All too often, executives make decisions based on the three Hs — history, hunches, and hierarchy. “If it worked before it will work again” is a familiar mantra. So is “Besides, I’m the boss and I got here because I made good decisions.”

Sound familiar? It’s so 20th century — early 20th century.

The way to run circles around your competitors these days, says Steve Sashihara, CEO of Princeton Consultants on Research Way, is to commission optimization models that use data for decision making and maximizing assets. “Businesses have way more data than they can deal with,” he says. “What other technology can you use to make decisions, when data is coming at you at such speed? Do you hide under your bed?”

Amazon uses optimization to suggest new purchases, trucking companies use it to reduce empty miles, and Walmart’s optimized supply chain helps to squash competition. Sashihara cites the example of Kmart’s use of traditional wisdom to sell, for instance, air conditioners. It bought ad space for the Fourth of July a year ahead to get cheap ad rates and set up shipping schedules. In contrast, Walmart kept track of weather patterns and moved the AC units to wherever there was a heat wave. If the heat wave came in June, Kmart was left flat footed with ad space but no stock.

Devising optimization models is the core business for Sashihara’s company. He and his high school buddy, Jon Crumiller, founded Princeton Consultants just after Sashihara graduated from Princeton in 1980. With 80 employees and 20 scheduled to be hired in the next year, they have a 5,000-foot office in Manhattan and are outgrowing their 17,500-foot space at 2 Research Way.

At the 30-year mark, Sashihara has written “The Optimization Edge” (McGraw Hill, 2011, $35) on how to use math and software to allocate resources, beat competition, and increase profits. It uses non-technical language to explain to executives what optimization is, the opportunities and challenges it poses, and why it should be an important item on their agendas. He recounts the history of optimization (starting with operations research for Winston Churchill in World War II), shows case histories (national examples and those from his portfolio), tells how to make it work in a business or even a nonprofit, and then does a futuristic riff on how optimization will pervade our lives.

Optimization is all done with algorithms, an intimidating idea — but think of it as software that doesn’t just present data but also makes explicit recommendations to help you achieve your goals. In this fast-moving business climate, the data changes rapidly. That’s OK, because the optimization software takes in the new data and makes new recommendations. “It quantifies things that have never been quantified before,” says Sashihara. “It slays sacred cows.”

In 2011 this doesn’t sound so amazing, but in the 1980s it was the math/science departments that were teaching the classic optimization then known as operations research, now sometimes called advanced analytics. Even today, Sashihara says, most companies have a “single use model,” which results in everybody coming to the board room to look at a power point presentation. At the end of the day the consultant delivers advice. Sashihara says that “my deliverable is the model, and you can change assumptions to get different answers. It is a more rigorous deliverable; the model tells what to do.”

Some of these new software models can react minute by minute. With optimization models, Princeton Consultants has solved such problems as the scheduling for an airline and layout complexities for a major international newspaper.

Models are expensive, right? Not necessarily. An optimization solution requires people with both software and business skills, but not a whole slew of them. Just two will do, at least in the beginning. Because it’s a creative process, using more people just isn’t effective. “You have to get the whole thing in your head,” says Sashihara.

A two-man team helped a 30-person wholesale flower distributor, the Jan de Wit Company, in Brazil. The firm needed to schedule the correct planting of the right bulbs during the right week in the right greenhouse environment to meet projected market demand — including fluctuations for seasonal holidays. The many constraints included the number of bulbs that can be grown in a pot, their spacing, and the limitations of each greenhouse. In the first year revenue grew by 26 percent and margin by 32 percent, while it had to add only one employee.

Because the head count stays low, you can often squeeze more profit from an optimization project than from the “big spend” IT projects, Sashihara says. The payback can begin even while the model is still under construction, so optimizers flaunt the metric TTP (Time to Payback) rather than the traditional TTD (Time to Delivery).

Take the example that Sashihara calls Argent Air. (He can’t reveal the real name of his client, an airline.) Object: keep the planes full, cut the number of empty miles, and reduce stress on the schedulers. “When our model first started to go online, the important number — fewer empty miles — spiked. Everyone noticed it. When it happened the second and third day, people were completely disbelieving. After about a month, it started to get a lot of notice.”

The pitch is never to reduce the company’s headcount. It is always to use the same people and get better productivity. “You see technologies come and go that are going to change the world but it’s the people who are going to make it happen,” Sashihara says. “We were working FOR the schedulers and we needed their cooperation.”

Optimization often gets confused with strategy, Sashihara says. Strategy is what high level executives use to make major decisions, but optimization works on any level. The client can be a lower level employee, like the layout editor of an international newspaper, which needed a better way to lay out the ads in its various editions. A small number of veterans ran its “playbook,” the limited number of available layouts, with inefficient results. In the worst cases, advertisers could not be guaranteed that their ads would run. In one infamous example, a CEO had planned to begin a keynote talk by opening to the ad in the centerfold of the newspaper — but the layout had gotten screwed up and the ad didn’t run.

