Process Enhancement and Optimization | Ahmed Fawzy | Skillshare

Playback Speed


  • 0.5x
  • 1x (Normal)
  • 1.25x
  • 1.5x
  • 2x

Process Enhancement and Optimization

teacher avatar Ahmed Fawzy, IT Transformation Advisor

Watch this class and thousands more

Get unlimited access to every class
Taught by industry leaders & working professionals
Topics include illustration, design, photography, and more

Watch this class and thousands more

Get unlimited access to every class
Taught by industry leaders & working professionals
Topics include illustration, design, photography, and more

Lessons in This Class

11 Lessons (51m)
    • 1. Learning outcomes (Change the Process)

      0:22
    • 2. What Is Process Improvement?

      3:02
    • 3. Statistical Process Control

      2:59
    • 4. Quality

      2:21
    • 5. Cost

      3:03
    • 6. Time

      2:15
    • 7. Organization Position

      4:52
    • 8. Process Improvement Practice

      12:01
    • 9. Pull System

      1:15
    • 10. Process Control Practice

      15:06
    • 11. Process outcomes

      3:31
  • --
  • Beginner level
  • Intermediate level
  • Advanced level
  • All levels
  • Beg/Int level
  • Int/Adv level

Community Generated

The level is determined by a majority opinion of students who have reviewed this class. The teacher's recommendation is shown until at least 5 student responses are collected.

28

Students

--

Projects

About This Class

In this short course, you will learn about 

  • Overview of Process Modeling.
  • How to Draw a process.
  • Analyzing the process results.
  • Determine areas of process improvement.
  • How to control the process quality, cost and time.
  • Determine the correct balance between quality cost and time.
  • How to handle the bottleneck in your process.
  • The drawback of the optimization you need to consider.

Meet Your Teacher

Teacher Profile Image

Ahmed Fawzy

IT Transformation Advisor

Teacher

Ahmed Fawzy, is an Advisor, Author, and Online Trainer. He has 18 years of experience in the fields of IT transformation. Utilizing a unique approach to achieve a better alignment to the business through solutions and processes. Also, how to transform IT organizations successfully from "Traditional to Digital."

Ahmed holds ITIL Expert certification and ITIL4 MP. He is also a certified Project Management Professional (PMP), TOGAF 9 Certified, and has a Master in Business Administration (MBA).  He has implemented improvement programs for a wide variety of organizations. His approach is unique because it doesn't require new additional software or hardware,  "It's simple few adjustments that yield a high return." Ahmed's goal is to help leaders transform their IT internal o... See full profile

Class Ratings

Expectations Met?
  • Exceeded!
    0%
  • Yes
    0%
  • Somewhat
    0%
  • Not really
    0%
Reviews Archive

In October 2018, we updated our review system to improve the way we collect feedback. Below are the reviews written before that update.

Why Join Skillshare?

Take award-winning Skillshare Original Classes

Each class has short lessons, hands-on projects

Your membership supports Skillshare teachers

Learn From Anywhere

Take classes on the go with the Skillshare app. Stream or download to watch on the plane, the subway, or wherever you learn best.

