Senior Design: Issues of consulting, how to run a meeting, value stream maps

Andrew Johnson Department of Industrial and Systems Engineering

Texas A&M UniversityCollege Station, TX

Consulting and services

Strategic

Solution

Technical

strategic consulting

solution consultingtransformational consulting

technology consultingIT services and integration

OR Consulting

Decision support v. decision automation

Decision support Giving managers and executives the insight

and information to make hard business decisions

Ad hoc, what-if, investigative analysis Incremental development Solve the problem, and the problem

changes Relatively inexpensive to design and build

Decision support v. decision automation

Decision automation “Hard-coding” of a prescriptive model for

some core business process; “Master Planning”

Runs periodically to determine well-defined business metrics

Static problem and solution statement Usually expensive to design, implement and

change

The OR consulting value proposition

Quantitative framework – If you can’t measure it, you can’t manage it

Business process abstraction for decision making –optimization, simulation and statistical

modeling

Confronting and taming complexity, uncertainty and chaos

Too many choices, too little information

“You’re not an expert in broadband hyperdrive infrastructures, so how

can you help your team’s company?”

The OR consulting process for decision support

compelling business questions

“How do I reduce inventory?”“How much should I buy?”“When should I build a new production

facility?”

analytical model/framewor

k

OptimizationSimulationStatistical analysis

actionablebusiness

recommendations

Recommendations that help managers and executives make decisions and answer the compelling business questions

analytical solution

Results of computational solution process; suitable for analysts, not managers

criticalsuccess points –requiresskill andimagination

Models and prototypes for decision support

1. Insight a wind tunnel for business –

test it before you build it uncovering non-obvious relationships;

get surprised early not just what and when, but also why

2. Collaboration* building analytical business models is a team sport operational prototypes are rallying points for

discussion, experimentation, assumption-testing, and validation

a “concept document” means something different to everybody;a team can collectively get their mind around an operational model

*Serious Play: How the world’s best companies simulate to innovate, Michael Schrage, Harvard Business School Press (2000)

Models and prototypes for decision support

3. Analysis• dealing with complexity, uncertainty and chaos• charts, graphs and numbers to help people make

decisions

Evaluating clients

Who’s in charge, and who is paying the bills?

Does the client already know the answer before the right question is asked?

What’s the client’s main concern: the problem or the process?

Does the client want to be your partner or your boss?

The client will be evaluating you, but you should also be evaluating the client

Designing successful OR consulting engagements

Know the engagement scope What will be included in the engagement? More importantly, what will not??!! “Scope Creep” produces failed engagements,

unhappy clients and hungry consultants Know what you know, and what you don’t know

Does this engagement use well-known technology and methodology? (And do I know how to do it?)

Or does it require inventing new ideas?

The Proposal Why it is important

Sets everybody’s expectations in writing

Defines the engagement scope Specifies the completion criteria

(when can I graduate?) Elements

Key assumptions and dependencies Consultant and client responsibilities Deliverable materials – content, format, level of

detail Schedule – for both client and consultant Change control procedure

Practical OR Issues1. Too many OR projects end up solving the wrong

problem. Why is that? And how can we avoid it?

2. OR is blessed (cursed) by the ability to solve both strategic and operational business problems. Why should we recognize and characterize this capability?

3. OR projects are data-driven. IT projects are data-driven. What’s the difference?

4. A successful OR project causes far-reaching ripples in an enterprise. Who in an organization is affected, directly and indirectly, by OR solutions? And why should we care?

Issue 1: Solving the right problem

In most IT projects, the objectives, requirements and deliverables are cast in stone before work begins.

In many Practical OR projects, determining the right objectives and requirements is part of the problem-solving process.

Solve the problem and the problem changes. Active modeling participation is an educational experience for the client. (“Ah-ha moments” alter requirements)

Scope creep: the tension between doing what’s expected and doing what’s right.

Issue 2: Data – IT vs. OR

For most OR projects, “data” means

data; I need the right numbers to make

my model operate properly.

OR data: engineering vs. business known vs. unknown digital vs. virtual (e.g., “promised”)

The Killer: level of data detail (what am I trying to model?)

Data: the really hard part of OR consulting

Most IT projects are concerned with the form of data; OR is about content: accuracy level of detail completeness

Define in the proposal the data the company will make available to you and the data you are responsible for gathering

Document in detail the data requirements: form and content of project data solution scenarios for testing and validating application

prototypes

Issue 3: Strategic vs. Operational With minor modifications an OR model can

often solve related strategic and operational variants of a business problem. (The technical difference is usually in the data.)