Sashihara’s consultants worked side by side with the playbook veterans and about 100 support staff to design software that made sense of the ad placement process. The result saved press time and added revenue, because more ads got into the paper each day.

Human bias can affect crucial decisions, but optimization tirelessly tries all the possibilities without respect to biases. (Daniel Kahneman, emeritus professor at Princeton University’s Woodrow Wilson School, won a Nobel prize for being able to quantify all the human biases that can cloud an asset optimization decision.)

To explain how optimization limits bias, Sashihara points to his client Quad/Graphics. With 28,000 workers it prints Newsweek and hundreds of other magazines. It has myriad problems, from postal regulations to scheduling. Sashihara’s team realized that no single algorithm would suffice, so they set up “an algorithm of algorithms” to bypass bias and help the human schedulers come up with quick solutions.

Universal acceptance of optimization is going too slowly for Sashihara. He attributes these big wins to optimization: McDonald’s vs. Roy Rogers, Walmart vs. Kmart, Marriott vs. Howard Johnson’s, Google vs. Yahoo!, UPS versus Airborne Express, and Amazon vs. Borders.

But Columbia Business School professor Rita McGrath demurs, suggesting that, like many business book authors, Sashihara might be giving his specialty too much credit. “Companies like Google and UPS and others use a LOT of very good practices and employ a LOT of sophisticated techniques across many of their operations, from human resources to product innovation,” says McGrath.

Indeed, there are some cases where optimization simply won’t work, where humans must work alone. An art appraisal, for instance, needs to be based on deep knowledge, where rules are not fully understood. Judgment must prevail in cases where not enough data is available in digital form at the time the decision is made. Humans shine where creativity is needed under rapidly changing circumstances or where the “buy-in” achieved by everyone’s participation is crucial to a good result.

Still, Sashihara emphasizes that even the big guys are buying into it. In 2009 IBM bought the top of the line software, CPlex Optimizer, and opened a 4,000-person Business Analytics and Optimization Division. In 2010 Accenture and SAS announced that they will jointly develop, implement, and manage a unit called Next-Generation Predictive Analytics Solutions. Earlier this year Deloitte bought software-as-a-service (SaaS) player Oco Inc. to enhance its managed analytics capabilities.

And there’s still time to get on board. If you missed the dotcom revolution, optimization “is a revolution that is happening right now,” he says. “I don’t necessarily want to promote it, yet I am frustrated by why it is not embraced more widely,” he says. “I honestly believe there are no alternatives to optimization over the next years. Only optimization has a prayer of making sense out of the tsunami of data that every important business decision is going to rely on.”

One reason it’s not “at the top of mind” might be the lack of a catchy description. It started as operations research in World War II and is often known now as advanced analytics. All four of those words scare off executives who pigeon hole it as “for operations” or “for academics and brainy mathematicians.”

Optimization has its roots in “optimism,” originally defined as “doing the most good at the cost of least evil.” Today, says Sashihara, it implies an active search for the best. “What we are doing is the practical best based on available data,” he says.

An executive with an “I’m the boss” complex will have trouble following the advice of a model. Executives, Sashihara writes, “are not likely to bet their farm or careers on a ‘black box’ or mathematical model they don’t understand, especially when the recommendations that emerge are counterintuitive and not fully explainable.”

“It’s a litmus test about how management views itself,” says Sashihara. “Other consultants might suggest ‘yellow, green, or red light’ but they don’t cross the line and get in your space and tell you what to do.”

It took Sashihara several years to realize the importance of this “people factor.” Mathematicians tend to focus on describing the problem, getting the data, writing the algorithm to fix the process, and presenting the result. “You might never meet the people who actually do the task for a living,” says Sashihara. “You are being hired to optimize the process, not replicate the current one.”

In contrast Princeton Consultants starts by interviewing and observing the “end-user,” the people doing the work now. They are not the only ones to talk directly to the workers, but this is unusual in the optimization world, says Sashihara, who is active in the trade group for this specialty, INFORM.

Now cognizant of the people factor, he looked for employees who could communicate and manage. In the mid-1980s he discovered (lo!) the field of management consulting. Somewhat naively he describes his discovery: “Almost randomly I found the solution: management consultants. I thought, ‘I have found the tribe of people who are fantastic at getting to top management and gaining their support’.”

