INFORMS Roundtable Meeting at San Antonio

February 1998

 

Administrative Items

Michael Alliston welcomed Roundtable members and guests to the winter retreat at San Antonio, TX. Mike described the gift and letter the Roundtable has already sent to Cynthia Rinaldi, the prior INFORMS Roundtable Administrative Support person. Meeting evaluation forms were provided to each attendee, and Mike encouraged each member to return the completed form to him or George Freestone during the meeting. Summary results from the forms which were returned are enclosed as Attachment 1.1. Detailed results can be obtained from George Freestone. Additionally, Mike provided the "Future Roundtable Programs" document to all, and it depicted the preference voting (based on feedback at the Fall 1997 (Dallas, TX) Roundtable meeting) for topics the Roundtable might discuss at future meetings. Mike encouraged all in attendance to record up to ten votes of preference, as well as up to ten votes for deletion, and return the list to him during the San Antonio meeting. The results of this Winter 1998 (San Antonio, TX) preference voting can be found in Attachment 1.2. Mike described the minutes of four past meetings provided to each member at the meeting (Dallas, TX (Fall 1997); Boston, MA (Spring 1994); St. Petersburg, FL (Winter 1994); and Phoenix, AZ (Fall 1993)) which had been edited (with strike-throughs for text which was to be removed based on George Freestone's and Chuck Beall's recommendations) for public availability from the Roundtable Homepage. Mike encouraged anyone with suggestions or corrections for the public minutes to provide that feedback to George Freestone during the San Antonio meeting. Those minutes (with any additional edits) will be posted shortly after the end of the meeting on the Roundtable Homepage.

Dr. Alliston asked Scott Parrish to introduce the evening’s keynote speaker.

 

Evolving Technology, Digital Technology and Implications for Operations Research, Led by Mr. Kevin O’Toole, Director, Strategy and Technology Intelligence, US West Advanced Technologies, Boulder, CO.

Mr. Parrish introduced Kevin O’Toole, who had been scheduled to address the Roundtable in Dallas, but experienced some very severe weather during his fall trip, which had to be canceled. Scott indicated that we are fortunate to have had Mr. O’Toole with us to provide an updated address. Kevin’s slides can be found in Attachment 2.0

Mr. O’Toole said that the rate of change in the telecommunications industry is truly phenomenal at this time. As an example of how significant some of this change is - he said that when the first three stories in The New York Times relate to your business, you know that something unusual is underway. The Chief Technology Officer (CTO) of U. S. West mentioned that the change will be unlike anything since the Industrial Revolution. Kevin asked Roundtable members to think about ‘internet time’ vs. ‘real time’.

Kevin said that his talk would be grounded in macro trends. He will cover

Where We Have Been

Kevin commented on some major technological forces in recent history:

Moore’s Law suggested that the capacity of an IC would double every eighteen months. The implication is that what cost $10,000 today will cost $100 in ten years. This is an example of an exponential, rather than a linear, world. Kevin added that many experts feel that Moore’s Law may have another twenty years left. Modems are another example of where the Moore’s Law type of thinking applies. It is amazing to think of all the places where ICs are now, owing to their economics - a Craftsman drill has a processor in it.

Kevin then discussed some recent events regarding fiber optic transmission and software. First and foremost, the cost to ship a megabyte has dropped like a stone. Second, software has very seriously lagged in matching the productivity improvement of chips and optical transmission. Some factors affecting software are:

In short, there have been few home runs and many heartaches.

Mary Calabro asked why software has tended to lag behind the other two areas so dramatically in terms of productivity improvement. Kevin responded that software is very complex and purpose specific. Further, software requirements are not given enough thought, so a lot of software just gets ‘jammed’ into use and is not terribly successful. Kevin also commented that many times there are overblown expectations. Mary commented that the other two items (chips and fiber transmission) are mechanical in nature, which helps too. Tom Spencer built on that observation by saying that the first two depend on physical attributes, while software depends on many physiological factors. David Heltne made this observation concrete by commenting that his group will provide a $100,000 piece of functionality for optimization and the IT team will add a $300,000 graphical user interface which will suck up the processor’s speed and capability. Gerry Hoffman said that many software companies exist now (which is optimistic for improved productivity regarding software development) since we will be much more prone to build once, and then use several times (across many companies).

Where We Are Headed

Kevin then switched gears to discuss where he thinks we are headed.

Kevin began by discussing access. Presently, from the home to the backbone modems limit data transmission to 56 KBPS. However, cable modems and digital subscriber lines will deliver 1.5 MBPS. Being ‘always connected’ will become the operational paradigm. Gerry Hoffman asked how the technology will deal with having many ‘always connected’ users. Kevin responded that communications companies will have to go to a packet network, peel off the data prior to the switch and become much more ‘packet aware’. This will imply multiple services over the same wire, e.g., it is not a phone OR a FAX or an ISP connection, but rather it is phone AND FAX AND an ISP connection AND a cable TV connection, with caller ID on TV.