Transcripts

1. Learning outcomes (Change the Process): now we reached the 50 station in the journey, how to implement the process and how to optimize it as well. In this section, you will learn what is a process improvement and how using statistical process control will improve on it to provide more stable and dependable process. 2. What Is Process Improvement?: Brussels improvement is all about finding what your organization uses to generate revenue and how to do it better by doing it cheaper, faster with good quality before deciding on the process improvement, you have to do commend the process. First, understand it and decide possible areas for improvement. I had created another course covering the business analysis and the process modelling. Check it out. If you are interested in such topic, you will be amazed that a lot of redundant steps and unnecessary loops show up on Lee when you recommend the process, if you are not seeing it and measure it, you might think things are okay while they are not. Moreover, you will not notice until the situation gets terrible. Every activity in the process need to be measured. Determined the biz line for it. In other words, what's normal? And if there is a problem, target that specific problem and solve it. Moreover, measure of your solution improve the outcome of the process or not. The critical question is, Do you have metrics? Do you measure all activities or not? Do you measure a monitor all activities in the activity chain or not? If you measure all activities. It was be overkill and will consume lots of resources on time. You start about analyzing the process outputs. Is it with the normal range or not? Afterwards, you begin investigating backward whenever you find an output to determine off output and what's normal. Try to compare it to the industry, standard or organization numbers in the past years for the same period. As he noticed, These things are not easy. If it's the first year of analysis, you can improve using only common sense and nothing else. Some institutes cell industrious, standard information. But I found these are very generic numbers and only help. If your case is extreme, it will not assist you to gain a competitive advantage. The only way to improve his benchmark against yourself, however, as the first step to improve a process, is simplification and reduction of steps to achieve the outcome. After you drove the process of steps using flow chart or swim lane, ask the following questions. One. Can any steps be deleted? Two. Can any steps a beauty mated or partially with meat? It three. Can any steps to be combined? Four. Can any steps be performed by one person instead of two five. Is it possible to run activities in Barrow? Answering these questions will open up lots of optimization projects that will generate value for the organization without much investment. Thank you. 3. Statistical Process Control: the broth of control is finding the ideal condition off the activities. If any off these activities changed from the ideal state off, the process, outcome would be less desirable. Maybe these activities will be focused on speed, coast or quality, or maybe even balance between all of these. The idea is finding the right spot on this triangle that suits the company best. There is no good or bad position. The location is determined by business needs to fulfill a specific market need by using the specific mix of resources. The impossible spot is the exact middle. This means 100 quality with 0% waste and 100% on time as well. This is almost impossible and if possible, it will be very expensive and resource intensive. This means there are almost infinite number of possibilities where the process will be performing. Ideally for the organization, the numbers on the triangle are to the ministry. The concept on Lee, for example, buoyant number one is high quality high coast with low output. This is apple, for example. The problem always is changing. This point from one point to another develops a gile and cloud help organization changed the position in the triangle fast to achieve better results for the business. Without the three, it's difficult to move the boat location off the organization, but with the three organization can select and change the position as bears the market demands. Otherwise, it might fall short or do toe slowness. It might even move too late toe allocations. That market already moved away from thinking Nokia with smartphones or Kodak with digital camera to decide where the process should be. You need to understand the quality coast and time. The issue was most example. You will find they talk about the final product, but you really find someone talking about services. When it comes to the triangle, you need to understand the both are the same. And instead of the production line to create a product, you have servers, people and devices. This means products for services is the outcome. It will be the email message coming out of the user machine or the mill you get in a restaurant or a product you will get from a supermarket. The product concept is obliged toe all aspects, including services. In this lecture, you learned about how specific mix of quality coast and Time will provide the organization value added for its customers and the position in the market. In the next few lectures, you will learn each aspect of the three and how to influence them. Thank you for watching and see you in the next lecture. 