Metrix-Blog

I remember being told a story that I really loved. Here is the version that I was told (best as I can remember).

In the days of the Soviet Union a People’s Commissar of the Commissariat for Education was doing his duty visiting schools testing the quality of the teachers by quizzing the students. Because he was the commissar he felt it was his duty(and necessity to justify his job, and negating the need to pay the teacher a stipend, because of incompetence) to prove that the education was lacking and the students double their efforts, by asking questions of increasing difficulty until every student had failed to answer a question. This strategy had served the commissar well, until one day he visited a rural school, and to his dismay one nine year old peasant boy correctly answered all of the commissars’ questions. Finally to shut down the trouble make the commissar asked the boy his trump question: “How many hairs are there on a dog?” he asked the boy, to the amazement of the commissar and the teacher, the boy replied without hesitation “3,571,962”. “How do you know this to be the correct answer?” asked the astonished commissar, to which the boy replied “If do not believe me count them yourself!” The teacher was worried, but the Commissar Broke into hearty laughter, complemented the teacher for his skill (Awarded him his stipend) and the boy for his cleverness, and vowed that he would tell the story to his colleagues at the Commissariat when he returned to Moscow, as they would really enjoy it.

When the Commissar returned the following year for his annual visit, the teacher nervously asked him how his colleagues at Commissariat responded to the story. Disappointedly he replied, “I wanted very much to tell the story, but I couldn’t. For the life of me, I couldn’t remember how many hairs the boy had said the dog had!”

I loved this story because; it spoke to me of many of the issues prevalent in measurement programs:

• Ulterior (insincere) motives to questions
• Meaningless answers to meaningless questions
• Ridiculous precision to measures
• Evil Autocracy
• Doubt that the number could be real
• Fabrication of Numbers (I am assuming the boy didn’t actually Count!)
• Mistaking the number for the message

This story is like the Swiss Army Knife of metrics stories, use it for whatever punch line you want. However, to my chagrin, I subsequently discovered the origin of the story, and …. I loved it even more! What I had heard was a version of a story (http://www.nwlink.com/~donclark/hrd/history/gestalt.html ) used by Max Wertheimer, founder of Gestalt psychology to explain the difference between learning based Gestalt principles, i.e. based upon understanding the underlying principles of the problem, and alternatives where facts are learnt without understanding them.

I think I just had a Gestalt Metrics moment, Have you? Can we be effective in a metrics program by learning hand applying the facts without understanding, or do we need to understand the underlying principles?

The World of Metrics and Decision Making has many stakeholders. Let’s see how the interesting game of Metrics is played within Organizations

Business Decision Makers – They need metrics at the speed at which they make decisions. An informed decision is so much more near reality and comforting than a gut-feel based decision.
Corporate IT – Team mandated to manage and provide for all electronic information needs of the corporation
DW/DBA’s – Database managers and teams who are mandated with the running of all electronic databases. The only teams to have knowledge of the data
Application Vendors – Providers of Applications which optimize Corporate time and capture valuable Data

Let’s look at how the desired Information finally reaches the Manager who needs it most
1. Manager – states the numbers required to make certain decisions (Markets, Customers, Quality, Delivery, WebAnalytics etc) GOTO IT
2. IT – defines the needs and sets up resources and the Project Plan. Databases and Queries are defined – GOTO DBA
3. Data is extracted from the Databases or Data warehouses and then cut and paste into some normalized formats – GOTO MS-XL
4. XL formats are sliced to present the respective Information in MS PowerPoint or MS XL – GOTO Managers

Now if a Manager wants a small change in any Information (A Marketing Manager requesting that the Sales per Region is to be changed to Sales per City) it’s a GOTO step1
If a Manager wants Information for another time period (A Delivery Manager looking at Delivery Quality or Resource Count of Last year – same period) it’s a GOTO Step1

Organizations are continuously debating what are the best possible processes to get the right Information on their desks when it is required most. The entire debate is around the very fundamental issue– IT wants the Business Managers to confirm to a specification of their information needs that they can deliver and maintain. Business Managers know that whatever Information needs they specify, will probably change by the time they hit the send buttons of their email boxes. Hence the gap in understanding of the Need and what is Delivered or should we call it perfectly “Lost in Translation”

Here is something anecdotally learnt from the field of Organizations and Metrics

Approximately 95% of Corporate Managers store valuable information in MS-XL and MS-PowerPoint
Over 80% of Organizations dealing in daily collection of data prefer XL sheets to any other automated solutions
About 5 % of the Organizations have a central Metrics Information office and  deicated teams – its normally to each his own

Interesting!! One can draw a parallel here with how we think. Good health is all about controlling weight, calories, body mass, Cholesterol, Blood Sugar etc. etc. Hey!!! where is the brain??? Do you need it to keep us healthy at all (smile)

I was asked the other day how managing was becoming more challenging in the current enconomic envioronment. I think this is a little bit of a tautology; it’s like being asked if something is “dead or deader”. Management has always been challenging, and in the IT industry a lot of the things that make it challenging are invariable. Having said that, the difference now in the current economic environment is the tolerance for failure and the expected time-line for success are very different. In more affluent times we could apply the very successful empirical approach (which served the Roman Empire very well for over 400 years) i.e.

1. Build a bridge
2. If it stands up, repeat this design next time
3. If it falls down try a different design

The empirical method has two characteristics: Catastrophic failure, and over engineered solutions. Right now we certainly cannot tolerate failure or over engineering. This brought us to the next part of the question. “How important is data to the role of managing?”.

