WHITE PAPER

BPM & MANUFACTURERS – Gaining Velocity & Responsiveness While Mitigating Costly Errors

ideas . consulting . technology

 

      

Download | Email

 

 

 

 

                                                                            

 

 

 

 

 

 

 

 

 

Copyright . Lean Productivity Solutions . 2008 . All rights reserved.

All of the definitions and special terms used in the Appendix and at various places in this paper are the original work of the author. Any copying or attempt to reuse these terms without all suitable citations and credits to the author of this paper will be considered a breach of its intellectual property and rights of authorship and legal recourse will ensue against all such plagiarisms and breaches.

Introduction

Business Process Management emphasizes managing processes through task definition, task instructions, task routing, and task assignment.

 

This means that no longer is a process flow map merely an elaborate drawing used by operations management to standardize and say what a process should look like, but is in fact, within a BPM framework, a sequence of automated events or steps which together comprise a fulfillment pipeline for work. This pipeline routes business objects such as service requests, emails, faxes, SMSs, business documents, even signals from external systems(- databases, software applications, programmable logic controllers), through various 'work' attendants as determined by process paths, and the business rules and conditions which control direction and flow. These automated processes occur within a workflow environment that constantly monitors all the activities occurring within it in real time. To use non-technical language, an automated workflow environment is both "entity-aware" as well as "process-aware" and can be requested to collect information on service times, the length of waiting lines, or to give instantaneous information about the size and location of bottlenecks, the disposition of an entity within a process, and everything that happened to the entity at each stage of processing. Because of this “awareness”, not only can the workflow environment be asked to provide real time data but it can also be asked to issue notifications about what is happening within it. Taken together all of these capabilities mean that:

 

1.)    The authors of processes can exercise very close to 100% control over process compliance at the task execution level (i.e. whenever someone or something interacts with a workflow).

2.)    Processes can be evaluated in real time for robustness, efficiency, and completeness.

3.)    Managers, supervisors, and task owners have central visibility over all aspects of the processes assigned to them, or to which they have been granted administrative access.

4.)    Since the workflow environment is capturing data such as service times in real time, any process is self measuring and fully capable of providing business intelligence about all the sub-processes it is running, and all of the entities that may be flowing through or interacting with the process.

5.)    Workflow processes are self measuring (equivalently self auditing), time consuming simulation and time studies are needed only for experimentation on large scale changes or highly complex operations, and where pilots need to be set up and measured quickly.

6.)    Because the core goal of a good BPM platform is to streamline and connect fragmented business processes, and diverse disconnected pieces of software running vital business applications, such a platform seeks to provide an automated workflow in which all of the processes of the organization can be defined and executed. Therefore all of the task attendants that must interact with each process whether humans, other software systems, PLC equipment, or communications devices, can connect to and exchange information within the process and across it as well (i.e. with the process acting as a communications bridge). Because there are some core things that all enterprises do over and over again – they must manage contacts and relationships, manage documents and content, secure content, manage and seek to mitigate risk, manage and successfully resolve issues that impact quality, and provide business intelligence reports – CRM, Document Management / Content Management, Group and User Security Management, Communications, Audit Management, and Reports form the core of a strong BPM platform. This forms a natural backbone much like the backbone of a fish which, with an open architecture, is capable of connecting external systems and databases, electronic input devices, and the web.

 

How Is A BPM Platform Important To Manufacturers

There are many factors that will drive forward thinking manufacturers to invest in BPM technology and to interpret and extend its capabilities to various applications throughout their enterprises. These factors include:

 

1.)    Training and re-training personnel from paper based manuals is impractical and expensive at best, and at worst, next to impossible given increasingly high rates of employee turnover, and frequent product changes necessitating rapid creation and updating of process and procedural documentation.

2.)    The need to capture and retain knowledge as a form of capital.

3.)    In order to ensure quality, manufacturers must constantly monitor the fitness of raw materials for production and ensure conformance of QA, Engineering, and Manufacturing processes to professional and industry standards. But with increasing competition in a global economy to produce better products, faster, with ever increasing levels of customer service they are being forced to condense the time it takes to execute all of the processes from design engineering, through supply chain operations, to manufacturing, production, and distribution, and to leverage the web to make their Customer Service and Support organizations more accessible and affordable to customers.

4.)    The impact of an increasingly competitive global economy has not just forced manufacturers to develop better products, faster, but also to compete with a product pricing trend that tends downward even as product quality and speed of delivery are expected to go up. This implies that manufacturers must invest in change not just as a program that attempts to transform operations, but actually as an infrastructure that supports 'newness' and innovation on a frequent and ongoing basis.