4. Quality: to determine quality. You need to understand two things prevention coast or time and the failure cost. So let's start by revenge in Khost on time, the cost on time consumed to ensure that every activity is done up to the standards. The more quality you demand, the more coast some time required. If you noticed the curve dramatically, go upward after the 50% point. If you also notice the graph stops at 90% because increasing the quality from 90% to 100% would require the line Toby nearly vertical. So we use a different graph for this remaining 10%. Next, we have the failure coast, the coast off time wasted due to low quality. We have to rework again and waste resources because the product was not up to the standard . The lower quality you have, the more expensive it is do to rework and the even lawsuits. In this case, the coast on time scale is a wasted resources trying to recover from low quality. Now, if you apply both grafts, you will end up with the lowest coast is where both curves intersect. But unfortunately, this is a point off commodity. This means you will have an average quality. And in today's business, this is not a very good choice, since most companies that focus on quality outlast the companies with focus on coast think like Apple or BMW. This will lead us toe another graph. Try always to stay on the Golden Zone zone one. If you go forwards on toe the gray zone, the Failure Coast will outmatch any profit generated from cutting corners. So always try to stay about 55% quality range. Now you understand quality. It is time to decide where you want your quality. Toby. This is done by calculating the Failure Coast and the prevention cost and select a point on a scale from 0 100 Now you have a quality equal 70%. In this movie, you learned about quality on how the organizations like the desired quality for the final product and services. Thank you for watching and see you in the next lecture 5. Cost: next we moved to coast. Since no organization want to increase the cost, this lecture will revolve around how to reduce cost reduced coast. You need to cut waste, wasting time, waste of materials, waste of facilities, defects, any defects in the product that render it useless or called the customer to return it over . Production units over produced and waiting in stock waiting the time wasted. Waiting for a product or half a product waiting for the next step. Not utilize the talent. Any improve with the skill not used is considered waste toe. Have someone only utilizing. Part of their talent is a massive waste and organizations today transportacion any movement of people or product or anything, the movement need to be optimized for the least transportation time Inventory access room materials work in progress. Finished materials. If it's not for a customer, it shouldn't be on your inventory motion. Waste. This is effort in people and the wear and tear of the machines. Excess processing over designed. The process to do think the customer doesn't need or care for eliminating always does not cheap. You will always have some sort of waste, which is OK as long as you know it and have full control over it. To calculate the cost. Combined all the mentioned waste if you have them in the product and reach a total number. Coast divided by coast blast waste were coast. Is the coast the company business expect to reach a specific market targets? This means the business prize, a product or service to become $10 or the product or the service will not sell in the market, since this is will exceed the expected value. How value is perceived is out of the scope of this court, but is discussed in my other course business analysis and the process modelling, I explained. How would do we create and on the life value? Let's have an example there. It cost is different from coast. The unit cost is the cost visible to the business. It includes direct and indirect coast and fix it blas variable costs as well for simplicity . I'm referring toa all of that by the Ward Coast. Let's assume our sales and marketing the side. It needs to be 10 U. S. Dollars so they can sell agonist competitors and sell the expected value off the market in this case. This means you will have the actual coast, the actual materials, the actual expenses off the actual device or product or service. Blow us. The coast of waste will be equal. 10. Let's assume the course. The six ends always this four. By feeding this into the formula mentioned earlier, you will end up with 60%. In this lecture you learned about coast and how waste impact organization. Thank you. 6. Time: Now we move to time other than waste of time, which already discussed and the cost lectures in this lecture, we will discuss lead time, the time it takes for the first product or service to go through the entire activities. If one activity is causing constrain or bottleneck or cues, this will call the waste of time to see if sometime have a better layout to move things between activities. Empress's Foster In case of I t. This is could be networked bandits. It could be direct connection Faster connectivity on the faster processing as well to each . Q is a bottleneck. Try to eliminate it or reduce the Q. If the elimination of it is difficult, let's have an example from the mention diagram How many units we can produce per hour. I will let you four seconds to try to solve it yourself. Without calculation, we will only be able to produce 60 unit per hour. The slow is tough to five that you will be between Track four and Truck five, while Task six will be underutilized. So how long for the input toe become output? Assuming this is not the first unit five minutes 15 seconds because, like you, Onley moves once every 60 seconds. So Task one will have to wait to take a new task for 60 seconds for the process. The task in but to become output task to and so on. If you are thinking of enhancing Tosca two or task for this will be a wasted capacity. The only improvement toe the line is to reduce the top five to which extent you plan to enhance the time is dependent on the mend. In our example. The time equal bottlenecks divided by total process time, which will equal 60 divided by 315 will equal 19%. This show how inefficient the time is utilized to produce the final require the product or services. 