘on the table’ will have a

strong influence on what people conclude and get ‘stuck’ with (e.g. the median and

average are often the enemy in process improvement projects)

6. Beware of the Analyst who is too certain

51

Key points

5. There is very positive ROI in getting these decisions right – small analytics teams can wield disproportionate influence on the bottom line

• Consider the hiring guidelines – there are few people who have the curiosity, sense of urgency, tolerance of ambiguity, and humility for this role

• The catalyst role is very powerful – get in, learn, analyze, win, get out

6. Good analytics drive positive action – in every organization we’ve done work with/in, simple/influential beats complex/impotent every time

• Consider each example we shared with you – in each case, we can point to one-time or on-going positive actions in our environment

• Moreover, this is why we’re planning to be fast-followers in Big Data …

Agenda

12:00 pm Scott Tip-off

The Need, the challenges, the issues — Jared

Intel Analytics “Triangle” Model/Framework—Scott

Foundational data role, where and how to begin — Jared

12:50 pm Q&A from webinar

1:00 pm Adjourn

Q/A From Webinar

Design for the Series of

Operational Analytics Webinars

(series of 5 at this point)

Webinar #1: Foundations 7 Dec 2017 (and GLR Conference)

Share the Framework, the Models, the Abstractions, the Principles

Management Systems Model

Intel “Triangle” Model

Webinar #2: Foundational Data Role--Measurement and Analysis

Planning 20 March 2018Measurement Planning using Value Stream Maps, Data Models derive from refining the

Management System Model, The Data Management Role of ISE’s in Process Improvement

Projects

Webinar #3: Best in Class ILSS Project Final TG’s 25 April 2018Showcase best in class projects, shine spotlight on Op Analytics

Webinar #4: Decision Support Role—M&A Execution 24 July 2018Feature and Knowledge Extraction, Creating Chartbooks and VSM’s, supporting the

evaluation phase of DMAIC projects and then also the Control Stage.

Webinar #5: Putting it all together Aug 2018

Revisiting the Management Systems Model with Case Examples

Upcoming Lunch and Learn Webinars

from Chapter #1 (Eastern Ohio and

Virtual Chapter for IISE)

Upcoming Lunch and Learn Webinars: FREE, over lunch time, Usually Tuesdays,

GoToWebinar Format!!

▪ 25 April: Best in Class ISE/ILSS Certification Project Final Tollgate Case Studies

• Gunnar Smyth, Mount Carmel and Maria Pandolfi, OSU Med Center—Improving Flow and Reducing Patient

Wait time in two Clinic’s/Practices

• Allen Drown, Transmet—Developing and Deploying a Standard Work System to Reduce Business Risks

• Joseph Wegar, Peerless Saw—Developing and Deploying an Employee Engagement Program to drive

Continuous Improvement

▪ 12 June: Operational Analytics Part III—Putting it all together More Case Studies

from Jared and Scott

▪ 6 Sept. CISE Career Choice Points #3 (Jim Dobson, Disney; Rudy Santacroce,

CarlsonRTKL (health care architecture firm); Kelli Franklin-Joyner, UPS.