The following is a transcript of a presentation given by Ralph Warchol, VP and CIO of Clayton Homes, at the recent IBM Virtual Global SOA Forum, "Work Smarter, Take Out Costs In a Tough Economy."
Ralph Warchol, VP and CIO of Clayton Homes:
My company, Clayton homes, is deploying SOA and BPM technology to satisfy the business growth objectives of our company.
First, let me start off giving you a little background on Clayton homes. In 2008, we remained the nation's leading homebuilder. This included the sale of over 28,000 new homes to customers in 48 states. Of these homes, over 16,000 were sold through 430 of our company-owned model home centers located throughout the country. We also provide consumer mortgage financing to the greater majority of these homebuyers. Unlike most new homes built, ours offer the advantage of being primarily constructed in climate-controlled factory environments that move to the building site for final assembly and completion. Because we build all out homes in a factory, we utilize labor-saving techniques not practical for the average site builder. With more than 28,000 homes built last year, we're also able to purchase materials at a more favorable price than contractors, due to volume. These significant savings are passed on to our customers in the form of very affordable housing.
Our 430 model home centers act as the point of sale for all new homes, each individual sales order is then subsequently processed by a team centrally located in our corporate headquarters facility. This team consists of 150 employees. Their daily work involves tasks such as order status tracking, processing of title, tax documentation, generation of plant surveys, mortgage processing, and all other activities required to move a new home sales order to from the initiation stage to the close stage.
The average cycle time of this process is not unlike what most of you experience when buying a new or used home. Approximately 30 to 60 days.
While our business has experienced steady growth throughout its 43-year history, none was so rapid as over the past five years, since becoming a part of Berkshire-Hathaway. And in all periods of growth, nothing is better when you can grow in the most profitable manner possible.
In late 2007, one of the areas we identified as a potential opportunity for enabling profitable growth was our centrally located 150-persopn sales order processing team. We identified that in order to grow profitably, we would need to find ways to improve the profitability of our sales order process. We recognized that our sales order processing team was A) spending increased amount of time tracking order status for model home centers and their customers; B) performing many processing tasks manually or with spreadsheets; C) redundantly entering identical data into separate systems, and D) unable to stay organized due to so much of their work activity being paper based.
So we recognized that addressing and improving this process could lead to more profitable growth, and just as importantly, that we could also increase our customer satisfaction and our position in the home building marketplace, through improved levels of accuracy and responsiveness.
....we had a very paper-intensive manual workflow. Our entire prior state(?) process was facilitated by the movement of paper sales order folders, one per sales order, through each step of the process until completion. Ultimately, these sales orders would be imaged for legal preservation, but not until all processing had been complete.
Imagine this if you will. Suppose you're a member of that 150-operson sales order processing team. At any given time, there are between 1,000 and 2,000 of these individual homes sales order folders being processed by your team. A model home center sales rep calls you on behalf of their customer to inquire into the simple status of an order. What happens next? Where's the order? Where's the folder? Which department is it in?
To say the least, we saw a significant opportunity to improve what we now refer to as the three VS. improving the velocity of the process, increasing the visibility into the process, and lastly, reducing the variability within the process.
To begin determination of what improvements were possible, we decided to start by modeling our prior state(?) process. To do this, we captured and analyzed several components that best represented the characteristics of our sales order process.
Among these were, first, defining the business measures that best represented our key performance indicators. For us, these included the process cycle time, wait time, and error rate. Secondly, we needed to identify our high-level processes according to six individual deal types. We sell homes in a variety of ways, according to the type of financing, whether the customer already owns land, etc. we also identified sub-processes within each high-level process. Typically, these represented the various individual tasks within the process, such as tax processing, title processing, loan processing, etc.
And lastly, we captured and mapped the various systems used to facilitate our process.
Next, over a period of several weeks, we physically observed our current state process, capturing the cycle times, wait times, and error rates of each process step. The sample size observed was approximately 1,000 sales orders from each of the six defined high-level processes. This observed data was plugged into the process modeler tool, and used to execute model simulations, and establish what was at that time a current state baseline.
One of the key learnings of this initial analysis was the determination that many process tasks could be performed in parallel. However, the ability to do to them so was being restricted by our paper-intensive prior state process. From this learning, we immediately began thinking about a more technology and business-centric process design that would address and improve our sequential priori state(?) process.
After identifying those process steps that could be executed in parallel, we rebuilt a new future state model, and began a series of simulations and refinements to quantify project saving and determine our ROI. As you can see in our future state model, many more process tasks are now executed in parallel.
Furthermore, through fundamental changes in our approach to the way sales orders are processed, we were able to employ the use of advanced BPM technology to enable much higher levels of parallel processing in our future state design.
As a result of performing simulations on our priori and future state models, using the actual employee labor costs from our sales order processing team, we were able to easily determine a dollarizer benefit. We determined that our total cycle time reduction opportunity was about 42%. Achievable through improvements to our total time required for each process step, as well as the queue time of work waiting to be executed.