7. Organization Position: now to find the position in our triangle, we need to do the following one. We draw three lines each lying cup, two circles in tow, half to withdraw, lying coming from the center off one of the circles and lend on the line on the opposite direction, splitting it in tow. Exact half lend on the exact center. The intersection point is the position the company should be, and this is the third ist. So let's clean things a bit. You will end up with this. As you see, this is the result off average quality, lots of waste and inefficient time management. The whole idea of finding the location of the triangle is to show you the changing it. It is not an easy task and will require lots of agility and might impact other aspect in an unpredictable way. Let's take a busty and demand as an example. If you try to reduce cost by reducing the call centers. In our example, there is no point off serving more than one person, but 62nd it will pile up before the bottleneck, and our total capacity is 60 people per hour. If our demand is only 50 units per hour and you receive a call in a predictable way or the statistical term is evenly spaced. There is no point in improving, but if your demand is 90 on it, it is worth investing in improving the current A bottleneck enough to service require demand. If you're the man does not evenly spaced and in the reward it is not even this based. Then you will need toe measure the fluctuation in the men and have a reserved capacity to serve it. Let me give you an example. In 30 minutes, you will receive a request off 40 units under another hour. You receive I three or 10 units in total. Both will average below 60 per hour, but both are different and we can not rely on average is in this case, the humans not evenly spaced is utilization divided boy one minus utilization. In the mentioned. The example, the 40 units is considered the 66% utilization of the line, so we will have 66 divided by 100 minus 66 it equal 1.94 killings. The average with time is 62nd. This is the average unit Wait time because off the bottleneck, So we will have 1.94 divided by 60 it will equal 116.5 seconds. This means each one will stay on hold for this time to eliminate accuse. You will get the cue to become one not one point my force, which is a person or the unit being served. Right now, it's example, you need to make the utilization 50% which mean bubble of the capacity to eliminate the waiting time. This is the capacity required to reduce the Q size. The problem is organization need to reduce the coast by increasing utilization and the customer need lower waiting time. You will have to find the balance between the two. This is why it's nearly impossible to believe the organization exactly in the middle of the triangle. The idea behind this example is, whenever you optimize your process, don't cut it to the bone to make it 99.99% utilized. You might think this is efficiency, but this will generate under effect to the customer and will impact the organization on the long run. What if a delay happened in task one or two then the delay is not actually a delay. It will only leave a gap in production. But since these tasks are foster than the bottleneck, it will catch up once the issue is resolved. This is considered a buffer, for example, to conclude the buffer if Task one was delayed for task. Five to Finney, the Q, which is four minutes. And for the unit to reach a task. Five. It will take 20 plus 30 plus 10 plus 25. So the buffer will be 240 minus 85 equal 155 seconds. This is a time allowed for task. Want tohave an issue without impacting the line. So the best breakfast is for the bottleneck is make sure it never breaks down, never run out of supplies and never supplied with damaged or defective parts. Thank you for watching and see you in the next lecture. 8. Process Improvement Practice: and this practice, we will learn how to process model on organization. But first we need to understand what is an organization. How the business functions. First, you will have the Brymer is dream off any organization. This is the core function of the organization. For example, if you have a riddle organization, the Green bar will be considered the ideal business, and every other biller is a supporting function. It could be I th our finance. Or if it's a financial organization, then the finance will be the primary pillar. So the green bar is identifying factor off the organization why the organization exists. So this is the core, and any other pillar is a support. Any organization have an input and output, the input could be materials. If it's a manufacturer, organization or skills and some other materials as well and few other materials. In this case, it could be a service organization. In all cases, most organization, the revenue seeking organization, will have an output over revenue okay and a profit. So this is the high level off what an organization is. Let's take an example in our example here you have the procurement process. This is the first process in your organization. Orderto generate a product at the end. You have warehousing, you have online Web store and you have delivery. You could have multiple other, but this is only for the administration for supporting function. You could have email file storage that Abby's recruitment at the HR accounting and finance . All of this is to drive and to generate and to convert the input, materials and skills in tow on out. But which is a revenue? So let's take an example how to do this forced. The real problem with process modelling is that you can take years to build all of these building parts. Your first step is to start to cover the entire organization by creating the simplest way off modeling the process. The easiest way to model a process and a practical terms is to create a service blocks, which is the steps that any service and process go through in order to generate embodied in tow. Albert, this mean all you have to do is to create take one step toe, step three and step for you're using. Get this for simplicity because you will have hundreds off processes inside than any given organization, you could if you model it with the deep technical details, you can take months, maybe a year, and during this year, the process could be easily changed