2007-06-28 (Thursday) by Gregory Tucker.

It is important that the help desk understand the requirements of the business. This is not as easy as it sounds for several reasons. The business consists of individuals who have different requirements. Additionally, there may be a gap between requirements as the business and the help desk understand them.

As a consequence, it is important to have a regular dialog between help desk and the business, both structured and unstructured. Unstructured communication may include making calls periodically to end-users to verify they are satisfied with the closure of a recent incident. It may include taking a department head out to lunch. Structured communication may include defining Service Level Agreements and providing regular reporting of help desk metrics.

Another critical success factor is defining the scope of services the help desk provides. I learned this from personal experience at Credit Suisse. When I joined my department was defined as “any IT-related work the developers don’t want to perform”, and I failed to redefine infrastructure support more proactively in terms that were meaningful to the business. In retrospect, what I needed to define was a service catalog, but a service catalog is not a prerequisite to running a successful help desk. At the very least, it needs some kind of formal definition to the services it provides (and doesn’t provide) that is communicated with the business.

The successful help desk also defines its processes, because its services must be professional and repeatable. This does not mean the help desk support members are little more than trained monkeys following a script—high levels of knowledge and professionalism are required too. However, it does mean the help desk organization strives for some base level of consistency that can be improved over time.

The final two critical success factors are ones I already mentioned here: reporting and training. Training of the help desk is important. The agents must be educated in the organization, its business requirements, its technologies, and its goals. Technical training is important, but not sufficient. Understanding the needs of the end users is equally important. Reporting is also important to the success of the help desk. Reporting goes beyond the simple generation of some reports. The reporting requirements must be analyzed. In some cases, new data fields or changes to data fields are required in order to create the required reports. The reporting results must be analyzed in order to draw lessons and make improvements. For example, the help desk may see that one department consistently has more incidents than others. This information can be used to recommend improvements that may reduce the number of incidents in the future.

Posted in ITIL, Technology | Print | No Comments »

2007-06-26 (Tuesday) by Gregory Tucker.

I have purchased a term life policy (actually three policies), and I believe in their value, but I feel the industry tends to oversell life insurance.

1) They prefer to sell you whole life, which attaches investment options to the term life policy, because they make A LOT more money on them.

2) They tend to sell you a single policy, even though your need for life insurance diminishes as your dependents get older.

I purchased three policies on myself of the same value: a) 10-year, b) 20-year, and c) 30-year. Hence my coverage diminishes over time, along with my wife’s need for coverage.

Posted in MBA, Family | Print | No Comments »

2007-06-24 (Sunday) by Gregory Tucker.

Bruce Schneier has written an excellent description of the Real ID system.

http://www.schneier.com/crypto-gram-0702.html#2

It is hard for me to understand the continued myth of security through a national ID card. Republican Cong. F. James Sensenbrenner of Wisconsin got the law passed in February 2005 to create a de facto national ID. Republican Presidential candidate Giuliani openly supports the creation of a newer and broader national ID card.

More information on the Real ID Act:

http://www.realnightmare.org/ 

Posted in Security | Print | No Comments »

2007-06-15 (Friday) by Gregory Tucker.

Live Science ran a brief article about the Full Moon phenomena. It is still unclear whether violence increases during a full moon, but the current research suggests there is little to the rumor.

The article makes the observation that our expectations influences our perception. If we expect that everything President Bush does is wrong, then we interpret all political news in a way that confirms our bias, and we ignore news that refutes the bias. Similarly, if a doctor sees a gunshot victim during a full moon, he or she automatically assumes it is because of the full moon.

Many doctors and policemen swear to the full moon hypothesis, even though there is no statistical proof to it. According to Bruce Scheier (here), humans are notoriously bad at calculating statistics. We are drawn to the new and novel, but we become desensitized to the ordinary. The fact that full moons are not ordinary probably makes them stick out when something unusual does occur.

Original article

Posted in MBA | Print | No Comments »

2007-06-09 (Saturday) by Gregory Tucker.

It is hard to distinguish the difference between use or misuse of statistical correlations.

