Saturday, September 18, 2021

Taking Operational Efficiency to the Next Level: Leverage the 95-5 Rule of Automation

Through an odyssey of over two decades helping clients in various industries solve hard problems, I have gained a deep appreciation of a pattern that can be leveraged to dramatically improve the quality and efficiency of the work and, ultimately, the return on investment of businesses.

Whether it is inventory planning, or financial fraudulent transaction detection, or finding costly insurance claims, it generally holds true that 95% of the work can be resolved by automated algorithms. The remaining 5% needs to be done by domain experts using their expertise, intuition, and creativity. I call it the 95-5 rule of automation.

The 95-5 rule is not simply a division of the labor between machines and human experts flatly in that proportion. There is a structural and temporal implication in it. Algorithms are first applied to a raw problem, which involves a large number of cases and big data and is hard or inefficient to solve manually. This step produces as an output a simpler problem where the work is greatly reduced, by 95% generally. Human experts then work on this reduced problem and make their judgment calls to reach the final decision.

Take as an example our solution to a worker compensation insurance claim problem. A company receives about 200 worker injury claims daily. Our algorithm highlights 10 (5% of the total) of them as potentially costly using a machine learning model based on factors including age, cause of injury, and injury body parts. Using these 10 cases as a starting point, analysts review them carefully and take proper action. The solution has resulted in a 40% reduction in claim loss.

To recap, in the real world the 95-5 rule of automation works this way: applying algorithms to a raw problem to reduce the work by 95% and subsequently having human experts take on the reduced problem.

Here are the benefits as reported by our clients that have adopted solutions based on the 95-5 rule of automation.

  • Improved outcomes. For example, a bank sees its fraud loss reduced by 70%. Another bank finds the bad debt rate dropping by 50%.
  • Increased efficiency. In a K12 education company, content tagging is 100 times more efficient than a manual process.
  • More jobs. A group in a bank hires more analysts because the operation there drives a good return on investment.
  • Improved employees' morale. This is because they work on the reduced problem where the same amount of effort generates more fruitful outcomes. ( I did not realize this point until I saw a report produced by an independent department from a client company.)

When the rule is applied to inventory planning, our advanced optimization algorithm generates a set of recommended safety stocks for all items which serves as the foundation for planners to make further improvements.

One lesson that we have learned is that, unless it is an exceptionally simple circumstance, domain experts should not work with the raw problem directly. Unfortunately, the violation of this principle is happening every day resulting in ineffective, inefficient, and unscalable operations and a stressful workforce. The whole situation is avoidable.

The 95-5 rule of automation has worked for us remarkably. I hope you make the most of it in your organization and take operational efficiency to the next level.

Saturday, August 14, 2021

Webinar Video: Holistic Safety Stock Optimization - Putting the Horse Before the Cart

The recorded video of the webinar "Holistic Safety Stock Optimization - Putting the Horse Before the Cart" delivered by Dr. Jay Zhou on August 12.



A significant amount of money can be saved without impacting the overall demand satisfaction by optimizing the inventory holistically. With the right approach, we can comfortably construct an inventory with less money while providing higher overall demand satisfaction.

Friday, August 06, 2021

Free Online Course: Oracle SQL for Random Sampling

In this course, competition winning data scientist and long time Oracle SQL practitioner Dr. Jay Zhou will share his expertise about performing random sampling using Oracle SQL. Students will learn practical skills that can be applied immediately in tasks that involve random sampling. These include: how to quickly view random samples of the data, how to generate exact number of records randomly, how to select random samples by groups, etc.

  • What you’ll learn? How to perform random sampling using Oracle SQL.
  • Are there any course requirements or prerequisites? Very basic Oracle SQL knowledge.
  • Who this course is for? SQL developers, data analysts, data scientists, statistician.
Please take the free course here.

Wednesday, August 04, 2021

Webinar Invitation: Holistic Inventory Optimization

If you are responsible for a large inventory, I would like to invite you to attend my webinar "Holistic Safety Stock Optimization - Putting the Horse Before the Cart" on Thursday at 2:30 pm, August 12, 2021. Please register here..

I have written an article ) describing the challenge facing companies with large inventory. Tens or hundreds of millions of dollars or more, in addition to their stock prices or market valuation, are at stake depending on the company’s inventory sizes. C-level executives and senior managers want answers to the following critical questions which are both sides of the same coin.

  • How to spend a fixed budget on safety stocks so that the average demand satisfaction for the whole inventory is the highest?
  • How to achieve the desired average demand satisfaction for the whole inventory with the lowest budget of total safety stocks?
In the webinar, I will demonstrate a powerful solution developed by Friesian Analytics. The total inventory value can be reduced significantly without affecting the overall demand satisfaction. The horse (budget) should pull the cart (inventory), not the other way around as it is done by many companies. I look forward to seeing you at the webinar! Thank you.