Example: planning and scheduling

Conflicts and disagreements begin when You don’t know if you are solving the strategic

or operational problem; You, the practitioner, thinks it’s strategic and

the client thinks it’s operational, or vice versa; It’s not documented

PlanningLook at the big pictureTells you what to do

SchedulingLook at the details

Tells you how to do it

Strategic Operational

Purpose

What to doBig capital investmentsEnterprise-wide strategy

How to do itProblem solving in immediate time or space

ScopeEnterprise-wideTouches many functions

Local effectsFocus on one or few functions

Model time horizon

Years or monthsVarying-size time periods

Weeks down to minutesUniform time periods

Data detailHighly aggregatedMany simplifyingassumptions

DisaggregatedPerhaps a few short-cuts

Model technologyAlgorithmicE.g., optimization –LP, some CP

HeuristicSome LP, CP, heuristics

Constituency

Executives, managers with strategic responsibilities

People responsible for operational functions

Issue 4: OR-based decisions: The Players…

OR-based decisions (“decisions”) often have enterprise-wide implications and ramifications.

Many people can be involved and/or effected by decisions.

Hopefully the enterprise will win as the result of a good decision.

But some people will win, and some will probably lose.

Assumptions

Who are the players involved/effected by decisions?

If you recognize and plan for these players, will it make you a better OR practitioner?

Questions

…OR-based decisions: The Players

Increased credibility Estimate project cost, time Estimate ROI for client Estimate chance of

success

As OR practitioners, why should we care about the players?

d m

p

DecisionOperator

Quick practical OR check-list

1. listen, and2. It’s not just models, it’s …

• workflows• infrastructures• user interactions• cultural acceptance and integration• people, money and resources• change (most people don’t like it)

3. Have a little fun

What is our last minute check-list before talking to a prospective client about a practical OR engagement?

How to run a meetingMost people don’t like meetings. They say they are boring, go on too long, and don’t get anything done. And often that’s the truth. So to have a good meeting, you need to make it interesting, keep it on track, and make sure something gets done. Here are seven steps toward making your next meeting a success.1. Set a goal for the meeting. Be very clear about why you’re having the meeting, and what needs to get done or be decided. Break that task into steps, or divide the discussion into sections—that’s the agenda for your meeting. At the start of the meeting say, this is our goal, and if we can get this done, the meeting will be a success. At the end of the meeting remind them that you achieved your goal. This lets everyone leave feeling successful, and they’ll be glad to come to your next meeting.

How to run a meeting (cont.) 2. Put decisions to the group. The

participants own the meeting. Let them set the agenda before the meeting, or at least add to it when you begin. If decisions need to be made about the process (whether to end a discussion that’s going too long, for example) then ask that question to the group.

3. Stay on schedule. Remember that every minute a person spends in your meeting, they could be doing other things. They’re with you because they’ve decided your meeting is important, so treat them like their time is important. Start on time and end on time!

How to run a meeting (cont.)

4. Pay attention to what’s important. Set a certain amount of time for each item on the agenda, based on how important it is. If the group starts spending a lot of time on details, ask them “Is this what we want to spend our time talking about?” A lot of details can be worked out by individuals or committees—meetings are for the decisions that need to involve the whole group.

5. Keep the meeting on track. Your agenda is the tool you use to make sure you’re on time and on the right topic. When side issues come up, help the group get back on track. If the issue sounds important, check with the group. “We’re talking about a new issue—is this something important that we should take time to discuss?”

How to run a meeting (cont.) 6. Have good facilitation. The facilitator is the

person who runs the meeting and acts on all the steps listed above. It’s a big job, and it usually doesn’t fit well with participating in the discussion. So if you need to have your views heard, let someone else run the meeting! Good facilitation doesn’t just happen—it’s a skill that comes with training and practice.

7. Make sure people participate. People think a meeting is useful based on one simple thing: whether or not they talked. So everybody should have a chance to share their ideas. It’s okay to ask specific people what they think! You should also be prepared to gently remind people when they’re talking too much.

Update ReportYour update should have the following 8 sections: #1 What were our goals for this week? #2 What did we accomplish this week? #3 Why are numbers 1 and 2 different? #4 What are my goals for next week? #5 How does this fit in to my bigger picture? How

does #4 fit into my longer term goals? (What are the longer term goals?) What deadlines are looming beyond the week horizon?

#6 What does the company expect from me this week?

#7 What do we need from _____ to achieve the goals we have set?

#8 What will we do or what can we do if we do not receive the things we need from ______

Order RecievedTraveler Created Fit and tack

assemblies

Cut Pipes

Pipe Supermarket

Vendor Supplied Inventory

Ready for Hydro Test?

Staging Area

Magazine Rack

DiameterOf Weld

TIGWelding

Fluxcore Welding

Tack or Weld?

Hydro TestYes

No

> 4“ < 4“

Material Handling

Material Handling

Tack

Weld

Material Handling

Material Handling

Manifold ProcessFlow Diagram

Value Stream Map

Value Stream Mapping is a Lean technique used to

visualize and analyze all the actions (value-added and

non-value added) of materials and information currently

required to bring a product or service to a consumer.

Queue Block

Process Block

Department

Transport Block

CT ?

BS ?

Res ?

Data Box

Sample Value Stream Map

Current State – Future State

Value Stream Mapping

Software Visio Excel Others available via the web

Its not enough to just identify the problem

To identify the future state you have to decide what to change

In order to decide what to change you need to estimate the impact of the change

This is an area where IE models help

Thank you for your attention

Good luck on your projects!