5.)    The cost of delayed detection and inaction concerning any instance of non-conformance is inversely proportional to visibility and a real time response capability. That is the lesser the degree of visibility as processes are executed, the lesser the ability of the enterprise to respond, and take action; and without a real time monitoring and intervention capability, responsiveness is drastically reduced. The lesser both of these are the larger and more numerous errors become. Please see the appendix section for a more rigorous explanation of these relationships. Now let’s consider the example below –

 

Manufacturing Case Scenario – Humans Always Advance Action

 

Let us imagine that Product A has just been re-designed to better compete with a competitor Product B in the marketplace. Hopefully the revisioning of product A has involved a series of well coordinated design changes - revisioned engineering drawings, new raw material specifications, new assembly or fabrication requirements - well coordinated communication of new raw material specs, and intermediate part assembly changes to supply chain vendors and outsource partners. These also have hopefully been communicated internally so that production re-tooling, new or revised assembly procedures and instructions have been drafted and personnel trained, and new computation or review of manufacturing cycle times, and capacity performed. Finally sales has received all the relevant information about the new product in order to confirm new pricing and allow marketing enough time to publish the latest and greatest about product A. Field support, and customer service have each received updated documentation as well. All of these events have occurred in tandem, in their correct sequence, and at the right time in order to achieve a successful delivery of product A to market. All of the new documentation including component part drawings or pictures, and manufacturing and assembly instructions has been delivered to the document control center ready for access by production control personnel. All of the assembly equipment and setups have been adjusted and tested to manufacture the revisioned part. Production of Product A is set to kickoff in a week and the arrival of certain time sensitive components with a limited shelf life has been timed to coincide with that time frame. A day before production is due to start, manufacturing engineering realizes that due to an error 2 pages from the old assembly instructions got mixed in with the new and need to be replaced and re-submitted to document control. The engineer in charge immediately advises the document control attendant, and the production supervisor, who stop the release of any assembly documentation pertaining to Product A’s build and delay production until the corrected instructions are delivered. But there is a problem; a typical production scenario plays out –

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 -   The regular production supervisor who knows everything there is to know about Product A has a car problem enroute to work on kick-off day.

-    The production manager assigns a cross-trained production supervisor from another area in the plant to fill in and manage Product A’s line until the regular supervisor gets in. The manager is more concerned with daily production yields and time performance than he is with other issues that affect quality in production. He is not really expected to get too engrossed in the daily details passed back and forth between engineering and production unless he is specifically alerted or involved.

-    The substitute supervisor has been advised of the delay in production and the reason for it, can see that everything else has been readied and is set to run, and that even the inventory of materials for the day’s production has been staged. He is gratified that himself and his team will be ready to roll the minute manufacturing and document control advise him that the assembly instructions are ready. He does not know however nor has anyone informed him that some of the raw materials involved in the production of Product A are time sensitive.

-    The stocking personnel who have already pulled the inventory onto the shop floor know that some of the materials are time sensitive but are unaware of any production delay. They arrived with the materials into the production area as usual, 10 minutes before the start of the shift – the required staging window whenever time sensitive components are involved.

-    Now the time sensitive material is out of its controlled environment and is already aging and beginning to degenerate.

-    It has taken some time to correct and re-submit three copies of assembly instructions to document control because it took a little time to remember which folder the file with the correct pages was in and since there was a large queue at the 11” x 17” color printer used for printing assembly documentation the ME assistant went back to his desk to try to get through the mountain of ECO’s there thinking that he’d just try again later in the afternoon.

-    The corrected documents are available in document control by the following morning.

-    The regular production supervisor never made it in to work the previous day but has resolved her transportation problems and is present and accounted for at the start of her regular shift the morning after kick-off.

-    She and her team work at an excellent pace, and by noon time, have produced 500 units of Product A at a cost of about $30 per unit. They will resume producing units after the lunch break.

-    The 500 units have consumed according to plan all of the time sensitive material from the start of the shift. The material must be delivered several times a day in order to maintain freshness for production. In between batches, a qualitative chemical analysis is performed in a random sampling of units belonging to the most recently produced batch.

-    When the supervisor and her team arrive from the lunch break, the QA Associate is waiting with the results of the qualitative analysis. All of the units in the random sampling have failed inspection and the entire batch must be discarded. The QA Associate is trained to consider several factors that would cause this high rate of sampling failure but must completely shut down further production until the root cause can be investigated and ascertained.

 

… Sounds familiar?

 

 

 

So What Exactly Happened Here

 

·        Ownership, retention, and dissemination of the most current and up to date knowledge of the enterprise were left almost entirely to humans.

·        The organization owns or can leverage only as much knowledge as the last piece of information typed and stored on some computer or printed and placed in some manual somewhere in the organization.

·        This organization is never able to convert instantaneously changing information and knowledge into central real time intelligence, which is where and when it is truly needed. At best it is always lagging behind by ‘when’, and ‘who’, knew ‘what’ so that action is mostly tardy – “too late”.

·        Left to themselves diligent personnel will make moment by moment decisions about what to do and how it is to be done based on their last best understanding of what their responsibilities are as communicated by the organization. In the absence of any recent instruction they will naturally either rely on the most recent information at their disposal (which apparently may not be current at all), or invent an approach, neither one of which is necessarily what the organization needs them to do under present circumstances.