or modified. So the easiest way for you to do it is to start building at using steps, and you keep rolling up Bates inside the process document so you will release in a week or two or a month. The entire process is off the entire organization, and later you keep building the flow charts and the swim lanes and the context and the abstract diagrams, which leaders toe the next step. You add logic to your process. The logic to the process is building a simple flow chart as mentioned before the flow chart is on Lee. For a single actor, this mean a single team or a single entity or a single person doing a specific flow off mosques. These are the most common flow chart symbols. You have the start and the end, and you have yes and no. You have the process. You're processing something in the square or rectangle, and you have the sub process. It's a tangle with a couple of lines and you have the data the most common that you will be using in any given flow chart and in 90% of the cases are the first Onley four. This shape is next page shape. You use it on Lee if your process is so complex and has so many steps that you need to create multiple bages. So this is used to clean things a bit and to redirect instead of having multiple arrows going through all over the place and confusing people who reading it. Because flow charts should be either top down or right left, it should have a single direction flow to make it easier for someone to read it. So the next step is to add logic is yes, no decisions, and you keep going until you reach the stop, which is a condition at the end. Your next step is toe. Add relationship between each flow chart on each other by adding multiple functions and multiple actors, for example, you really have a single team reforming an entire process or performing on entire service. In this case, you need to see some interactions between the processes and the multiple teams. In this case, we introduce the shared the profits that swim lane. Basically, you will have multiple lanes. Okay, Each of them You will have multiple processes, so each block off these should be a complete flow chart off these and this is Onley. Demonstrate that interaction between functions. This is the primary purpose off the share the process end swim lane diagram. Next you add a little bit information. You add a little bit off information to the share of the processes you add for each process , the input and the output, and you can list the owners as well. This is all in a single diagrams. This is the most advanced and sophisticated level off the swim lane diagram. Your input could be a forum. Okay, and you're out, but could be a process female or a processed function and the owners off this out. But who will process and who will do the flow chart off the input and the processing to generate the expected out? This is the most advanced stage, and this is typically handled in a very mature process. So let's take an example. We have here a part of a process that a user need to create an email. Okay, so let's go through your first build. The process details. This is you have to get right from the beginning of the process and you build the steps or the main factors and the main blocks off the process. So let's read it together instead of his disk. Receive a request from HR that the user joined our company. So first, the owner of this process is clearly the service desk received a request from HR. That user joined our company. We send the detail, the form or the reactivation form. If the user was working with a company before, so the first step is to collect information. This is the first step. So the process objective in this case is to provide user with the means off productivity and timely manner for the HR request toe. Very activation or the creation for the user required the resources. Next, you have the collect user information and requirement. This is the second step Collect user requirement. We contact the user manager to get approval for the hardware required and the software required. They use their manager get back to us. With hardware and software required. We could eat two forums. One first asset management team toe Acquire the new laptop and another one for the licensing team toe. Approve the license usage for that specific user. So this is our acquisition stage. Once approved both the hardware and software received. We we worked with the security team to create the user identity and access to our system. This is the fourth step. The identity. Once he created, we create an email and stole the software packages on user laptop and hand over to the end user to get the USAR sign off. This is the email, laptop and the handover process. So now, after reading the entire process, is script The Interviews service desk. You generate the details, the main details, the process objective to provide user with the means of productive ITI and timely manner and governance of user creation. User asset provided in accurate and timely manner. The processes scope is to provide user with a laptop and identity and access on finally email account. The process. Constrain a lot off manual actions using paper forms because we cede multiple times a year that there is multiple forms keep going back and forth between multiple people in the organization process Assumption Noah Soleil in this process. But soon it should be implemented. This need to be highlighted because something like manual paper form will impact the process else. L A significantly last part is involvement apartment. Clearly, this is the involved department will be the i. T. This is the mean actor of this process. So let's translate the steps in tow. A flow chart. This is the next step. So following the same script you have the HR request is this is a new or old user. We added a little bit of logic in this. If a new user and you user form will be sent If on old user every activation formed will be sent, both of them will collect user information. Then we contact the manager for his approval. If the