John Lucas asked about those who are not tethered to a network, and how these changes will affect them. Kevin answered that low earth operating satellites will play a key role. Also, battery life is improving to assist these users. Kevin said that wireless phone communication while traveling down the highway at 70 mph is still a problem in terms of quality of transmission. Kevin also stated his opinion that the growth in this ‘personal’ (rather than ‘business’) area would be as significant as for business purposes. What’s missing is some compelling wireless application. Gerry Hoffman described a business application catering to the home market, which is heavily networked where a firm provides ‘over the net’ consulting regarding potential menus and complementary wines for dinner parties. The material for the dinner, along with preparation instructions and complementary wines are drop-shipped to the requester’s home. Gerry Hoffman also observed that the ‘wired’ data transmission rate will set the requirements for ‘wireless’, because some folks will occasionally be tethered and frequently mobile. Kevin described a business application where the network provides access to experts who will provide several potential answers to your problem - and the requester pays $50,000 only if he or she uses one of the recommendations. Kevin mentioned that cross subsidization of cost will probably be involved. In general, Kevin feels that business will be willing to pay more for comparable service. Personal applications could include spectator virtual reality games. David Brinkley’s memoirs asks what to do with television, given these changes in telecommunications and the proliferation of the Internet.

Kevin continued by discussing some potential opportunities with the communications network backbone; such as SONET (the synchronous optical network), ATM (the asynchronous transfer mode), and IP switching. Chuck McCallum said he assumed ATM switches will end up in the backbone to enhance the quality of service. Gerry Hoffman inquired if there is a distance limitation for using SONET. Kevin answered no, but there are factors which are dependent on the type of fiber that you are using.

Kevin described the growth in voice vs. Internet data throughput. Voice is growing at about 4% per year while Internet traffic is growing very rapidly. Kevin expects Internet traffic to overtake voice traffic in the year 2000. As an interesting example of this - presently on the U.S. / U. K. link - data traffic exceeds voice traffic. Kevin described how he thinks the Internet will evolve, and this is shown in his Attachment 2.0 also.

So What?

Kevin commented on the implications of what is forecast.

First, Kevin quoted Andrew Grove - "only the paranoid will survive".

He also suggested the importance of predicting the strategic inflection points (where the growth turns to exponential increases). Kevin thought that some significant outcomes include:

Kevin forecasted highly interconnected high speed data networks that will flourish during continued periods of deregulation and lead to new models for computing resource distribution and information access. Presently, companies are spending on average $14,000 per year on physical PC related costs. This will drive down the use of network computing machines to manage these costs.

Some "losers" will include Blockbuster Video (and their fellow video store entities). Incidentally, it is amazing how much of a video store’s income is generated from late fees. Video on demand will take a deep toll here. Software and book purchases will shift to on-line activity. Newspaper classified ads may also be at risk.

AOL and other ISPs will be big "winners". Incidentally, AOL adds 6,000 customers daily.

Implications for OR

Kevin concluded his comments with a few thoughts on new OR questions and new OR opportunities.

New questions for OR:

New OR opportunities:

Kevin suggested we all "stay tuned" because the trends leading to these opportunities are all going exponential or vertical.

Discussion

Gerry Hoffman inquired about the rise of the "stupid network" (attributed to David Eisenberg at Bell Labs) where intelligence will move to the periphery of the network. Kevin responded by suggesting he expects a rubber banding effect where the intelligence goes out to the extremes of the network, but then returns to the "center". Mike Alliston suggested that others could then manage software updates for me on my personal PC at home. Kevin answered that this could take the form of the ultimate of dumb networks where all of the processing capabilities are remote from the home PC, or where the smarts for updating a "smart" home PC is centrally controlled. The question remains as to how many of us would trust others to update our home PC software.

Kevin contrasted telephone companies (high capitalization, large cash flows, but restrictions on what they can do) with cable companies (debt laden, terrible reputations, but good potential).

Kevin also commented on the demand-driven nature of the WWW, vs. the highly structured "take it when I want to show it to you" network broadcasting approach. A large portion of WWW published material may not be what I want to see anyway, published by authors I may not want to read. However, I can control what I choose to read.

David Heltne asked what percentage of households are involved in this activity today. Kevin answered with some rough estimates - ~40% have PCs, ~7% are ISP-active. Kevin forecasted that within 5 years 50% of the homes will be ISP-active.

Chuck McCallum asked if banks would disappear. Kevin O’Toole suggested that some functions of a bank will become more readily available on the Internet, while some functions will likely be retained by the bank.

 

Roundtable Monday Morning Introductions and Administrative Items:

Introductions: Mike Alliston

Dr. Michael Alliston encouraged each Roundtable member or guest to introduce him/herself to the group and to comment on recruiting activity at the firm or institution they represent.

Mike concluded by soliciting the number of firms currently growing their staff vs. those maintaining or downsizing their staff. For those that replied, there were six firms expanding and six contracting or maintaining their staff size.

