Eliminating the 8 LSS Wastes (Muda) in Your Digital Teams | by Will Keefe | Sep, 2022

Lean Six Sigma practices are industry-standard in manufacturing, but the same principles are readily transferrable and transformational when applied to your digital teams in software development and production

Lean Six Sigma and related total quality management systems have been around for decades, and have had multiple predecessors stemming back from W. Edwards Deming and succeeding to specific quality systems in individual sectors such as Agile Project Management in tech.

While the clear emphasis on measurability and functional working technologies is easily apparent in Agile, some of the wastes recognized in LSS may not be as apparent, but still deserve significant attention in our development teams and production environments. In this article, we will break down each of the eight LSS wastes, also known as Muda and easily recalled with the acronym “TIMWOODS”, and how they can be readily applied to our work digitally.

T — Transportation

This waste originally described the movement of people, machinery, and other resources across factories and other manufacturing environments. Why have your conveyor belts send widgets across the entire factory between unit operations if it is possible to have the operations right next to each other? After all, it reduces risk, uses less power, and saves time.

This concept is readily transferrable to the movement of data too. For example, if you have a database in the terabytes of customer data and are searching for a dataset to model the habits of one town, SQL queries selecting everything will use much more time, effort, and computation resources than specific, manageable, streamlined queries with keys.

-- Do not do (may take hours to run):
Select * 
From CustomerData-- Instead do (will run much faster):
Select CustomerID, CustomerAddress, CustomerTown, LastOrderDate 
From CustomerData
Where CustomerTown = 'Nashville'
And LastOrderDate >= date('2022-01-01')

I — Inventory

Inventory management problems can traditionally borderline hoarding, and having a “just-in-case” reserve while necessary in some industries, can lead to a crisis in others when just-in-time inventory management is ideal in most cases. Most people will remember bare supermarket shelves, specifically shelves with toilet paper, at the start of the pandemic, when COVID didn’t even justify such a drastic increase in toilet paper need. Similar problems in inventory management throughout the pandemic caused shortages in silicon chips, pharmaceutical equipment, and other materials and parts we today may take for granted. Real estate and storage space are finite, just like supermarket shelving.

Digital space is also finite, even though sometimes the costs of maintaining data can seem so small in the micro incident. However, at scale, data and storage limits can have a significant impact. The Y2K bug caused a serious fear in banking and technology as a whole (despite not having an egregious impact), and similar data dimension problems can happen all of the time. Integer overflow problems, such as with the Ariane 5 rocket, can cause hundreds of millions of dollars of damage from a seemingly simple arithmetic error. On the flip scale, wasting inventory, such as allowing for fake social media accounts to exist on platforms owned by Meta and Twitter, can cost individual companies millions in maintenance fees and storage of fake photos and posts. Proper digital inventory management needs to be addressed at the micro level in our teams so that incidents don’t have a macro impact.

M — Motion

Wastes in motion are common place in everyday life, let alone in either the manufacturing space or in our digital teams. Having a tool on the top shelf, requiring a step stool to reach is a waste in motion because extra steps are needed than should be. In manufacturing, perhaps the equipment needed to add further value is on the other side of the room instead of adjacent to your station. In technology, similar wastes in motion are applicable with data. Overly complicated user interfaces or entries in data can cause wasted time. Even minutae such as “confirmations” or manual checks can be redundant. In an ideal program environment, only the right data can be entered. An example is allowing alphabetical keyboard strokes in a program looking for a phone number. This waste causes a poor experience for everyone involved.

W — Waiting

Arguably the most aggrevating waste on the list is waiting — recall how many times just this week we have sat in a line, in traffic, waited for someone to get us something we needed, etc. Waiting in manufacturing is an unnecessary expense to the value of product, and waiting in the tech sector can cause poor developer flow, poor backend timing, and poor user experience. Ultimately many cases of waiting digitally can be avoided by changing capacity. Perhaps there is a server problem where there are either: A) not enough servers or B) the servers are not fast enough. Keeping our code bases clean and efficient can help reduce this problem, as well as building functional code that uses modern parallelization techniques and vectorization wherever possible. We want our software to run smoothly and not exceed the capacity of our hardware.

O — Overproduction

Overproduction on a physical level is exactly what we would expect — manufacturing too many items than we have a need for. Digitally, overproduction can related to having too much service. Yes, too much service. Too much service is an unnecessary expense because the service available is not being adequately utilized. Perhaps this is using desktop computer to perform the task of a cheap microcontroller, or having multiple dedicated servers perform the same work a serverless architecture could perform. Ultimately, this is a conversation on costs and the amount of work that needs to be performed. Our code can also be overproduced with using improperly sized variables than necessary and being inefficient in our data structures, leading to similar costs on too great of digital inventory.

O — Overprocessing

Overprocessed manufactured goods provide more value than necessary to a customer. Some great examples of this are using rare-expensive metals for parts when cheaper ones would suffice, adding more layers of paint to a car than noticeable for people, and designing TV’s or other screens with higher framerates than are even detectable by our eyes. If nobody will notice the “added value”, how is it even adding further value?