manager disapproved, then we send it to HR. The manager disapproved the user. In this case, it will be an end for us. If yes, we get the hardware and software requirement, we send the hardware forum and we send the software forum. We have the licensing team process and we have the asset team process. Each of them are our sub process that outside the scope off this flow chart we see here a custom diagram. This is oclock. This is not a standard flow chart. You can add whatever diagrams you think will deliver the meaning and will deliver the ideas that you are trying to communicate. But the really important thing is to add a key to your diagram. This is will translate what the simple will mean, but tried to avoid having too much custom symbols because this will confuse people a lot. So now back to our flow chart. When I say we have the acid team process and we have the licensing team process, each of them deliver the hardware and the software. So once delivered, it would be moved to another beach because now we run out of space. So now we have this diagram covering toe another page because otherwise the process will be very long. So we need to segment it in order to make it more logical. So now we continue our process. Now we send to security team the security team, create identity and access. If not, then you had the clock once more, which is a cell a escalation. And finally, you have the sub process off email creation, the Sapporo's off laptop preparation and hand over to end user on. Finally, you will get the sign off. This is, if you see will map toe each step we mentioned earlier. So we have Step one collecting information. It will end with manager approval. Step two. Requirement. It will end with hardware and software requirement. Step three. Acquisition. It will end with hardware and software received and stiff for identity. It will end up with identity and access created Stay five email. It will end up with a male creation step Sex Laptop was lacto preparation Step seven handover. It will end up with the handover to the end user and receiving the sign off. Thank you for your time and see you in the next lecture. 9. Pull System: now the pool system. Another way. To others, the bottlenecks is the pool system. Poor system is when you take on Lee what you want for the activity. You don't push on. Pile up. Let me give you an example. Buying from the online store is pulled wired. Your email inbox is push. The idea is in our production line. Example. One. Task five. Request from dusk for six units to task for produce them 36 units delivered to Soft Fife. Four. All trust Go and help Task five to Brussels Z six units Foster five. Task five Place another order six. Everyone goes back to their position to create more output for task five. So this missile cannot be applied. In some cases, in other cases, it will improve drastically. Example is coffee shops, fast food restaurants, etcetera. The idea is toe have better teamwork and balance of capabilities. 10. Process Control Practice: In the previous breakfast, we created the steps summing up the process. Now you need to start analyzing, get a little bit deeper toe, understand? Where can you optimize such process? Some processes will not require optimization, but if you feel and your users feel there is something missing and you have unpredictable results out off this process in this case, you need to start analyzing it toe find Where is the problem and where is the problem area to analyze the fix later on down the road or arrange a project for it. So first you start creating that table for Time, coast and quality for each of these steps. By collecting samples, you can do this yourself manually, or you can use a software to collect all of the results. But at the end, you need at least seven off each in east seven samples off time, cost and quality. The time cost and quality for each process will be a little bit different than the other. They will not be called time, cost and quality because they are different from one process to another. In this case, in our case, let's fill out the table. We will end up in this. We have seven samples which assemble one to assemble seven and it's repeated. This means that the first sample here is the same sample we collected. But we extracted three factors and three results from it regarding time, cost and quality. In our case, it will called for the time. It will be the lead time. How long it take for the first item to be through the process toe getting out of the process. Next we have the waste of time. How long did it take us actually to repeat the action and the third is prevention? How long did it take us to educate the end user to do a specific task that actually helped us a little bit in the lead time and reduce the waste? You do this, you collect these information for each and every sample and you collect this information for each and every step for simplicity purposes. We will do it on Onley, collect information the first step. So based on the collect information, we start the building a graph. This graph is ah, high level analysis off the process off the specific collect information process for the three factor for the time, cost and quality. We will be calling them now, Lead time, waste and prevention. So let's analyze the graph for a little bit to understand a little more about our process. First, we see a little bit of a trend off our process. Lee time going upward. This is the first observation. The second observation when the prevention was born. Five. We found it. The actual lead time will be a little bit lower, and the waste is almost nothing. We observe also that on the sample number six, when we have the lead time of three, we actually have a waste of two, which means there might be a correlation between those two values. Maybe the waste is related to the actual lead time, but you found year on other values that contradict this observation, which is the observation and the sample Number two. We found a very highly time, and actually you