Obviously a key improvement in the empirical method is to measure. Even a simply tally of success can be an important step to management improvement, (then for example engage the project manager with the largest ratio of success to failure). Data gives us a way of representing the material environment quantitatively, this allows us to develop and change conceptual models of the subject. Once we have such a model it can be manipulated and experimented with intellectually, and more importantly optimized. We are able to accelerate the development of the empirical model, without the material losses. This does sound very theoretical, and 800 year ago it was, this was when we began to make the transition to a universal measurement model. Prior to this transition, numbers were important but were simply statements of quantity, the idea of developing conceptual models of reality and then using these to base our strategies was new. The fundamental differences between the two models is best illustrated side by side (top and Bottom) comparison. The Venerable model is characterized by:

Symbolism

Numbers were chosen because they had mystical meaning

Mysticism

Things can only be known if “the Gods” deigned to let it be known. Predictability derives not from the reality itself  but from “the Gods

Heterogeneity

Whatever the circumstance today would not preclude other circumstances yesterday or tomorrow.

Destiny

It’s not possible to understand therefore we cannot control

The Pantometric model is characterized by:

Realism

Numbers are chosen because they best characterize a specific characteristic

Determinism

Predictability derives from reality itself

Homogeneity

The circumstances today are similar and comparable to circumstances yesterday or tomorrow

Destiny

 We can control those things we understand

It can be seen that under the old model, data was not very important, because things were what they were, and we had very little control over it. Even the most skilled artisans were the playthings of the gods. If a bridge feel down it was more likely a whim of the gods rather than a fault of the design. If it stood it was because it pleased the gods. In the new model the data is most important because not only does it explain what happened it provides us with a crystal ball to the future, and allows us to predict what is likely to happen, and take actions to drive to the result we want to achieve.

It is clear that a fundamental belief in Realism, Determinism and Homogeneity and the ability to control our own destiny, is essential common ground for establishing a successful measurement program, and our ability to manage effectively.  However, in my experience it is rare that these four pillars of success are fully present in any corporate measurement program. Take a minute to consider your own experience. How many of you believe for instance that the current economic situation was predicatable? that the future performance of the stock market is in our own hands?  Do you believe that it predicatable? What about your corporate performance mangement program? – Did your review last year accurately reflect your performance?  - is it driving you to better performance this year?  Would most people in your organization say that project data from 5 years ago can be used to predict the your current programs?

Managing becomes more challenging  not because of the constraints of  the economy, but when you take issue with the fundamental properties of the Pantometric model – because  in doing so by definition,  every situation becomes unique, every decision has an unknowable outcome, and you are no longer in control.

It is more insidious than this however. Management can keep the venerable model alive and well through its own actions. Consider the effect of time tracking systems that require 40 hours of work to be logged irrespective. Consider the impact of arbitrarily imposed schedules. Consider the impact of poorly executed incentive systems.

When I review the state of many corporate measurement programs I have to ask myself time and time again “Do we really believe in Measurement?”

Hmmm!! Someone asked me this very interesting question in what could be called a “Gotcha!!” moment in an argument. It started a thought process which was interesting to start with.
The very common myth is that common sense has some dotted line direct link to IQ. In school when a student gets an 8 out of 10, he/she is graded an A or an 80%. It is a clear measure on how much the student can think or how much of the subject matter is known to the student (Or how much he/she could cram up the night before). Did it reflect on any valuation of common sense?? Strangely, there are many who do not get the A grades but solve real life problems much better than the GradeA student . Was it that they had more Common Sense or IQ? The net conclusion of the argument was that mayb IQ is what can be measured and quantified and Common Sense is what cannot be quantified but is knowledge everyone should have to carry out their lives. Example is we all know day comes after night, there are 24 hours in a day, there are 12 months, you need money to exist, you need to eat to live etc. etc. This is common knowledge which takes you through life and are the senselet’s which one gains over time. Knowledge of algebra and geometry is not required to live life and hence it is not common sense.

Now here is the more interesting part!! On digging deeper into all pieces of information which probably constitute Common Sense, it suddenly came up that they are nothing but commonplace measurements. 24 hrs in a day is a measurement known to everyone. Ask a person on the North Pole who cares less about day and night or the number of hours in a day. What is his/her common sense knowledge then?? So now isn’t Common Sense a set of standard measurements we live our lives out of. Are these Senselets (measurements) contextual to what one does and where you are?
Here is the killer one!! Lets extend the hypothesis that Organizations are no different that Organisms (Human Beings) co-existing in an business eco system. So the common sense of the eco system is all about goods and services being manufactured as per the skills of the organization and being distributed to multiple customers for the goods. But what is the Common Sense of an Organization??? Let’s say a new Manager walks into an Organization taking up a responsibility. What “common sense” of the organization should he/she know? Or is it the Gut Feel and Experience which is all that is required? What information is there in an Organization to tell its employees and constituents of what are the standards, benchmarks, optimum levels, goals etc and how historically they have become a culture and common sense. An interesting argument one can make is that 20% of the Organization deals with increasing its IQ ( the top lines, the growths, the optimizations)and 80% run on the Organization’s Common Sense(business as usual). The big question always is, what is this all-pervasive Common Sense of an Organization??? Can it be quantified like any other Common Sense??

When you have mastered the numbers, you will in fact no longer be reading numbers, any more than you read words when reading a book. You will be reading meanings.

 - Harold Geneen, Legendary President of ITT and Founder, MCI Communications