 

INFORMS Roles Coming Available: Dr. Arthur Geoffrion

Art Geoffrion, Past President of INFORMS, reported that Randy Robinson will be leaving INFORMS August 1st. His role is that of Executive Director, and there is a search committee trying to come up with a person to replace Randy. The candidate could be a person making mid-career change, or someone who is recently retired. Interested candidates should approach Art Geoffrion or any other INFORMS leader.

An INFORMS nominating committee is looking for a President-Elect, Treasurer, Meetings V.P., IT V.P., and several other positions. The President position is approximately three days a week, while the others require about one day a week, Art reported. Please direct any potential nominations to Art.

INFORMS is also looking for an Associate Editor for Practice for INFORMS On-Line. Mike Trick runs INFORMS On-Line, with eleven Associate Editors. Art suggested that INFORMS needs an Associate Editor to develop content for practitioners, and it is estimated to require about ˝ day per week. Please contact Art or Mike Trick with the name of any interested candidates.

 

INFORMS Opportunities for the Roundtable: Dr. Thomas Spencer

Tom Spencer, the INFORMS V.P. of Practice, mentioned that Roundtable members should be on the lookout for nominations of firms for the INFORMS Prize. Any nominations should be made to Tom Spencer or George Freestone, and self-nominations would be welcomed.

Secondly, Tom mentioned that we are tentatively considering having a facilitated meeting at Montreal where Roundtable members could contribute to the general INFORMS effort to improve national meetings for practitioners. One alternative being considered is to change the format of the two annual INFORMS meetings such that one of the meetings would be heavily focused on practitioner needs. Tom said that we might have a professionally administered and facilitated meeting at Montreal to accomplish this. George Freestone suggested that the Roundtable has taken on this issue many times in the past and that it would be a mistake to do this again without having a specific proposal from INFORMS in the hands of Roundtable members prior to such a meeting.

 

OR and Recruiting Session 1: Panel Discussion - What Works Our Groups Are Doing - The Demand for New OR Recruits, facilitated by Mr. Scott Parrish, Distinguished Member of the Technical Staff, U. S. West Advanced Technologies, Boulder, CO.

Scott Parrish introduced the four sessions that will be held on Monday to discuss OR and recruiting. The first session will set the stage for recruiting by describing some typical work that is going on in OR functions within industry. Scott introduced each of the four panelists in this first session.

Mr. Bruce Nemlich

Director of Management Science

Pfizer Corp., New York, NY.

Bruce Nemlich overviewed some strategies his group at Pfizer uses:

Mr. Shane Logue

Senior Industrial Engineer, Management Engineering Services

Eastman Chemical, Kingsport, TN.

Shane Logue began by reviewing the matrixed organization. Each manager has two roles: one for a group of people, the second for a specific technology domain.

The group has five technology platforms:

Mr. Tom Bowen

Manager of Management Science

American Greetings., Cleveland, OH.

Tom Bowen began his discussion with an overview of American Greeting’s business and some information about the OR group there. Tom mentioned that of all greeting cards purchased, 90% are purchased by women. This is a key fact of relevance to marketing and sales efforts at American Greetings.

The OR department at the firm presently has eight people in it and helps solve business problems with actionable and creative solutions using operations research and statistical techniques. There are two sub-groups within the OR function at American Greetings - one that serves marketing and sales, and a second that focuses its attention on operational problems.

Mr. Bowen suggested that while recruiting for staff to serve in this group, a key factor he looks for is one of a "business perspective".

Dr. Thomas Spencer III

Head, Analytic Modeling and Applications

AT&T Labs, Somerset, NJ.

Dr. Spencer’s slides are found in Attachment 3.1. An overview of his group is included as Attachment 3.2.

Dr. Spencer began by describing the mission of his group - the Business Operations Analysis group, its organization, and the business model in place at AT&T Labs.

 

OR and Recruiting Session 2: Panel Discussion - What Kind of People We Seek and How They Are Introduced to Our Organizations, facilitated by Mr. George Freestone, Manager of MIS and Co-Manager of Materials and Systems, Air Products and Chemicals, Inc., Allentown, PA.

George Freestone introduced the second session regarding recruiting of OR candidates to firms. In this session, each panelist will describe (1) what kind of people are sought and (2) how they are introduced into the firm. George introduced each panelist prior to their presentation.

Dr. Charles McCallum

Operations Research Director

AT&T Labs., Holmdel, NJ.

Dr. McCallum’s slides can be found in Attachment 4.1.

Chuck initiated his discussion by reviewing a mission statement for his group.

Dr. McCallum then covered some details concerning his staff, breaking up the issues into employee items and contractor items:

Dr. McCallum then moved on to how his group recruits:

Chuck commented that entry level staff generally work with senior staff members on two to three person projects for some of their early assignments. Early on training generally occurs for these new members of the team in the following ways:

Chuck feels that the following lessons have been learned regarding recruiting from his group’s perspective:

Scott Parrish asked how entry level staff’s consulting skills are trained. Chuck said that is one area where his group does not do well.

Dr. Mary Calabro

Engineering Associate

Exxon Research and Engineering., Florham Park, NJ.

Dr. Calabro’s slides can be found in Attachment 4.2.