find near zero waste, which mean there might not be a correlation between both values. So we start analyzing it a little bit deeper by finding out where is the center of our values. What value considered the normal on what value considered extreme. So we started by analyzing the mean value, the center value, which all values revolve around. We find it out by first summing up all values for the old sample in lead time, waste and prevention. So if we actually some of the values starting from 1 to 7, we will end up with 13. Our next is to find out the mean, which is actually the some the value of the 13 divided by a number of values one through seven miners, one because this is a sample. So to avoid statistical buys with abstract one value, which mean now we have our mean This is the number that all other values revolve around. We draw a line here to see it. If you can see it, it's a little bit higher than expected. This is the number you're trying to hit with your easily. If you're SLE considered, this number is not an acceptable number. In this case, you need to start thinking how to make it better, how to make it closer to your business objectives. So next we start analyzing a little bit deeper to understand more about our process. Our next target is to find the deviations. This is the numbers from our means, the differences in the values from our mean. But the problem is with deviations. When use obstructed from the actual mean you will come up with negative values. So when you sum up all the negative values, all your results will be negative and this is will not be possible. Toe have a negative value. So in this case, we square it to get the deviations. This is a square number that actually cover our lead time differences from the mean we go a little bit deeper to find the variance. The variance is a deviation, the some off all deviations divided by the number of samples. This is the squared value off the differences between this, The next step is to find the sigma, which is called standard deviation. The standard deviation is a measurement off death, a dispersion. So this is a measurement off. How close the data in your process is grouped together across the mean. So now you will have to measurement. The first is the mean. This is to measure your S L. A. Against the second is the standard deviation above or below the mean. This is the grouping off your data. So let's see our graph. So now we have here the mean we have here two lines. The 1st 1 is going from the highest value, which is three, and goings toe the lowest value, which is buoyant. Six. This is called the range. This is not the standard deviation. The idea behind the range is governing all your data and so abstract from the highest to the lowest. And this is will give you your range. This is will not give us a lot of information other than your process. Have something wrong because actually you're mean is a little bit higher than the half off the range, which means your process is a little bit. It's cute. We will see this a little bit later. Now you have your standard deviation. This is the bluffs and miners from your mean what the standard deviation tells us. The standard deviation tells us that actually the boy in six results is abnormal result. This is not the correct result that we actually should target. The more common results are the results falling in this specific noted board, which means the 2.3 is considered normal. The 2.8 is considered the normal as well. Even the three and the two boys. Sex is considered normal, but the boy in 6.7 and the one is not considered normal in our process. This means if this results is not acceptable to you, that you actually need to start working to lowering your mean to change your average value . So now let's dig a little bit deeper to find out more information about our process. This is how our process actually look like. At this point, if you draw a graph and we draw a curve around our data, So the blue dotted line is actually our current process. Our target toe make it the red bell curve, which means this is called normal distribution. The idea is, the normal distribution is a predictable output over process. But you can see from the dotted line that actually our process keep going up and down, up and down, up and down this mean our process generating unpredictable results which mean we cannot actually rely on our process. So your first objective is to start regardless off the mean, regardless of anything to start grouping all your results under the bill curve shape because this means the results coming out of your process is predictable. And then you'll start working on actually moving the entire shape right or left, increasing or decreasing based on the business requirement. So now let's understand the Sigma range and how we optimize it further. So now the target is to make our results coming out of the process a little bit predictable , but how we can optimize it at the end. First, you start with the big red curve. This will be your basic target because it will mean a predictable result. But over time you will start squeezing it a little bit. You will have a lower range. So how we do this first you find out what is the range you are trying to target. You do this by finding out your sigma, your standard deviation, the upper limit and the lower rim. You find the upper limit and the lower limit by a very simple rules. It's the mean bluster minus the size of the bell curve multiplied by the sigma, which is a standard deviation. So what is the required rule if you want it toe have one Sigma Tau, sigma, three Sigma or even six Sigma. And this is the leanest possible in any process with acceptable coast, of course. So in our process, since our process is completely unpredictable, if we actually use the seven samples we already got, you will have a range off 1.14 until 3.18. This is your range. The majority of your data should fall between the strange. Your target is to start lowering extreme values to make your standard