Similarly in technology, our code can be overprocessed and Agile does a great job combatting scope creep, through structured and measured cadence with sprints. Managing software projects focused on “minimum viable products” and well defined essential features based on user feedback and requirements is the best way to combat overproduction digitally. Why add a feature to product, spending developer time, effort, and even pay for something not asked for or desired? While it is true that some features can be pleasant surprises that enhance a user’s experience, some users may never notice. Overproduction digitally is a waste of a team’s time and energy, and that time and energy should be better utilized elsewhere.

D — Defects

Defects are measured differently in different industries and even for different products in the same industry. A defect in printer paper, may not be noticeable to most consumers unless it is physically nearly destroyed. Slight changes in the appearance may not be detected. However, a defect in litmus paper used in the laboratory setting may totally ruin an experiment, which in turn can affect their reputation. Typically, the Six Sigma portion of Lean Six Sigma gets goals for defect reduction, the sigma being the variance of a statistical collection. “Three Sigmas” covers 93.32% of a population, four covers 99.35%, five covers 99.98%, and six reaches 99.9997%. Within the binary classification of defects, with a six sigma process 99.9997% of the units produced should be defect free, or around three defects per one million units produced. This is a hefty goal to reach, but a six sigma process is extremely reliable and consistent, and could definitely have the trust of a customer.

Software can have defects too. Think about how many times an app has crashed on a smart phone, or how many times a “bug” in a game or program prevented you from moving forward in a task. At its most insignificant, these bugs can have insequential, or even humorous consequences. At its most serious however, software defects can be life threatening. There are plenty of incidences where even one small bug uncontrolled and unpredicted cost millions of damages and threatened peoples’ livelihood. Ultimately there will be a tradeoff on every software team of anticipating bugs and getting a project done on time — however the goals are clear. The best code may only fail 3.4/1,000,000 runs.

S — Skills (The 8th Waste)

The newest identified waste is based on skills of employee. At first thought it may seem that the waste is a risk in not having the right skills, however the larger risk is not recognizing a workforce as a company’s greatest asset. This waste, or not enabling, recognizing, and supporting employees is the greatest waste. Especially in the tech sector.

With each new graduation year, a class and generation of new excited employees enter the workforce with new skills and backgrounds, hungry for experience. Micromanaging and a lack of creativity and trust in employees can destroy a company. Encouraging growth and ownership in the newest employees will build loyalty — and recognizing and rewarding newly developed skills and experiencing will have a domino effect on a team’s success.

Lean Six Sigma practices are industry-standard in manufacturing, but the same principles are readily transferrable and transformational when applied to your digital teams in software development and production

Lean Six Sigma and related total quality management systems have been around for decades, and have had multiple predecessors stemming back from W. Edwards Deming and succeeding to specific quality systems in individual sectors such as Agile Project Management in tech.

While the clear emphasis on measurability and functional working technologies is easily apparent in Agile, some of the wastes recognized in LSS may not be as apparent, but still deserve significant attention in our development teams and production environments. In this article, we will break down each of the eight LSS wastes, also known as Muda and easily recalled with the acronym “TIMWOODS”, and how they can be readily applied to our work digitally.

T — Transportation

This waste originally described the movement of people, machinery, and other resources across factories and other manufacturing environments. Why have your conveyor belts send widgets across the entire factory between unit operations if it is possible to have the operations right next to each other? After all, it reduces risk, uses less power, and saves time.

This concept is readily transferrable to the movement of data too. For example, if you have a database in the terabytes of customer data and are searching for a dataset to model the habits of one town, SQL queries selecting everything will use much more time, effort, and computation resources than specific, manageable, streamlined queries with keys.

-- Do not do (may take hours to run):
Select * 
From CustomerData-- Instead do (will run much faster):
Select CustomerID, CustomerAddress, CustomerTown, LastOrderDate 
From CustomerData
Where CustomerTown = 'Nashville'
And LastOrderDate >= date('2022-01-01')

I — Inventory

Inventory management problems can traditionally borderline hoarding, and having a “just-in-case” reserve while necessary in some industries, can lead to a crisis in others when just-in-time inventory management is ideal in most cases. Most people will remember bare supermarket shelves, specifically shelves with toilet paper, at the start of the pandemic, when COVID didn’t even justify such a drastic increase in toilet paper need. Similar problems in inventory management throughout the pandemic caused shortages in silicon chips, pharmaceutical equipment, and other materials and parts we today may take for granted. Real estate and storage space are finite, just like supermarket shelving.

Digital space is also finite, even though sometimes the costs of maintaining data can seem so small in the micro incident. However, at scale, data and storage limits can have a significant impact. The Y2K bug caused a serious fear in banking and technology as a whole (despite not having an egregious impact), and similar data dimension problems can happen all of the time. Integer overflow problems, such as with the Ariane 5 rocket, can cause hundreds of millions of dollars of damage from a seemingly simple arithmetic error. On the flip scale, wasting inventory, such as allowing for fake social media accounts to exist on platforms owned by Meta and Twitter, can cost individual companies millions in maintenance fees and storage of fake photos and posts. Proper digital inventory management needs to be addressed at the micro level in our teams so that incidents don’t have a macro impact.