Monday, July 26, 2021

Holistic Inventory Optimization

The following was first published by me as a LinkedIn article.

Tens of millions or hundreds of millions of dollars or more, in addition to stock prices or market valuation, are at stake depending on the company’s inventory sizes. C-level executives want answers to the following critical questions which are both sides of the same coin.

  • How to spend a fixed budget on safety stocks so that the average demand satisfaction for the whole inventory is the highest?
  • How to achieve the desired average demand satisfaction for the whole inventory with the lowest budget of total safety stocks?
To illustrate the point, consider the following simple inventory with 10 stock items. Factors determining the total value of safety stocks including lead time, unit price, demand standard deviation, and demand satisfaction (DS). Among these factors, DS for the individual stock items is the only one that we can adjust readily.

Figure 1. Common Approach - Setting the Same DS to Each Item

A common approach is to simply set the DS for all parts to a single number. Some managers adopt ABC analysis to divide stock items into several groups based on their values and assign the DS to each group. When we set DS to 91% for all items as shown in Figure 1, the average DS is 91% and the total value of safety stock is $129,985.

However, we can do much better. By setting the DS for each item more intelligently, we can keep the average DS of 91% and reduce the total safety stock by about 16% (Figure 2). Or by assigning a different set of DSs, we can maintain the same total safety stock budget and increase the average DS from 91% to 93% (Figure 3).

Figure 2. How to Keep the Average DS of 91% and Reduce the Total Safety Stock by About 16%?

Figure 3. How to Keep the Same Total Safety Stock Budget and Increase the Average DS From 91% to 93%?

Interested readers may do an exercise. Here is a downloadable spreadsheet file that embeds the formula to calculate DS. One simply fills in whatever DS she/he desires for each item. The average DS and total safety stock will be calculated for you.

If you can produce the results as shown in Figures 2 and 3 or better, congratulations! If not, I would invite you to attend my webinar "Holistic Safety Stock Optimization - Putting the Horse Before the Cart" on Thursday, August 12, 2021

Even if you are able to solve the 10-item problem by trial-and-error, in the real world many businesses have far more stock items than 10, I would still encourage you to attend the event.

Thursday, July 22, 2021

Upcoming Webinar: Holistic Safety Stock Optimization - Putting the Horse Before the Cart

I will give a 45-minute webinar on Thursday, August 12, 2021, at 2:30 PM Eastern Time. If you are responsible for a large quantity of inventory, I would encourage you to attend the event. Please register here.

The following is a summary of the webinar.

The CFO of a company wants to know, given a fixed budget, how to determine the safety stocks of all inventory items to achieve the highest overall demand satisfaction. In the webinar, Dr. Jay Zhou will show that the above goal can only be achieved by optimizing safety stocks holistically. He will demonstrate a powerful solution developed by Friesian Analytics (https://friesiancorp.com). The horse (budget) should pull the cart (inventory), not the other way around as it is done by many companies.

Friday, July 02, 2021

Free Online Calculating Inventory Safety Stock Calculator

I made a youtube video at https://lnkd.in/d3hECWU to introduce a free online tool for calculating inventory safety stock. A spreadsheet file that implements the calculator is downloadable. Here is the link to the calculator. Enjoy!

Friday, May 28, 2021

Oracle Function Returns Two Values

There is a table in a schema that contains three columns, p, low and hi. In the table, p is the primary key. I want to develop a function to return low and hi based on an input variable p. First I create a type.

create or replace type t_low_hi as object ( low number, hi number);  

Then I create a function that finds low and hi based on p, constructs a type object and returns it.

create or replace function f_prob (p_p number)  
return t_low_hi is  
p_Low number;  
p_Hi number;  
Str_sql varchar2(2000);  
begin  
Str_sql := 'Select low, hi from t_lookup where p=:1';  
Execute immediate str_sql into p_low, p_hi using p_p;  
return t_low_hi(p_low, p_hi);  
end;  
/ 

I call the function and retrieve low and hi for p with a value of 0.99.

select x.v.low , x.v.hi from (select f_prob(0.99) v from dual) x; 

Tuesday, January 12, 2021

Updated Online Chinese Document Analytics Tool

We have updated our free online tool for analyzing Chinese documents: https://aistrike.us/text-analysis.html A user fills in a textbox with the content and click Submit button. The tool identifies words, displays a word cloud picture and calculates a sentiment index for each sentence. Enjoy!

Identifying words in a sentence is necessary. Chinese words in a sentence are next to each other without spaces separating them, e.g. Chinesewordsinasentencearenexttoeachotherwithoutspacesseparatingthem. And yes, the division of words could be ambiguous. For example, "结婚和尚未结婚的" could mean "married and unmarried" ("结婚 | 和 | 尚未结婚的") or "married, monk, unmarried" ("结婚 | 和尚 | 未结婚的") .