Mary indicated that Exxon has a formalized corporate program for recruiting which includes considerable communications material. Exxon hires more candidates from campuses than from the experienced job pool. The recruiting effort is a team-based approach. There are specific requirements for employment candidates (in terms of GPA by discipline). Early assignments at Exxon are ones that provide a broad set of experiences. Many new hires are chemical engineers.

Dr. Calabro described some of the requirements for positions at Exxon. They include:

The recruiting effort at Exxon is a very organized effort, which includes:

Mary went on to describe entry level positions in operations research at Exxon. A new recruit will work as a protégé with an expert, senior team member in operations research. Typical projects include: optimization, simulation, scheduling, or statistical analysis. Extensive training takes place each week over the first six months of employment. This training is a mix of formal and informal. Some travel will also be required during the first six months of Exxon employment.

Lynne McCullough asked about the level of training in the Management Science group, specifically whether it makes sense to invest heavily in training an individual who may leave the group after two years. Mary answered that some of the training involves areas which will be of general utility.

Gerry Hoffman asked if Mary’s experience is that chemical engineers either know or can readily learn many of the skills required in the management science group there. Mary answered that this is particularly the case with optimization tools - the chemical engineers rapidly can apply these tools and techniques. Mary further explained that she targets Masters level candidates for work in the management science group.

George Freestone asked what is good about the Exxon recruiting approach. Mary replied that Exxon does tend to get top candidates from schools, the company has an effective network of new hires, and that Exxon will spend money to ensure effective recruiting is accomplished.

Dr. Michael Alliston

Senior Vice President, Risk Prediction

MasterCard International, Purchase, NY.

Dr. Alliston’s slides can be found in Attachment 4.3. He explained that his presentation will reflect the experience he has attained at several firms, but he has a bias in his comments towards smaller groups.

Mike suggested it was important to consider the context of the position that is to be filled. The approach will be different if the organization is:

  • Building.
    • More experience is required.
    • A ‘teacher’ is called for.
    • An independent project manager is desirable.
    • The ability to partner is also key.
  • Growing.
    • A journeyman capable of persistent hard work is useful.
  • Continuing.
    • An apprentice can be assigned a buddy to learn.

Mike continued by describing desirable characteristics and effective screens for those attributes.

  • Personality.
    • Have a large number of persons conduct the interviews to assess this.
  • Organization / Culture. (particularly for a small group, you cannot tolerate an unhappy camper)
    • Will the candidate be happy in the role being considered?
    • Will others like the candidate?
    • Have a frank discussion about any workplace realities ahead of time to head off false expectation problems.
  • Technical knowledge.
    • View degrees (academic history) and projects (applied history) as a package.
    • Statistical background is useful - should know how to deal with dirty data, cleaning up data, and algorithmic processes.

Dr. Alliston offered some thoughts about how to recruit.

Mike said that entry level positions do not really exist in a small firm. However, in a large firm they are prevalent. Small groups tend to be like tricky puzzles where the pieces are difficult to fit together. Mike indicated it is helpful to do at least vague career planning so that there is a rough way forward for a three-year horizon.

Mike completed his presentation with some comments about early training for new employees in a small group.

Jim Hobbs inquired how extensive should the interviewing by multiple people be. Mike answered that he has seen most effective interviewing with seven interviewers. Make sure there are about three that the candidate will work with and be sure to include others outside the group also. Mike cautioned against more than seven interviewers, claiming that they do not provide much additional value or information. Jim Hobbs mentioned that Boeing sometimes asks candidates to give ad hoc presentations during the company visit.

Dave Heltne asked if it was necessary to look for candidates with IT skills, rather than relying on the MIS function for that. Mike said that he has seen MIS staff struggle when digging into data (possibly using SAS). His opinion is that MIS groups are typically not very good at trying something to see if it will work. John Lucas commented that he would prefer a second rate algorithm and a superior interface, owing to the need to have people use the computer tool in question.

Art Geoffrion asked how important resume design was. Mary Calabro answered that there is an Exxon form to help structure that data.

 

Big Waves Which Will Provide Opportunities for MS/OR Practice, Led by Dr. Gerry Hoffman, President, Senior Strategy Group, Chicago, IL.

Art Geoffrion introduced Gerry Hoffman. In addition to the many notable items listed in the agenda for the meeting, Art mentioned that Gerry Hoffman was one of the founding fathers of the Roundtable.

Gerry said his presentation would do the following with big waves: find them, catch them, surf them, and also not get wiped out by them. Gerry provided an outline of his talk:

1. Why Should Operations Researchers Care About Big Waves?

Gerry asked what is OR. OR is what people do. Good OR practice is about enhancing business actions. That definition unfortunately does not differentiate OR practitioners from others involved in business. OR is also about mathematical models. OR people turn worries into structured problems. In Shane Logue’s presentation, he spoke of a senior manager wanting comfort that the OR group there was going to be able to meet a commitment. Another way of describing OR is in terms of chaos theory. There is order if we can identify the generating function, the model. The trick for those of us in operations research is to make a model a useful abstraction of reality.