The net result being an overall increase in the productivity of each sales order processor, defined in terms of the number of transactions or process steps they were able to execute each day. After annualizing these savings, and comparing to the cost of procuring the appropriate BPM software, subsequent execution of a deployment project, and ongoing maintenance and support, I found a very attractive ROI with a payback period of well under 24 months.
In early 2008, we gained approval and initiate a business project to move forward wit our future state deployment.
...the individual technologies that we selected to enable our future state project design.
First, imaging. Historically, we had imaged and archived all of our sales order documents and other legal artifacts on the backside of the business process. We had been using imaging products from Colfax in an integrated manner with IBM enterprise content manager for several years. But key changes made in our future state process is that we now do all of this imaging up front, as step one, not the last step.
Secondly, these images are then dropped into our WebSphere process server-based sales order workflow solution, where they are then processed from sales order initiation to close.
To provide ease of connectivity to legacy systems, we also deployed the WebSphere HATS product, which delivers AS/400 emulation server within our solution's browser-based client. Similarly, our solution was also required to integrate with the point of sale system used within our 430 model home centers.
You know, its said what you cant manage what you cant measure. So to ensure process visibility and the opportunity on continual improvement within our future state process, we also elected to deploy WebSphere business monitor. This now allows us the ability to measure process cycle time, wait time, and defect rates electronically, versus the old way through physical observation.
To build a solid BPM and SOA foundation using these products, we also elected to engage syscom incorporated, an IBM partner....
An additional objective of this initial deployment has also been to ensure that we build a scalable infrastructure, able to be leveraged against future WebSphere process serer-based workflow solutions within our business.
'
I'd like to share with you what I'd consider the most critical success factors in a successful BPM deployment. These types of projects sometimes results in radical business transformation, as it did in the case of ours.
The ability to capture the full return on investment depends highly on the competency of being able to facilitate and manage business process change. To best accommodate this, several things are required.
First, business stakeholder buy-in. we had this early on, and it was a key ingredient to the project's success.
Second, competency and BPM and SOA solution technology. Early engagement with IBM and its partners had helped to accelerate the development of our solution, as well as accelerate our own learning.
And lastly, well managed project methodology. I'm happy to tell you were currently deep into user acceptance testing and beginning preparation for final deployment and production in the next three months. Our attention will then turn to focusing on monitoring process performance via the WebSphere business monitor solution, to help ensure that we can manage the process, and more importantly, achieve the project's full ROI.
There will be plenty of case studies on SOA-based implementations at IBM's upcoming "IMPACT" conference, to be held May 3-8 in Las Vegas.
Ralph Warchol, VP and CIO of Clayton Homes:
My company, Clayton homes, is deploying SOA and BPM technology to satisfy the business growth objectives of our company.
First, let me start off giving you a little background on Clayton homes. In 2008, we remained the nation's leading homebuilder. This included the sale of over 28,000 new homes to customers in 48 states. Of these homes, over 16,000 were sold through 430 of our company-owned model home centers located throughout the country. We also provide consumer mortgage financing to the greater majority of these homebuyers. Unlike most new homes built, ours offer the advantage of being primarily constructed in climate-controlled factory environments that move to the building site for final assembly and completion. Because we build all out homes in a factory, we utilize labor-saving techniques not practical for the average site builder. With more than 28,000 homes built last year, we're also able to purchase materials at a more favorable price than contractors, due to volume. These significant savings are passed on to our customers in the form of very affordable housing.
Our 430 model home centers act as the point of sale for all new homes, each individual sales order is then subsequently processed by a team centrally located in our corporate headquarters facility. This team consists of 150 employees. Their daily work involves tasks such as order status tracking, processing of title, tax documentation, generation of plant surveys, mortgage processing, and all other activities required to move a new home sales order to from the initiation stage to the close stage.
The average cycle time of this process is not unlike what most of you experience when buying a new or used home. Approximately 30 to 60 days.
While our business has experienced steady growth throughout its 43-year history, none was so rapid as over the past five years, since becoming a part of Berkshire-Hathaway. And in all periods of growth, nothing is better when you can grow in the most profitable manner possible.
In late 2007, one of the areas we identified as a potential opportunity for enabling profitable growth was our centrally located 150-persopn sales order processing team. We identified that in order to grow profitably, we would need to find ways to improve the profitability of our sales order process. We recognized that our sales order processing team was A) spending increased amount of time tracking order status for model home centers and their customers; B) performing many processing tasks manually or with spreadsheets; C) redundantly entering identical data into separate systems, and D) unable to stay organized due to so much of their work activity being paper based.
So we recognized that addressing and improving this process could lead to more profitable growth, and just as importantly, that we could also increase our customer satisfaction and our position in the home building marketplace, through improved levels of accuracy and responsiveness.
....we had a very paper-intensive manual workflow. Our entire prior state(?) process was facilitated by the movement of paper sales order folders, one per sales order, through each step of the process until completion. Ultimately, these sales orders would be imaged for legal preservation, but not until all processing had been complete.
