**The principle, graph and Pareto analysis and ABC analysis**

Very often, when analyzing the results, we wonder which causes had the greatest impact on the situation, the problem arose. Methods that can help us to correctly interpret are many, but the most well-known and most commonly used is the Pareto analysis and its variant – ABC analysis.

According to this approach, only (or as much as 20% of factors affecting a given feature) should be in the sphere of our interests.

To understand the specifics of this analysis, it is necessary to familiarize yourself with its three key elements, i.e. the principle, the graph and the interpretation, i.e. the final Pareto analysis.

Now we’ll begin….

**1. The Pareto principle**

**The Pareto principle**, or more **precisely Pareto-Lorenzo**, is the most common principle applicable not only to industry but to virtually every area of activity and even our life. This rule says that** 20% of causes are responsible for 80% of the effects.**

So examples can be multiplied:

1. 80% of defects are caused by 20% of causes

2. 80% of turnover is generated by 20% of clients

3. 80% of complaints come from 20% of dissatisfied customers

4. 80% of sicknesses are provided by 20% of employees

e.t.c.

In this post I will focus on the first example, i.e. 80% of defects are caused by 20% of reasons to continue the quality path of problem solving methods.

In addition, I will show how the Pareto principle is used in the design of full-factor experiments and the analysis of material resources.

**2. Pareto chart**

An indispensable element of the Pareto analysis is the Pareto chart. It is on the basis of his interpretation that a summary of TOP key factors influencing a given situation is made.

To generate a Pareto chart, you can use the Minitab program or simply with MS Excel.

**How to make a Pareto chart?**

*On the example of MS Excel*

- First, prepare the data in the form of a table and rank the results from the highest values to the lowest (column Number of occurrences). Then calculate what percentage of the total is a given defect and% cumulatively.

- Based on the data in the table, a graph should be generated. For this purpose you will need columns described as: Defect, Number of instances and% cumulatively.

After selecting the required columns, select from the Excel options:

**Inserting> Column graph**

You will receive a graph:

- The next step is to change the series with data% cumulatively to the line chart.

To do this, just mark the series of data we want on the chart, and then click the right mouse button to select from the available options Change the type of the serial chart and indicate the line graph.

In addition, you need to select in the options that the line graph is an auxiliary chart. Then the% data will be marked on a separate axis, to the right of the graph.

In this way, we get a Pareto chart:

The axis on the right side of the graph indicates the percentage increase in the share of individual factors, which is reflected in the line graph. On the left, you can read the number of occurrences of a given feature / factor – a bar chart.

All that remains is to interpret the Pareto chart correctly and take effective action on this basis.

**3. Basic interpretation of the Pareto chart**

According to the previously defined definition of the Pareto principle, 80% of the effects cause 20% of causes.

Let’s make Pareto’s analysis from the prepared example.

Guide the line on the graph indicating the 80% value on the minor axis. The point of intersection of this line with a line chart determines the cumulative value corresponding to 20% of the factors that cause the effect.

In the example cited, the main problems are film bends and filings. You could also take scratches under the microscope, because they are on the border of 80%, which may indicate a similar effect as the occurrence of filings.

The elimination of these three factors should significantly improve the quality indicators.

But in order to do this effectively, you need to ask the right questions for the analyzed results, i.e .:

– How was the data collected?

– What causes bending of the film, the appearance of filings or scratches?

– What period of time do the data apply to?

– Which type of defects is the most critical from the point of view of customers?

– Is Pareto stable in time?

**Application of Pareto analysis in the process of designing Full Factorial Designs**

**Designing Experiments (DoE)** is an extremely important method, allowing to understand which factors chosen for the study have the greatest impact on the result we are interested in. The DoE assessment is based on a number of elements, including:

– practical, graphical and quantitative analysis

– main effect and interaction graph

– calculation of the main effect

– **Pareto chart**

In the interpretation of the Pareto chart there is a very interesting assumption that allows you to eliminate the so-called noise, interference in the process, statistically non-significant signals.

According to this approach, **5% of results should be omitted in the analysis, as it is the so-called**** interference, their importance is small and has no effect on the result.** We should not deal with this.

On the Pareto chart this is marked as so-called **water level.** Everything above has an effect on the result.

Now all you have to do is apply the 80/20 rule and make the final analysis.

**Pareto analysis = ABC analysis in the planning of assortment and material inventory**

In the area of material management and product range planning, the Pareto principle is also applicable.

Thanks to it, we can easily link the Pull system, the map of connections and production forecasts.

Often this approach is called ABC analysis. It helps to understand how our production system works. Using the customer’s order schedule and the need for different part numbers, we can generate a Pareto chart and refer to our production or inventory level.

Interpretation of results:

A – 20% of elements including 80% of a given feature

B – 30% of elements including 15% of a given feature

C – 50% of elements including 5% of a given feature

The exact use of this approach will be mentioned in the entries on the pull system.