Operations research and management science are two words that we have struggled with. Creating models tends to favor science, while applying solutions to the real world conveys engineering.

Unfortunately for the OR field, the world takes our models as they move from OR specialists to the generally undifferentiated world of MBAs. This has happened with spreadsheets, scoring models, financial analysis, mathematical programming, fuzzy logic, statistical analysis, "reengineering", and logistics. Gerry said that OR is good at bringing new models into the world and then passing them on. Gerry’s first project at Amoco was a primitive spreadsheet which would aggregate data and provide basic planning tool functionality. It was used to create a corporate plan. Gerry described some railcar optimization work in which he had been involved with Dave Hirshfeld. The size of the problem was 10,000 rows and 10,000 columns with 1,000 integer variables. Minolta’s autofocus camera is now employing fuzzy logic, as is the military. All MBAs are now capable of rudimentary statistical analysis. OR has also unfortunately blown an opportunity of business process reengineering. F. W. Taylor had been doing that since the 1890s, and industrial engineering is closely related to OR, even academically at many schools.

Why should OR people care about these big waves:

Art Geoffrion has made a statement regarding big waves and it is posted within the Shared Destiny Initiatives web page.

OR/MS has mostly missed all the "big waves" and management mantras (AI, TQM, BPR, etc.) even when we could ride them to greater prosperity during their time of favor. We need a way to sense them early, study their merit, determine what OR/MS has to offer, publicize them to the membership when this would serve a purpose, and help practitioners take advantage of these natural episodic opportunities.

For starters, Gerry said that we need to sense these big waves early. The value of this is clear. BPR is now owned by Index in Cambridge. Who has the sense to identify the big waves early? Airline magazines are often leading indicators. Forbes is good, but Business Week is often wrong. Watch the IPOs that serious investors are investing in. Watch the current big waves to identify what comes next. For example, Jim Champy will tell you that 70% of BPR efforts fail, so think about the market for how to clean up the mess after reengineering.

…sense them early…

Gerry listed and the group supplied some potential new big waves, some of which may be too near the beach for a good run:

…study their merit…

Gerry described how to spot some of the important ones. All candidates may be easy to spot. Some of the many opportunities include:

Gerry encouraged the use of the Willy Sutton rule: Go where the money is (especially important for bank robbers). If your candidate is not beginning to get some publicity now, either you are premature or the big wave is not going to happen.

The evaluation of the quality of the big wave is harder. Gerry described his model of a big wave:

Gerry asked if that was what happened with MBO. Gerry recommended the best time to get aboard the big wave is when the first failure is publicized. Gerry suggested you budget for making mistakes, because you will.

…determine what OR/MS has to offer…

This is a very creative activity, but OR people should be good at this because of many of the basics of creativity are part of the systems approach that we have learned.

Gerry warned that some of the more recent graduate students who can only do a limited set of things. Some of their research is limited to typing in search criteria on the Internet.

…publicize them to the membership when this would serve a purpose…

The Edelman Prize has demonstrated that money motivates OR practitioners to talk about their work. This example can be leveraged.

…help practitioners take advantage of the naturally episodic opportunities…

Gerry pointed out that although the big waves may be naturally occurring, the opportunities have to be found. The field of dreams does not happen, you have to market it.

2. The Information Revolution as El Nino:

Gerry focused his discussion on a few the big waves caused by the information revolution. The information revolution is not a big wave itself, but many of its consequences are, and many of the big waves we listed earlier are related to the information revolution. Some of these consequences are just emerging (these consequences that may be big waves) and they are frequently viewed with awe. Big bucks are spent on IT, it is the largest single corporate capital expenditure category. It has a large impact on the workplace, operations, lives, and organizations.

Gerry encouraged us to look for big waves that present big problems, particularly where expenses grow faster than revenues, where the problems are not solved yet, or where we really can provide some help - based on our skills set:

3. Three Big Waves of Importance to Operations Research:

Big Wave #1: The Value of Information Wave

The Worry

Known, budgeted IT expenditures range from 2% to 30% of expenditures. John Lucas suggested that 7-8% is typical, based on his experience. Actual budgeted expenditures may well be double what is public. IT is the largest capital expenditure category in the U. S., accounting for about 18% of all capital. Most senior management hates IT, frequently saying they don’t know where the money goes and feeling they have no option to control or change the amount spent. Worse yet, IT expenditures are growing faster than any other category of expense.

Steven Roach at Morgan Stanley has looked at the value of this IT expenditure. Paul Straussman, a first-class statistician, has also. Steven concluded (roughly) that all this money is wasted; firms are not getting anything back. Some of this analysis, Gerry mentioned, had been based on dubious SIC code data. Paul looked at annual report data and searched for the correlation between IT expenditures and "success". Unfortunately, he found no significant correlation.

The Problem

The IT profession has already found the low hanging fruit, the easy cost displacement opportunities have been identified and capitalized on. Some new initiatives center on joint efforts with the business where the benefits cannot be readily allocated between the IT and the overall effort. An example of this is sales force automation where field reps would have direct access to inventory levels, for example. Success of the overall effort depends as much on the users of the system as the system itself. Success is consequently hard to measure. Business based metrics linked to the IT resources required to make success happen would be helpful.