1) There is a desire to try to manipulate the independent variable to control the dependent variable. Many people believe (falsely) that antioxidants cause us to live longer, so they start popping antioxidants pills. This is a misuse of the correlation. Antioxidants are present in fruits and vegetables, which are beneficial for good health and longevity for a variety of reasons.

2) There is the temptation to use the correlation as a predictor. This is what the British resort town appears to be doing. In my opinion, it gets down to the confidence in the correlation. If you could prove there is a 0.0001% chance of this correlation occurring randomly, then I would use the correlation as a predictor. If there is a 10% chance of random occurrence, then I would ignore it. How far back did the resort town analyze the data? How strong is the correlation? What is the probability of random occurrence?

I think that everyone in this class now has enough theoretical background on statistics to ask these questions. (I doubt most of us–myself included–will be cranking out statistical reports when this is all done.) We should not accept the opinion of a single statistician trying to make a name for himself, or a single news outlet trying to sell a headline. Let’s face it: running numbers all days is dull business. If you discovered a correlation, even a weak one, that was novel and capable of getting coverage in the press, wouldn’t you do it? And the press loves to run novel and interesting stories like that, even if the correlation behind it is weak–in part because the journalists have no clue how to analyze the statistical data or question the statistician.

Specifically, I believe there is probably truth to correlation between full moon and behavior. I think it is not purely coincidence that menstrual cycles are equal to the lunar cycle. I am not saying the full moon causes either one–blowing up the moon with ten-thousand nuclear warheads won’t change anything. I am saying that we evolved with the lunar cycle and it is embedded in our DNA.

Posted in MBA | Print | No Comments »

2007-06-09 (Saturday) by Gregory Tucker.

Problems are issues in the infrastructure or applications that cause (or potentially cause) one or more outages, or Incidents. The root causes or Problems can be all over the map and just about anything.

For example, a printer consistently jams the paper. The root cause may be an improper paper type is loaded, or the roller needs to be cleaned every 3 months, or it may be a consistent problem with this model printer. The solution may be training, periodic maintenance, or replacing the printer. Based on the cost of lost productivity, and the costs of the potential resolutions, you can determine how you should proceed.

Sometimes the Problem resolution is merely a firmware update, which is free. This is a relative no-brainer. Sometimes the resolution is upgrading the memory, or expanding the hard disk storage. Sometimes more training is required. Sometimes you decide to tolerate the situation and document it as a known defect–the downside is less than the cost of fixing it.

The main sources of Incidents will vary from one company / environment to the next. That is why you want to track your Incidents and analyze them. There are some rules of thumb. About 60% of all outages are self-inflicted–IT makes a change. That is why a strong Change Management process is required, in order to review, prioritize, authorize, and schedule changes before they occur. Change Management goes hand-in-hand with Release Management, which is automating changes and the monitoring of changes in the production environment. IT engineers will often make changes without authorization, because they think it is quick and dirty and won’t break anything. Most of the time you cannot catch them, unless you have well-developed systems and processes for Release Management.

Posted in ITIL, MBA | Print | No Comments »

2007-06-09 (Saturday) by Gregory Tucker.

The latest issue of Businessweek discusses the plight of Pepsi and Coke in India. Both are experiencing public relations issues surrounding contamination of their products with carcinogens that is somewhat above acceptable levels in the EU.

There is some abuse of the measurements provided by an independent, watch-dog group. The Indian government does not set standards, which is why they quote EU standards. Moreover, the levels of contamination are far below what most ordinary citizens receive drinking tea (or water) from tap water, which is a regular occurrence. Therefore, singling out Pepsi and Coca Cola products appears to be a xenophobic or nationalist reaction.

On the other hand, when you buy bottled water, as opposed to drinking tap water, what are you really buying except the assurance that it is controlled for quality? And if it isn’t controlled for quality, shouldn’t you feel deceived?

Businessweek Story

Posted in MBA | Print | No Comments »

2007-06-09 (Saturday) by Gregory Tucker.

I put up some new photos of the recent visit of Yoshiko to Portland. Yoshiko is an intern at the Japan-America Society of DFW who stayed with us for several months. Her internship is up soon, and she is visiting friends throughout the U.S. before she return to Japan (possibly for good).

Link to photos

Posted in Portland, DFW, Family | Print | No Comments »