deviation a little bit acceptable to your organization. Now we learned how to calculate your standard deviation. Your next step is to start applying this on every step off your process. So let's take an example. If you actually take one part of your process and start doing a little bit deeper, he will have three steps. It's animal creation and the provisioning. The step. You will have a step one, a male creation step to sending and receiving and step three the deep provisioning you calculate for each step, the standard deviation for time, coast and quality. You will have the sigma off each and start analyzing it a little bit deeper. For example, you will find that the standard deviation for time in Step two is actually 1.52 which means you're results is widely varied. This mean by applying the bottleneck rules we learned in previous lectures. This means this is our bottleneck. At this point in time, we need to start working on the highest standard deviation because it's mean this process is the one with the most unpredictable results. So we take this a step and start breaking it down, monitoring the sub components inside it. So now we have Steptoe. We have Step One American and sent Steptoe email received by the internal server. Step three email process. Security checked. Stay for a millrate into the cube and Step five email sent each one of them. We start drilling down on finding out there Sigma until we started noticing that actually, in step four, we have steps, Reese, therefore and step five, we have a very high in time. We have a very wide range in time in each one of these steps. So our target focus at this point will be to focus on analyzing Step three stiff four and stick Fife by analyzing the time on Lee, we found out that actually each step have a sub component. In our case, we have the span filter and we have the sending to the Q. We have the spam filter. Actually, we have a very wide range, which is 1.9. And in step four, we find that, actually the main contributors that increased our standard deviation. Our Sigma is actually written to this quiches 3.7, which mean we have a very wide range when actually the servers trying to rating something to the disk. And actually, in step five, we have the open session to the remote host with have a very huge wide range. So what this graph tells us it is that we have some possible issue in spam filters. We have a possible issue in our disk in our storage, and we have a possible issue in the remote connectivity and maybe even Dean s resolve. This will help us later on the side on the actual functional and nonfunctional requirement to solve our problem that we actually discovered by applying statistical process control toe our processes. Most of these data is already generated automatically by monitoring software's. But you need to understand from where these days is coming from on how you build upon it, because this will help you down the road. To start analyzing and deciding and breaking down the problem, go find a real solution that will solve your business problem based on your given information and giving data. Thank you for your time and CNN's Annex lectures. 11. Process outcomes: this is critical in any process and especially in the case of develops deployment. If your deployment failed or generated undesirable outputs, how can you measure what's normal and what's not? This is where statistics come in. To understand this when you see the process, outcomes or activities multiple times. In other words, this is a process measurement. Over a period of time, you have three options. Perfect scenario type one. No grouping of results. Type two results are grouped. Perfect scenario is a process. Generate the same results each time. And this is the ideal scenario by one no grouping of results in this, the processes outcome widely very h process Ron. This means the process is unpredictable. You cannot predict where the next outcome will be. In a statistical term, the results have too much variation. To do that, we take a sample reading and do our plot. This will give us an idea off. What kind of process do we have? For example, user creation. How long does it take from you to get that new user request for you toe actually create the user Type two results are grouped. The process outcome will vary, but with a predictable range, and this the average is drifted in a statistical terms. This is called MMR. We take type one and make it type two, and before the process off type two, we try to shift it toe the ideal scenario. In some cases, we cannot reach the ideal scenario or situation, so we decide the variation limit. We think it's OK and improve based on that. What if your process was well and suddenly things started to get worse. How do you intervene? You will need to recalculate your sample five to make sure that that is accurate. However, as the rule of some, you need 1.2 billion where the billiard is a typical reporting period, usually weekly. This means you need to wait 8 to 10 days to decide the things went bad in the process. If your process was ideal or Neil Ideal state, this means that things will not fall out of control yet, but it's still in the predictable level. This is your process now, where the solid bars are the ideal scenario, the process turned into this. Now you have four types of outcomes. The 1st 1 if the process is working ideally within the solid in herb lines. Next we have the process has a little more variation in the outcome. The small dotted line in this case intervene after 1.2 periods. Three. The process has more variation and outcomes. The sick dash line. In this case, intervene after boy and five period. The third option is the outcome. Fall outside the thick dashed line in this case, intervene immediately. In this lecture, you learned about how to measure the process and identify the processor status in your organization. Also, you learned about when to intervene on went to hold back and simply wait because this is normal variation off the process. Thank you.