Sam Savage Professor (Consulting) Management Science and Engineering

Research

* Fields of Specialization:
Embedding analytical techniques in spreadsheets, data bases and the OLAP environment, Risk Minimization in Petroleum Exploration, Stochastic Modeling in Accounting and the Law.

He has an article from Oct 8, 2000 in the SJ Mercury.

Published Sunday, October 8, 2000, in the San Jose Mercury News

 

The Flaw of Averages

If you count on the stock market's average return to support you in retirement, you could wind up penniless

BY SAM SAVAGE

``The only certainty is that nothing is certain.''

So said the Roman scholar Pliny the Elder. And some 2000 years later, it's a safe bet he would still be right. The Information Age, despite its promise, also delivers a dizzying array of technological, economic and political uncertainties. This often results in an error I call the Flaw of Averages, a fallacy as fundamental as the belief that the earth is flat.

The Flaw of Averages states that: Plans based on the assumption that average conditions will occur are usually wrong.

A humorous example involves the statistician who drowned while fording a river that was, on average, only three feet deep.

One of the points Professor Savage makes is

While many of today's managers still cling tenaciously to ``flat earth'' ideals, the innovators are abandoning averages and facing up to uncertainty. Those who dare discover a New World of managerial tools including simulation, decision trees, portfolio theory and real options.

And what happens when one of these innovators is confronted by someone cloaking themselves behind a single number? The story of the emperor's new clothes says it all.

I am constantly amazed how many people hold up PUE as a single number.

Quarterly energy-weighted average PUE:	1.20
Trailing twelve-month energy-weighted avg. PUE:	1.19
Individual facility minimum quarterly PUE:	1.15, Data Center B
Individual facility minimum TTM PUE*:	1.14, Data Center B
Individual facility maximum quarterly PUE:	1.30, Data Center H
Individual facility maximum TTM PUE*:	1.22, Data Center A

* Only facilities with at least twelve months of operation are eligible for Individual Facility TTM PUE reporting

What we need are more graphs showing the range.

Figure 2: Daily average PUE data for a new Google data center currently in bring-up

Part of greening a data center means you have staff who think about optimization. At first glance, this may seem like an analytical skill set, but daydreaming and pushing the edges are need as well. There are some extremely talented people who totally get into the algorithms of how power and cooling systems should operate in an adaptive system. Trade-offs are made all the time as they think about how to save power while keeping acceptable conditions.

But, here is an interesting question is this skill analytical or insight (daydreaming)?

WSJ.com has an article on Insight.

A Wandering Mind Heads Straight Toward Insight

Researchers Map the Anatomy of the Brain's Breakthrough Moments and Reveal the Payoff of Daydreaming

• By ROBERT LEE HOTZ

It happened to Archimedes in the bath. To Descartes it took place in bed while watching flies on his ceiling. And to Newton it occurred in an orchard, when he saw an apple fall. Each had a moment of insight. To Archimedes came a way to calculate density and volume; to Descartes, the idea of coordinate geometry; and to Newton, the law of universal gravity.

Eureka Moments

Five light-bulb moments of understanding that revolutionized science.

In our fables of science and discovery, the crucial role of insight is a cherished theme. To these epiphanies, we owe the concept of alternating electrical current, the discovery of penicillin, and on a less lofty note, the invention of Post-its, ice-cream cones, and Velcro. The burst of mental clarity can be so powerful that, as legend would have it, Archimedes jumped out of his tub and ran naked through the streets, shouting to his startled neighbors: "Eureka! I've got it."

In today's innovation economy, engineers, economists and policy makers are eager to foster creative thinking among knowledge workers. Until recently, these sorts of revelations were too elusive for serious scientific study. Scholars suspect the story of Archimedes isn't even entirely true. Lately, though, researchers have been able to document the brain's behavior during Eureka moments by recording brain-wave patterns and imaging the neural circuits that become active as volunteers struggle to solve anagrams, riddles and other brain teasers.

I liked this article because it reminded me of a team I am working with who get it.

By probing the anatomy of 'aha,' researchers hope for clues to how brain tissue can manufacture a new idea. "Insight is crucial to intellect," Dr. Bhattacharya says.

Taken together, these findings highlight a paradox of mental life. They remind us that much of our creative thought is the product of neurons and nerve chemistry outside our awareness and beyond our direct control.

"We often assume that if we don't notice our thoughts they don't exist," says Dr. Christoff in Vancouver, "When we don't notice them is when we may be thinking most creatively."

Do you have a team who can work in an insight mode as well as analytical?

If your team thinks about models, then there is a high probability they have the capability for insight.

The WSJ has a list of more reading on insight.

Recommended Reading

Daydreaming is more demanding than it seems, researchers reported in "Experience Sampling During fMRI Reveals Default Network and Executive System Contributions to Mind Wandering" in Proceedings of The National Academy of Sciences.

A positive mood makes an insight more likely, Northwestern University researchers reported in "A Brain Mechanism for Facilitation of Insight by Positive Affect" in the March edition of Journal of Cognitive Neuroscience.

In the journal Neuropsychologia, Drexel University scientists reported on "The Origins of Insight in Resting State Brain Activity."