Some new initiatives are for systems to support corporate strategy, for example - to establish world wide connectivity to "make us one company". This IT expenditure will be very difficult to determine the value of. The same problem occurs where IT efforts support the corporate vision. It may be easy to quantify the value where the IT effort supports the company’s operations.

Succinctly, the problem is how to measure the value of information so that we can do the cost / benefit calculation required to make sound decisions about new IT initiatives, and be comfortable with the decisions. Also, this is much more difficult to do before the IT effort commences, rather than retrospectively where it may still be tough.

Kinds of Models Needed

Gerry suggested that models which link business needs to information are required. These models should be logical (to us OR folk) and credible (to business executives). These models should create and evaluate business based metrics that are influenced by the availability of information. Finally, these models should evaluate projects in ways that encompass the above and are easily related to the conventional time value of money models.

The value of information to reduce uncertainty model may be useful in some decision situations, but does not deal well with most of the transaction systems IT deals with. Gerry said he has been working on a vision-strategy-operations model. The thinking is: improve operations to support the corporate strategy to support the corporate vision. The major difficulty is the decision becomes a "bounding" decision - (would you spend $$$$ to prepare to achieve this rather fuzzy strategic objective?). Another difficulty is that you can evaluate these things on alternate criteria, and many times the probability of success is very far out on the tail of the curve.

 

Big Wave #2: The Right Size for the Firm

Gerry highlighted this topic in relation to our Sunday night presentation, where U. S. West described getting in, and then subsequently getting out of the cable business. Gerry handed out an article by geneticist J. B. S. Haldane entitled On Being the Right Size. It is included in Attachment 5.0 and was written concerning the domain of biology yet has relevance to this big wave in organizations today. The article was written circa 1931. The prologue to the article was written by Galileo in circa 1600. Gerry included a commentary by James B. Newman on Haldane from The World of Mathematics.

Gerry also described an article by Ronald Coase concerning the outsourcing problem. It is not attached, but is found in Economica volume 4, published circa 1937. He described the outsourcing problem by commenting on what affects the make vs. buy decision. The article stated that there is a clear market driver of price and cost when you buy outside the company. Gerry suggested it would be interesting to quantify the Michael Porter value chain in Coase terms The external cost now becomes an internal cost. The costs of coordination are grossly affected by IT. Gerry said that outsourcing is a state of mind. For example, we buy word processors and optimizers without thinking twice. When the operation is retained inside the company, the firm becomes unrealistic about the costs to the company for the product/service.

The Worry

Many questions are involved here:

How do I structure my business?

What’s in and what’s not?

What strategic alliances, strategic partners, or outsourcers should I associate with?

For example, Hewlitt-Packard outsources the final assembly of some of its printers to a trucking company. Change is a constant in this area, with companies divesting an operation, then later recreating it. The focus on core competencies sounds good, but it is very easy to lose control if these questions are not answered and executed well. There are several examples of the information revolution making this possible: code sharing between airlines, outsourcing warehousing and logistics (UPS plays a key role for Gateway computers), and JIT manufacturing.

The Problem

The problem is how can we perform a comprehensive evaluation of the effects of major changes in corporate activities or structures of the firm.

Kinds of Models Needed

New models are not always the best, which is why Gerry cited Haldane, Galileo, Coase, and others. The Porter value chain consists of five functions:

Staff functions clearly underlie the five functions above, such as two of the largest:

Any of the above is a candidate for outsourcing:

Dave Heltne pointed out that many times an outsourcing decision is erroneously made by focusing alone on the benefits and ignoring all the true costs. Gerry commented on this point by describing the situation with Continental Bank, where the networking was outsourced to IBM and the applications development and support to Arthur Andersen. Forty people stayed on the Continental Bank payroll to administer the outsourcing of the original 600-person group. Additionally, another 20 people stayed on the Continental Bank payroll to worry about new technology and initiatives - encouraging what should be done. Gerry agreed that you cannot write a check and walk away.

Big Wave #3: The Knowledge Management Wave

Gerry cited the experience of A. T. Kearny where as we move to a service economy there are "tons of data" available, yet people cannot find what they need. Being stuck with keyword searches can be a problem, because you need to predict how people will want to access the knowledge later. Gerry reported that there is some exciting software coming down the pike that may assist this problem.

The Worry

The transition to a service economy is making knowledge workers more important. Some of the worries include:

Some dilemmas are raised:

The Problem

How can we capture, store, and retrieve the information and knowledge we need to run our enterprise?

Kinds of Models Needed

Gerry suggested some of the following in this area:

4. How Can Operations Researchers Participate?

Gerry stated that the field of dreams strategy will not work. You must actively market your services. As an aid, Gerry reviewed some Marketing 101 concepts as applied to custom built products:

Gerry then provided additional detail about each of the above.

Market Research

Find the big wave, pick one, and identify the business problem.

Preliminary Product Design

The creative process is to discern the order underlying the visible chaos.