Rather than hire experienced management consultants with MBA degrees, he continued to choose science, math, or engineering majors, one-third with PhDs, from the Ivies or other selective colleges. But now he also requires emotional intelligence, consisting of likeability, empathy, and sales ability. His project managers need to collect data, not from laboratory animals, but from recalcitrant humans, and they need to be able to sell to executives making 10 or 20 times their salary.

Sashihara turned to a psychologist, Jim Weitzul of Banks & Weitzul. From hundreds of resumes Weitzul tests those who make the initial cut. The requirements are described with the acronym SWAN: Smart, Works hard, Ambitious, Nice.

Nice? That trait is not necessarily valued in business. For a Wall Street trader, says Sashihara, “nice is weakness. For us it is one of the big four. We want someone you would trust your kids with.”

As for his own “niceness,” Sashihara is an elder at Princeton Presbyterian Church. However, except for the firm’s donation policy (it supports a charity rather than give holiday presents) he scrupulously separates his religion from the business. He calls his goal “the life well lived” and attributes his upbeat, optimistic attitudes to his family background.

A memoir written by his grandfather, who came to Los Angeles from Japan speaking no English, reveals a remarkably sanguine reaction to the way Japanese-Americans were treated after Pearl Harbor. An entrepreneur who owned three stores, he was stripped of his assets and sent to internment camp at Heart Mountain in Wyoming, where he was promptly elected mayor. “He didn’t see these things as betrayals, just bumps in the road,” says Sashihara.

Like most in that generation, the grandfather thought of himself as American no matter what. After the war, owning nothing, he took the family to Ohio where he had been told there would less prejudice. He started working in a pharmacy but was not allowed to be a pharmacist because he was not a citizen, and he was not allowed to get his citizenship until he was an old man. “But when we read his memoir there is no bitterness at all,” says his grandson. He was mostly concerned that this decision would put America on the wrong side of history.

Sashihara’s father, born in America, fought in the Korean War and had a life-long job as a chemist at DuPont in Wilmington, Delaware. His mother, also Japanese, grew up in Hawaii and has a master’s in child psychology. “Later we found out she used it on us. The anti-tiger mother, she realized that if she pushed on something, we would push back, and so she pretended she didn’t care.” (He has two brothers; one went to Harvard and works for IBM and one is a deputy attorney general for the state of New Jersey.)

Sashihara and Crumiller met each other at a youth program, Indian Guides, and after graduating from Brandywine High School, Crumiller majored in computer science at the University of Delaware while Sashihara went to Princeton.

Sashihara was a philosophy major because the university had no computer science department. But game theory — as espoused by Institute for Advanced Study fellows John von Neumann and Oscar Morgenstern (founder of Mathematica on Alexander Road), and Princeton University’s John Nash — was “at top of mind” among the undergraduates then. Game theory used math to make behavior models.

“My philosophy professors were interested in it and one of my roommates was an economist and we spent a long time talking about it.” The Nash Equilibrium, the idea that games can result in a good outcome for everyone, is a form of optimization. “Everything that would eliminate waste, for instance, qualifies as a win-win game.”

In 1980 they started their firm, incorporating in 1981. “This was my first job out of school,” says Sashihara. “Lots of people compliment me about being an entrepreneur, but I didn’t have kids or a blue chip job that I had to quit to start something new. We started a year before Jean and I got married. She commuted to Philadelphia to teach at Friends Central and for the first two years I didn’t take any money out. There wasn’t any money to pay us, but it wasn’t like we felt a huge sacrifice.”

“If there had been an Internet, maybe I would have found a job and gone for it, but at that point I wanted to take computers and put them in businesses to flow information and help make decisions. In the early days we had to assemble our own computers. I would go through businesses and wire them, and we would pop in our own chips.”

One of his first clients, Hein Besselaar, had founded G. H. Besselaar Associates in 1976 and built the company into a worldwide leader in outsourced clinical drug development, now known as Covance. Sashihara built all of Besselaar’s first computer network on College Road. “He didn’t like computers early on, saying, ‘You do pharmacology with your heart, your experience, not your damn computer’.”

Princeton Consultants’ current long list of clients includes three of the four largest railroads in the U.S. (Norfolk Southern, CSX, and Union Pacific), Quad/Graphics (the second-largest printer in the Western hemisphere) and a health insurance firm, Ameritas.

One of its nonprofit clients, Boston Rising, aims to narrow the unusually large gap between the haves and have-nots in Boston. It wants to energize the wealthy who have made their money in hedge funds and technology and turn them into philanthropists and volunteers. “Our goal is to have a profound and measurable impact on poverty across the city by using evidence-based strategies to drive results,” according to the website.

Sashihara believes that, instead of sending nominal feedback, such as an annual report or a form letter, Boston Rising could use social media to help potential donors and volunteers focus on a worthy place to put their dollars and time.

