3-2-12 analyzing systems: Humanization

Humanization is the ultimate skill of analysis, the gift that sooo few of us are capable of doing and maintaining it with long times of dealing with abstract numbers

That is most impotent with systems that affect people. Humanization is the skill of translating the numbers to actual human beings, understanding how these numbers stand out in a world full of humans with feelings and emotions

People with this skill need to be around people all the time, not just setting on front of their laptops … as it is kind of skill that fades with time if you keep out of the customers and people for some time.

This empathy part is where A LOT of the best data analysts miss. It is easier for them to control numbers. The more they understand what these numbers really are in the real world, the more useful the data becomes and the better they can use it to improve the system

3-2-11 analyzing systems: Segmentation

A very useful way of adding context to a measurement is to segment it: take a measurement and split it up into groups

For example: segment customer data by age, social level, location, how long they stay in the store, how they know about us …etc.

Segmentation is how you add meanings to the number by looking into details. It is adding new dimensions to the data to help you understand what is actually changing in detail

The more data you can collect, the more ways of segmentation you can figure out, the more ways you can interpret, understand and get insights in the data

Raw, unorganized and un-segmented data are very low-value data 

3-2-10 analyzing systems: Context

There is no such thing as "The awesomeness factor" or something; the one number that if you look at it you will understand if your business is doing good or bad and what needs to be worked on to improve things …ect.

You actually can come up with one, but the toooo much abstracting you will do to get that one number will actually take out any possibility to learn anything from this number or get any details from it

You have to look at more than one measurement, interrelated pieces of information to understand what these number actually mean and what is going in the right direction and what is not and what needs to be done to improve your business 

3-2-9 analyzing systems: Proxy

Is the method you use when you have some important to measure but you can't measure it directly e.g. Customer Satisfaction

So what you do? You look for things that are directly correlated to this thing and measure them. The more tightly they are connected to the original thing and the ore they are, the more accurate representation you will have

For example; you can measure the customers' satisfaction by measuring the number of returns/refunds, the number of complaint calls etc.

Know that this won't give you a 100% perfect measurement, but if you do it right it will give you pretty good representative numbers 

3-2-8 analyzing systems: Norming

It is "comparing your business now with itself earlier" to see if you are improving or not.

-          It needs a significant amount of historical records

-          For your Norms to be valid, you have to have the same measurements. If the measurements changed, you can't compare

-          One important benefit of this is to compare the same period/season over the years to get a better understanding of the seasonality of your business 

3-2-7 analyzing systems: Ratio

At certain point you have to compare two results (numbers) with the same nature

Ratio is the relationship of two numbers of the same nature. It gives us a sense of direction; where thing are going and a sense of comparison between stuff

Ratio can be a single number, a percentage, a fraction

So when you need to compare, to have a wider look or just want fewer numbers to look at; think of using ratios 

3-2-6 analyzing systems: Counterparty risk

People staring highly-interdependent businesses - where they have to participate with other companies of have partners that they don't have control/influence over – usually are the most ones that will deal with this term

"Counterparty Risk" is the risk of someone you're working with not doing what he is supposed to do & how does that impact the system as a whole

What makes this so important is that you don't have control over him, and people tend not to expect or consider any delay or problems happen with these partners  

You have to keep that in mind all the time, watch and keep track of their progress and assess that risk and keep a B-plan in your pocket just in case

3-2-5 analyzing systems: Sampling

If the system you are trying to analyze is not very simple and small, you will find that it is impossible to collect all the data, so you need to take a sample

Sample is: a subsection of a whole that can be representative of that whole the best way possible; so that you don't need to collect the whole data pool

How to select a sample, and how much sample you need to get, etc….. this is a whole other realm of science labelled "Statistics", "Sampling", "Data warehousing"

One last thing, "Confidence Level" is the certainty of how accurately the sample represents the whole system … and "Principle Of Statistics" let you answer this question and other questions of sampling 

3-2-4 analyzing systems: Garbage in – Garbage out

So you are collecting data to feed into your measurements systems to get results. If you enter crappy data, your measurements will be crappy and you will get crappy results 

When you are collecting data/info; beware of any human error, data entry error, outdated data, process errors, biases, spams and malicious data. Also keep your measurements equations and processes updated all the time with all important factors included 

The quality of your measurements depends on the quality of the data fed into the measurements processes and how good are your measurements' processes

-          Sometimes you will find that you or the people working with you knowing that there is something wrong with the data they are entering of with the measurements but they are ignoring them because correcting the error will make them look worse; for example lowering the number of website visits, cut the sales numbers in half …etc.). A very basic human nature is that we want to look good, and looking worse puts us out of our comfortable zone.

However, know that "Honest Systems" don't give a damn about your feelings! All they care about is getting true numbers and giving you true, objective results. 

3-2-3 analyzing systems: KPIs

"Key Performance Indicators" are the things that you can count that are actually count for something, things that can make great difference if you make use of them

If you can't use a measurement, then it is not a KPI

-          Some measurements are more important than others.

-          There are things that are easy to measure, and things that are important to measure :) .Sometimes they intersect but most of the time they don't

KPIs are the things that help you measure what the system is actually trying to do

You can figure out what your business's KPIs are. KPIs are numbers describing:

-          How much you put in

-          How much you are getting out

-          What is the speed of the process

The previous three numbers or measurements are for the 5 core business processes: value creation, sales, marketing, value delivery and financial stuff

Notice that not all of these 15 measurements are the KPIs of your business. You will find that one or two numbers (maximum 4 numbers) are the real indicators of how you are doing; if they are rising up then you are right, if they are going down then there is something wrong 

3-2-2 analyzing systems: Measurement

To understand anything you must analyze it

To analyze anything you must measure it

A very important part of analyzing a system is to figure out whether if its parts are doing what they are supposed to do or not, and the only way to do this is to look at different parts of the system and asking that question

-          You must keep track of your business all the time, especially of you are starting a new business and don't have a complete analytical system; as it is very easy to ignore stuff

-          This works as a verification of quality and progress … you compare what you are doing against competitors, standard reference … and the only thing you can compare them with is numbers. That's why you need to measure

Notice that:

-          A lot of things that are easy to measure aren't the important things to measure

-          As Albert Einstein said: "Not all the things that counts can be counted, and not everything that can be counted counts"

-          The hardest things to measure is the qualitative stuff, but there are ways to do it

Measurement is a critical part of understanding your business. "What gets measured gets managed" … and if you can't measure things you are then completely blind; you can't figure out if you are doing right 

3-2-1 analyzing systems: Deconstruction

The skill we use when we want to analyze any system is called "Deconstruction":

Taking a complex system and break it down to its simpler systems, studying them as if they stand alone

-          Any system consists of smaller simpler systems that are connected to each other

-          You take the small, easy-to-understand part and:

-          Examine it

-          Know its inflows and outflows

-          Know the connection between it and other parts

-          Know its order in the process

-          Whether it is a slack, constraint or not … etc.

You keep doing that with any system you want to analyze till you are finished with analyzing all its parts, and then you make a diagram/ flowchart of the system to get a whole eagle-eye look at it