Sale

This is what operations researchers do particularly badly. There is a difference between features and benefits. Why does a college boy want a Porshe 911 with a feature that will do zero to sixty miles per hour in three seconds. The benefit is that he will pick up women. Remember, we sell features but we should focus on benefits. What does it look like from the customer’s perspective? Features are characteristics of the product. Features cost money. Benefits are characteristics of the end use of the product. Benefits make money.

Nobody wants an OR model. For a MIP with 20,000 rows and 120,000 columns with 3,800 integer variables, the user sees money flying out the window. If you sell it right, someone might want the output owing to the benefit potential.

So, how do you sell?

Gerry then described characteristics of a successful salesman:

Gerry described one of the worst hiring decisions he ever made. He had hired a bright, talented woman to work at Amoco. Unfortunately, she hated oil companies and she did not succeed in her role there.

Gerry asked can every OR analyst be turned into a salesman. Probably not, but they can clearly be more effective than they currently are. Further, not every analyst needs to be a salesman. You do need a few to keep the projects flowing.

What can the manager of an OR group do to improve this general area of weakness?

One time marketing is usually a waste.

 

Gerry concluded his presentation with the following statement:

If you catch one of the Big Waves we have been talking about - or some other one - a lot of marketing and selling is already done for you.

The biggest single mistake you or your people can make is to be seen to be putting the features of the product ahead of the benefits it will produce.

Don’t do that!

 

OR and Recruiting Session 3: Panel Discussion Graduates to Meet OR/MS Group Needs, facilitated by Dr. James Dyer, Foster Parker Centennial Professor of Management and Finance, the College of Business Administration, the University of Texas, Austin, TX.

Scott Parrish introduced Jim Dyer, and called attention to the recent OR/MS Today article describing the MBA program at the University of Texas.

Dr. Dyer pointed out the ‘techno’ MBA program provides an alternate source of candidates for firms seeking OR/MS candidates. The techno MBA includes an MBA core and adds a strong set of IT and other technology disciplines. There had been a redesign of the Texas program prior to 1992 that had eliminated a core requirement in Management Science. This brought the total technology requirement for an MBA to one hour. The elimination of the Management Science requirement spawned a creative response (we did not see this as a problem, but rather as an opportunity). Industry representatives were brought in to figure out what to do in concert with the academic leaders. At the same time, we noticed the trend to decentralize the corporate IT and MS groups. The MS faculty at the University of Texas looked for a change in direction, and they adopted a new format. The program is described in some detail in the OR/MS Today article.

 

Jim Dyer introduced the panel he had arranged from the University of Texas. He began with Dr. Tom Davenport, who leads the techno MBA program. Tom has a Ph.D. in Sociology from Harvard, business experience in reengineering and knowledge management, and then returned to academia.

Dr. Thomas Davenport

Professor of Computational Finance

University of Texas, Austin, TX.

Tom provided an overview of the current program.

Tom Spencer asked Tom Davenport for a definition of knowledge management. Professor Davenport responded with:

Three years ago eight students were involved with knowledge management. This past year there were fifty-five students. Tom added that some consulting firms have hundreds of employees in this area.

Art Geoffrion pointed out that the great Achilles heel is that we are terrible at integrating OR/MS into functional areas of the university. Tom Davenport commented that we do not have a separate course for OR/MS within the program; instead the concepts are integrated into many of the other courses. Leon Lasdon’s course is available for techno MBA students to take, however, and he will describe it later. Tom added that unfortunately many of the marketing professors do not know much about OR/MS.

Tom Spencer asked if data mining is in the curriculum. Dr. Davenport indicated it was, and that it was co-taught so students can analyze marketing data within the framework of data mining.

 

Dr. Dyer introduced Leon Lasdon, who had been at Case Western until 1977. Leon now teaches MS to MBAs. Leon is an author of one of the optimization solvers in Excel. Jim said that Leon would describe where Ph.D.s are going and what the course is like.

Dr. Leon Lasdon

Professor of Management Science

University of Texas, Austin, TX.

An overview of Dr. Lasdon’s management science course is found in Attachment 6.0. The attachment covers course objectives, instructional methods, course materials, the grading process, and course topics.

Leon provided a few highlights of the course:

George Freestone asked how softer interpersonal skill are reinforced in the course. Dr. Lasdon responded that each session involves several presentations to enhance consultancy and communications skills.

Bruce Nemlich asked how data analysis is taught. Leon answered that it is only covered on the fringe and principally through the use of Excel.

 

Jim Dyer introduced Dr. Anitesh Barua. Anitesh teaches a course with IT emphasis in electronic commerce. It stresses the business value of IT investments. He also teaches an exciting course in advanced data communications.

Dr. Anitesh Barua

Professor of Electronic Commerce

University of Texas, Austin, TX.

Anitesh focused his presentation on the "game course" offered in the area of electronic commerce. Some of the highlights are described below:

Art Geoffrion asked if the products are predetermined. Anitesh answered that there is some predetermined nature of the products that the students ‘sell’.

In summary, Anitesh indicated he is quite pleased by pulling together technical concepts with business concepts.