Imagine this if you will. Suppose you're a member of that 150-operson sales order processing team. At any given time, there are between 1,000 and 2,000 of these individual homes sales order folders being processed by your team. A model home center sales rep calls you on behalf of their customer to inquire into the simple status of an order. What happens next? Where's the order? Where's the folder? Which department is it in?
To say the least, we saw a significant opportunity to improve what we now refer to as the three VS. improving the velocity of the process, increasing the visibility into the process, and lastly, reducing the variability within the process.
To begin determination of what improvements were possible, we decided to start by modeling our prior state(?) process. To do this, we captured and analyzed several components that best represented the characteristics of our sales order process.
Among these were, first, defining the business measures that best represented our key performance indicators. For us, these included the process cycle time, wait time, and error rate. Secondly, we needed to identify our high-level processes according to six individual deal types. We sell homes in a variety of ways, according to the type of financing, whether the customer already owns land, etc. we also identified sub-processes within each high-level process. Typically, these represented the various individual tasks within the process, such as tax processing, title processing, loan processing, etc.
And lastly, we captured and mapped the various systems used to facilitate our process.
Next, over a period of several weeks, we physically observed our current state process, capturing the cycle times, wait times, and error rates of each process step. The sample size observed was approximately 1,000 sales orders from each of the six defined high-level processes. This observed data was plugged into the process modeler tool, and used to execute model simulations, and establish what was at that time a current state baseline.
One of the key learnings of this initial analysis was the determination that many process tasks could be performed in parallel. However, the ability to do to them so was being restricted by our paper-intensive prior state process. From this learning, we immediately began thinking about a more technology and business-centric process design that would address and improve our sequential priori state(?) process.
After identifying those process steps that could be executed in parallel, we rebuilt a new future state model, and began a series of simulations and refinements to quantify project saving and determine our ROI. As you can see in our future state model, many more process tasks are now executed in parallel.
Furthermore, through fundamental changes in our approach to the way sales orders are processed, we were able to employ the use of advanced BPM technology to enable much higher levels of parallel processing in our future state design.
As a result of performing simulations on our priori and future state models, using the actual employee labor costs from our sales order processing team, we were able to easily determine a dollarizer benefit. We determined that our total cycle time reduction opportunity was about 42%. Achievable through improvements to our total time required for each process step, as well as the queue time of work waiting to be executed.
The net result being an overall increase in the productivity of each sales order processor, defined in terms of the number of transactions or process steps they were able to execute each day. After annualizing these savings, and comparing to the cost of procuring the appropriate BPM software, subsequent execution of a deployment project, and ongoing maintenance and support, I found a very attractive ROI with a payback period of well under 24 months.
In early 2008, we gained approval and initiate a business project to move forward wit our future state deployment.
...the individual technologies that we selected to enable our future state project design.
First, imaging. Historically, we had imaged and archived all of our sales order documents and other legal artifacts on the backside of the business process. We had been using imaging products from Colfax in an integrated manner with IBM enterprise content manager for several years. But key changes made in our future state process is that we now do all of this imaging up front, as step one, not the last step.
Secondly, these images are then dropped into our WebSphere process server-based sales order workflow solution, where they are then processed from sales order initiation to close.
To provide ease of connectivity to legacy systems, we also deployed the WebSphere HATS product, which delivers AS/400 emulation server within our solution's browser-based client. Similarly, our solution was also required to integrate with the point of sale system used within our 430 model home centers.
You know, its said what you cant manage what you cant measure. So to ensure process visibility and the opportunity on continual improvement within our future state process, we also elected to deploy WebSphere business monitor. This now allows us the ability to measure process cycle time, wait time, and defect rates electronically, versus the old way through physical observation.
To build a solid BPM and SOA foundation using these products, we also elected to engage syscom incorporated, an IBM partner....
An additional objective of this initial deployment has also been to ensure that we build a scalable infrastructure, able to be leveraged against future WebSphere process serer-based workflow solutions within our business.
'
I'd like to share with you what I'd consider the most critical success factors in a successful BPM deployment. These types of projects sometimes results in radical business transformation, as it did in the case of ours.
The ability to capture the full return on investment depends highly on the competency of being able to facilitate and manage business process change. To best accommodate this, several things are required.
First, business stakeholder buy-in. we had this early on, and it was a key ingredient to the project's success.
Second, competency and BPM and SOA solution technology. Early engagement with IBM and its partners had helped to accelerate the development of our solution, as well as accelerate our own learning.
And lastly, well managed project methodology. I'm happy to tell you were currently deep into user acceptance testing and beginning preparation for final deployment and production in the next three months. Our attention will then turn to focusing on monitoring process performance via the WebSphere business monitor solution, to help ensure that we can manage the process, and more importantly, achieve the project's full ROI.
There will be plenty of case studies on SOA-based implementations at IBM's upcoming "IMPACT" conference, to be held May 3-8 in Las Vegas.
Leave a comment