Together, the two research teams reported that people who solved problems through insight had different brain wave patterns than people who don't. In PLoS Biology, they documented "Neural Activity When People Solve Verbal Problems with Insight" and the "Neural Basis of Solving Problems with Insight."

At the University of London's Goldsmith College, researchers reported in the Journal of Cognitive Neuroscience that brain waves heralding an insight can be detected 8 seconds before we become conscious of it.

WSJ has an article about Chrysler’s use of www.organic.com

Modeling Tools Stretch Ad Dollars

Chrysler Uses Digital-Response Data to Adjust Commercials, Drive Web Visits

By EMILY STEEL

With a reduced advertising budget and a desperate need to increase sales, Chrysler is relying more heavily on new technologies to predict how ad purchases will translate into sales.

A team of statisticians, economists, software engineers and media planners at Chrysler's digital marketing agency, Organic, has designed a "media modeling" system that helps the company calculate the best ways to allocate its marketing dollars. The system calculates how much ad spending is needed to meet certain sales targets and then analyzes how both online and offline ads affect Web activity and, ultimately, sales.

Car makers and other companies have used forecasting tools for years, but digital ads have ramped up the systems' sophistication and accelerated reaction time to the data gathered.

Chrysler is using digital-ad agency Organic to try to make the most of its marketing. Organic's technology was used in Chrysler's campaign to promote the new Dodge Ram truck, shown above in a video.

"As a marketer, it helps me be smarter about the dollars I need to reach the sales goals we are responsible for," says Susan Thomson, Chrysler's director of media and events. "It gives you some science."

What does this have to do about data centers is the economic use of dollars to get value. Organic is creating models of how people will  interact with media.

In refining its model, Organic learned how certain ads spur people to visit the Web. It then figured out which Web activities translate into actual auto sales. Some actions, such as scheduling a test drive online or entering a ZIP code to locate a dealer, are a good predictor of sales. Other actions, such as pricing a vehicle or playing with the colorizing features on the site, occur earlier in the shopping process and aren't a direct indicator of serious buyer interest.

The result was a system that predicted 2008 sales within one percentage point of actual sales figures for its Jeep brands, Chrysler says.

And, these ideas are good to think about how models can be created for how users interact with data centers.

Organic modeled the allocation of money

vs. the old way

Mike Manos kicked off his “chiller-side chats” concept on his blog.  And, this is good to see someone bringing up the conflicts in data center design and operations.

With this post, I am kicking off a series of posts in which my sincerest wish is to help all three groups during these stressful times.   Having spent significant time in all three camps I will offer up my own personal take on the issues at hand.  I am calling them Chiller-Side Chats.   From time to time I will post my thoughts on various issues aimed at bridging the communication between these organizations.  I strongly encourage anyone reading these posts to drop comments or offer up suggestions so we can have a lively discussion on these topics.

The three groups Mike refers to is real estate, business users, and data center eng/ops. Mike goes on highlights an issue.

Three worlds have collided and its never pretty.   In my experience and in conversations with many customers in all three categories its a time that fosters frustration, mistrust, and stress.  Its also a wonderful time for less than scrupulous vendors, contractors, and consultants to take advantage of the situation and cause poor decisions to be made.   I am not saying that all consultants are bad or ill intentioned, in fact, there are some phenomenal organizations and products out there.   Its just that you need to be aware of the biases and “religious” debates in this space. 

…

Different firms have different biases and religious affiliations. 

Sometimes the firm that wins the deal is the one who has the best sales team.

Mike is helping to create awareness for a problem I’ve seen for many years in the data center industry. 

And, the good thing I’ve already thought of an answer/method to address this issue.

Modeling.

Modeling enables Trust of a technical solution.

For a trustful and friendly use of technology, the user must be able to have a clear mental model of its use and functioning (way of working), being it partial, superficial and even wrong, but at the same time sufficient for having precise expectations and for knowing how and what to do, i.e. sufficient for reducing uncertainty and perceiving safety and reliability.

So, why model the data center? It increases trust in the data center system including its users. Higher trusts promotes knowledge sharing.

It is clear how trust is a precondition for knowledge sharing and a result of it or, more precisely, that trust is a mediator, a catalyst of the process: it is a mental and interpersonal (cognitive, dispositional, and relational) precise condition for the two crucial steps in the organisational flow of knowledge.

The relationship between trust and knowledge sharing is circular: in order to trust Y, X must either have information about Y, helping him to evaluate Y's trustworthiness, or having knowledge in common with him that encourages the establishment of a trust relationship so as values sharing; on the other hand, in order to share knowledge, it is necessary to have a trust relation or atmosphere.

While caring of making knowledge capital explicit and circulating, an organisation should care of what are the beliefs of the actors about the knowledge itself, about the organisation values, authority, infrastructure, and about each-others, and what they expect and feel on the basis of such beliefs. In knowledge management organisations should monitor and build the right expectations in their members. Knowledge management entails a cognitive, affective, and structural "trust management" in organisations.

The trick is to get the right modeling tools, and this has been a difficult search.  The good thing is I’ve found a partner on this topic, and we’re working on data center modeling solutions.