 

OR and Recruiting Session 4: Facilitated Discussion - Managing the OR/MS Recruits Supply / Demand Gap, facilitated by Dr. Mary Calabro, Engineering Associate, Exxon Research and Engineering, Florham Park, NJ.

Jim Dyer introduced Mary Calabro who, in this final session regarding recruiting, attempted to focus discussion on what gap exists between the demand of OR graduates and the supply that is being presented as potential for hiring.

Dr. Calabro said that her goal for the late session on Monday afternoon was to prepare a list of actions that either the Roundtable or INFORMS could take to address any gap that exists between the demand for operations research candidates and the supply, principally coming from schools.

Informal Summary Survey Results

Mary summarized the results for her informal survey on this subject. The detailed, Excel spreadsheet of the results can be found as Attachment 7.0. Some summary figures from the survey:

Roundtable: What Is the Gap Between the Demand for and Supply of Recruiting Candidates?

Mary then encouraged discussion from the group about describing the gap between supply and demand. She began by suggesting that the gap exists in working independently and in use of a structured methodology for getting things done. Bruce Nemlich said that since Pfizer recruits heavily from consulting companies he sees problems with too much depth in one area, rather than a broad background in candidates. Scott Parrish mentioned that the gap exists in broad based business skills. Frequently, new recruits will be well trained with tools, but lack skills in problem definition and how to form a response. David Heltne agreed, adding that the education is generally satisfactory in recent graduates. What is lacking is the ability to work independently on an unstructured problem. Shane Logue said that since Eastman Chemical Company hires mainly B.S. Industrial Engineers. The most significant gap is in certain tools and techniques (education) and in experience solving business problems. Carey Shaw disagreed, saying that there is no gap and the only difficulty in attracting recruits is that the economy is good and opportunities (and starting salaries) are plentiful. Tom Spencer said that historically he did see a significant gap, but through refining the screening and recruiting activities the gap has been closed for hiring into AT&T Business Operations Analysis. Bill Mann of Raytheon suggested that there is a gap and it is that new recruits do not know how to complete a study plan. For example, for simulation modeling, the new recruits are unfamiliar with determining:

Jim Hobbs of Boeing added that the most significant problem is hiring people who have the potential to acquire business skills, computing skills, and system performance assessment skills. John Lucas of EDS echoed Tom Spencer’s thoughts by saying that through improved screening EDS has not noticed a gap recently. One recruiting technique that has helped is requiring employment candidates to complete a presentation during the company interview visit. In particular, John added, the most significant skill missing is that of listening. George Freestone mentioned that the gap he has seen in recruiting for Air Products and Chemicals is the inability of recent graduates to handle an unstructured problem, an over enthusiasm to apply particular techniques (at the risk of not addressing the problem), an unwillingness to get their hands dirty, and an inability to listen well. Art Geoffrion said that he is actively "producing" PhDs at UCLA and most are not inclined to work in industry. He is eager to have the material from this session made available on the "Student Union" for INFORMS.

Discussion

Gerry Hoffman was surprised that functional business knowledge was not mentioned as a gap. Scott Parrish said that limited domain knowledge is not a problem because he expects people to have to learn that. What he finds more significant is the lack of scientific methodology. Instead, Scott finds, in recent graduates, great technicians who have not been trained to think "outside the box". Jim Hobbs agreed with Scott’s point and added that there is a difference between understanding techniques and understanding modeling.

Dave Heltne took an opposing view by pointing out that there is a finite amount of time that a student spends at university. He said that schools, students, and employers all have a responsibility to make a new hire productive and well trained at a firm. Dave contended that it is much more important for schools to teach non-linear programming than "people skills".

Tom Spencer observed that some things may not be teachable - he gave an example that everyone will not learn to become a golf pro even if they receive individualized instruction from someone of that caliber. In some regards, consulting skills may be like pro golf skills.

Art Geoffrion said that perhaps we should be content to place our Ph.D.s where they can teach, and we should not pretend that MBAs can be great OR practitioners.

Jim Dyer commented that The University of Texas MBA program has taken some recent criticism that it has lost some of its selectivity on people skills.

Art Geoffrion added that there has been a very successful INFORMS Doctoral Colloquium which has had some success. The last four are up on the web. The main orientation of the Colloquium is for students who are headed to academia. Art proposed that possibly what might be successful is to share this material at a Masters level Colloquium on Practice Issues that is being planned for the Montreal meeting. Art suggested the web site - http://www.anderson.ucla.edu/informs/career/ - especially the section on Career Planning. He concluded by saying that he or Donna at UCLA could be conduits for this material to that practitioner oriented colloquium.

Mary concluded this final session on recruiting of the San Antonio meeting by brainstorming what the Roundtable or INFORMS might be able to do to close the gap and by summarizing key observations from this session.

What INFORMS or the Roundtable Can Do to Close the Recruiting Demand / Supply Gap

Summary Observations on Recruiting from This Session

 

 

The attendance list for the San Antonio Roundtable Meeting is included as Attachment 8.0.