“We are saying — what would a full E-commerce platform look like, a kind of Zagat rating for charities,” he suggests. “Let’s say I am a well-intending person, and I want to give $100 or a few weekends of my labor. What am I looking to do with my time? What am I using to make my optimization decision? How can Boston Rising show all the opportunities?”

A charity can come up with an idea that needs a sponsor. It submits its story to Boston Rising. Editors vet the stories and release them to the public. For the next three days the fund (begun with $15 million in seed money) will match dollar for dollar up to $1,000. Donors need not be totally altruistic — they may want to use this as an opportunity to meet other people like themselves.

Princeton Consultants will be able to use the software for other charitable organizations. “I am hoping to make it open source,” says Sashihara. A small college could create a vibrant alumni website to try for 100 percent participation to a giving campaign. Donors could see the goal happen, then communicate with their favorite professors.

Boston Rising is beginning with five charities. “Imagine the success with hundreds of Boston charities,” he says. Getting people into a thankful charitable mode is the real heart of what our country is about, says Sashihara, quoting Alexander de Tocqueville who, in 1835, made the pronouncement that “America is great because she is good. When she stops being good, she will stop being great.”

America can reclaim that goodness, believes Sashihara, the grandson of the mayor of Heart Mountain, by becoming a nation of givers again. “We have become cynical about ourselves, and the answer is for Americans, as individuals, to become philanthropic. But when we donate we want it to be a personal thing.”

Steve Sashihara has no regrets that he has spent his whole career in his business. He sees optimization opportunities everywhere, and this could be the key to his success. Johns Hopkins neuroscientist David J. Linden (New York Times, July 24) defines a successful leader as someone with a physical addiction to it — the feeling of success in improving something, to never be satisfied with the status quo.

“I don’t characterize myself as indomitable but I don’t give up,” he says. “When people have a view that they are in the wrong place, I say, ‘if it is not as good as it should be, you can make the party happen where you are.’ This world view is a gift from my parents and grandparents — a nice positive way of looking at life. “

The root of optimization, after all, is optimism.

Princeton Consultants, 2 Research Way, Princeton Forrestal Center. 609-987-8787; fax, 609-987-0033.

Optimizing Your Life

To be dedicated to optimization is almost a curse, says Steve Sashihara, because everything looks like an optimization opportunity: “From the moment I get up, I am always trying to improve something.”

He lives in New Hope with his wife Jean (maiden name Makita) and their four sons. Jean is a concert harpist with a degree from Westminster Choir College and also a helicopter pilot. Their oldest son, Nathan, works in his father’s firm. One of the middle sons is a Marine and another is at Colorado College. The youngest is at Solebury School.

“In optimization you are usually trying to take a scarce resource and make decisions about it,” he says. On a personal level, the scarce resource is usually either time or money. For instance, Sashihara is always in search of a better way of organizing, so he’s always on the lookout for the next gadget to add to his iPod, iPhone, and Blackberry stable. Among his tips:

Organizing notes: Sashihara veers between the traditional (sketching on white boards and large sheets of paper) and the high tech (the Live Scribe pen, which takes the notes, digitizes, and uplinks them. The in-between solution is to take pencil and paper notes but summarize and store them in Evernote.

Optimizing time: “If you have a computer, it doesn’t matter how much memory is available, it’s how much contiguous memory is available. You are looking for blocks of time, trying to arrange your day so as to get as much time available for the areas where you are most productive.”

Focusing on goals: Optimization calls for figuring out what are the most important things and just focusing on those. “That is not my nature,” admits Sashihara. “I collect lots of random things and I clip out little articles. But if you are a real organizer, you have to strip away the clutter, to be cold blooded and limit yourself to information that is important for decision making.” To counteract his tendency to accumulate mental clutter, he reads from his collection of self-improvement books, specifically the time management shelf, and he even teaches in-house time management classes. “Rather like dieting and exercise, reminders help,” he says.

Optimizing decisions: Even without access to the algorithms, you can use elementary principles of optimization to make decisions. First define the objective function — what you are trying to accomplish. It may seem simple, but then you realize your primary objective has some nuances. Optimizers are good at eliciting all the nuances of what you want to accomplish and then quantifying them.

Sashihara has built personal optimization models for himself, most recently to allocate charitable donations. For example, if he had $600 and had chosen six charities, should he give $100 to each? That’s not optimization. To build an optimized model, you could define such judging criteria as prestige, administrative costs, impact, and relevance. Then judge the charities on each criterion, against each other in pairs. “You try all factors against each other How much better do I feel Charity A is over Charity B on this particular criterion?” Then the model assigns a point count to everything. A personal model could have from 6 to 12 variables.

Says Sashihara: “The exercise helps you quantify, and make explicit, the things buzzing around in your head.”

Facebook Comments