M — Motion

Wastes in motion are common place in everyday life, let alone in either the manufacturing space or in our digital teams. Having a tool on the top shelf, requiring a step stool to reach is a waste in motion because extra steps are needed than should be. In manufacturing, perhaps the equipment needed to add further value is on the other side of the room instead of adjacent to your station. In technology, similar wastes in motion are applicable with data. Overly complicated user interfaces or entries in data can cause wasted time. Even minutae such as “confirmations” or manual checks can be redundant. In an ideal program environment, only the right data can be entered. An example is allowing alphabetical keyboard strokes in a program looking for a phone number. This waste causes a poor experience for everyone involved.

W — Waiting

Arguably the most aggrevating waste on the list is waiting — recall how many times just this week we have sat in a line, in traffic, waited for someone to get us something we needed, etc. Waiting in manufacturing is an unnecessary expense to the value of product, and waiting in the tech sector can cause poor developer flow, poor backend timing, and poor user experience. Ultimately many cases of waiting digitally can be avoided by changing capacity. Perhaps there is a server problem where there are either: A) not enough servers or B) the servers are not fast enough. Keeping our code bases clean and efficient can help reduce this problem, as well as building functional code that uses modern parallelization techniques and vectorization wherever possible. We want our software to run smoothly and not exceed the capacity of our hardware.

O — Overproduction

Overproduction on a physical level is exactly what we would expect — manufacturing too many items than we have a need for. Digitally, overproduction can related to having too much service. Yes, too much service. Too much service is an unnecessary expense because the service available is not being adequately utilized. Perhaps this is using desktop computer to perform the task of a cheap microcontroller, or having multiple dedicated servers perform the same work a serverless architecture could perform. Ultimately, this is a conversation on costs and the amount of work that needs to be performed. Our code can also be overproduced with using improperly sized variables than necessary and being inefficient in our data structures, leading to similar costs on too great of digital inventory.

O — Overprocessing

Overprocessed manufactured goods provide more value than necessary to a customer. Some great examples of this are using rare-expensive metals for parts when cheaper ones would suffice, adding more layers of paint to a car than noticeable for people, and designing TV’s or other screens with higher framerates than are even detectable by our eyes. If nobody will notice the “added value”, how is it even adding further value?

Similarly in technology, our code can be overprocessed and Agile does a great job combatting scope creep, through structured and measured cadence with sprints. Managing software projects focused on “minimum viable products” and well defined essential features based on user feedback and requirements is the best way to combat overproduction digitally. Why add a feature to product, spending developer time, effort, and even pay for something not asked for or desired? While it is true that some features can be pleasant surprises that enhance a user’s experience, some users may never notice. Overproduction digitally is a waste of a team’s time and energy, and that time and energy should be better utilized elsewhere.

D — Defects

Defects are measured differently in different industries and even for different products in the same industry. A defect in printer paper, may not be noticeable to most consumers unless it is physically nearly destroyed. Slight changes in the appearance may not be detected. However, a defect in litmus paper used in the laboratory setting may totally ruin an experiment, which in turn can affect their reputation. Typically, the Six Sigma portion of Lean Six Sigma gets goals for defect reduction, the sigma being the variance of a statistical collection. “Three Sigmas” covers 93.32% of a population, four covers 99.35%, five covers 99.98%, and six reaches 99.9997%. Within the binary classification of defects, with a six sigma process 99.9997% of the units produced should be defect free, or around three defects per one million units produced. This is a hefty goal to reach, but a six sigma process is extremely reliable and consistent, and could definitely have the trust of a customer.

Software can have defects too. Think about how many times an app has crashed on a smart phone, or how many times a “bug” in a game or program prevented you from moving forward in a task. At its most insignificant, these bugs can have insequential, or even humorous consequences. At its most serious however, software defects can be life threatening. There are plenty of incidences where even one small bug uncontrolled and unpredicted cost millions of damages and threatened peoples’ livelihood. Ultimately there will be a tradeoff on every software team of anticipating bugs and getting a project done on time — however the goals are clear. The best code may only fail 3.4/1,000,000 runs.

S — Skills (The 8th Waste)

The newest identified waste is based on skills of employee. At first thought it may seem that the waste is a risk in not having the right skills, however the larger risk is not recognizing a workforce as a company’s greatest asset. This waste, or not enabling, recognizing, and supporting employees is the greatest waste. Especially in the tech sector.

With each new graduation year, a class and generation of new excited employees enter the workforce with new skills and backgrounds, hungry for experience. Micromanaging and a lack of creativity and trust in employees can destroy a company. Encouraging growth and ownership in the newest employees will build loyalty — and recognizing and rewarding newly developed skills and experiencing will have a domino effect on a team’s success.