·        Lastly humans can only take action to the degree that they are ‘aware’ of what’s happening. In today’s large dynamic enterprise with ever changing events and information it is simply impossible for the human worker or even hundreds of human workers to reconnoiter and respond to this dynamism in any controlled systematic way.

 

 

Enter BPM Best Practice & Workflow Automation

 

 

  1. Since the BPM workflow model emphasizes quality as a process element and not necessarily a personnel element, business rules have been imposed to set up conditional attributes on all issues affecting quality in production as a result of changes to the Bill of Materials on any part.

 

  1. There has been a specific process defined for upgrading an existing product or introducing a new one.

 

 

Control Flow Logic

 

 

  1. Included in the process is the setting of a time sensitive flag on any product that utilizes materials that are time sensitive. The default setting is flagged or checked. Anyone updating or creating a new record against the BOM is required to double check whether this flag should remain checked or unchecked.

 

  

 

 

  1. Once a product has a time sensitive flag associated with it every other object or record related to it inherits the flag.
  2. Since the BPM system has a Document Management System and content repository, document control is not a physical department where paper based documentation is controlled, instead, design, quality, and industrial engineers draft, check in, and submit documentation into the DMS with all the requisite permissions for viewing and interacting with it. Production personnel either print fresh instructions at the start of each shift in extreme environment work areas, or else their production personnel have the instructions displayed before them on smart terminals.
  3. There is a BPM process defined for check-in / check-out out operations into or out of the document management system. The checkout dialogue form requires the engineer to indicate whether a stop / delay production flag must be imposed as a result of the checkout. The workflow environment is configured to monitor the checking in and out of documents from the DMS.
  4. Any time a document related to a product with a time sensitive flag is checked out, a notification is sent by the workflow environment to all supervisor and manager work stations in any work center that has open work orders for the product.
  5. A document related to a product with a stop / delay production flag imposed on it cannot be printed on the production floor and a notification is sent by the workflow environment to all supervisor and manager work stations in any work center that has open work orders for the product including all warehouse and materials work-centers. A special advisory is included in the notification if any time sensitive or other special materials have already been released out of inventory to the production floor.
  6. The advisory contains the production floor staging area name and other pertinent information along with the move request code. Since material stockers and movers rove about all throughout the facility they wear a special SMS pager which is paged via the same notification event as well.

 

 

 

 

 

  1. Personnel receiving such a page must remove the material from the production area and pull it back into inventory and into its appropriate storage or use special stickers to flag it in its production staging area.
  2. Time sensitive or other special material related to a product with a stop / delay production flag on it, cannot be released to the shop floor from inventory for that production.
  3. A work order cannot be created or released for a product with a stop / delay production flag on it.
  4. A production run cannot be commenced on any product until a production supervisor or other personnel fires the clear-to-start signal by initialing a special field on the production advisory.
  5. A clear-to-start field is color coded red and disabled if a stop / delay production condition exists and requires further action to be taken, or a clearance from engineering to be re-enabled.
  6. No computer visuals or instructions can be displayed at smart terminals if the clear-to-start has not been fired or if a stop / delay production flag exists.

 

Conclusion

 

The difference between the two scenarios is obvious. The former consists of a well-intentioned effort to design and implement process controls where knowledge is exclusively retained at the personnel level at least as often as it is retained in some system that is static and non-central to the wider organization. The latter consists of a robust BPM platform with automated workflow into which process and procedural documentation is centrally edited and published, well thought out processes have been defined and automated, the right control conditions have been established and imposed, the correct activities, events, and communications occur at the right time and in the correct sequence between the right task owners and systems, and knowledge is retained and transferred by the BPM system independent of personnel.

 

The very real costs associated with both scenarios can be appreciated without delving into any detailed cost analysis. This is just one example of the power, promise, and versatility of BPM in manufacturing.

 

… Discover the difference.

 

 

Appendix

It was stated in an earlier section of this paper that the cost of delayed detection and inaction concerning any instance of non-conformance is inversely proportional to visibility and a real time response capability and that the lesser the degree of visibility as processes are executed, the lesser the ability of the enterprise to respond, and take action.

 

Let us consider now a more rigorous explanation for these relationships.

 

Let   ν                 

 

 

 

Let   t response        

 

 

 

 

Let   ų                   

 

 

 

be some factor representing central visibility of process knowledge and changing events within an open system  – the  organization

 

represent the response time taken by a task attendant to initiate and complete an action that leads either to minimization or mitigation of some issue that costs an organization some thing

 

represent a measure of response capability within the open system

 

The maximum non-conformance cost varies as

 

   ~C α (1/v + 1/ų) where ų = 1/ t response

 

Response time as a function of an entire process can be approached by considering the effect of each individual capsule of time taken to initiate and complete a response to some incidence of non-conformance.

 

 t response = η PRC(∆t) + η PNRC(∆t)

The two terms of the response time expression are explained as follows:

η PRC – is the number of possible instances of a non-conformance occurring throughout all the steps of a process for which a response capability exists to either minimize or mitigate the overall cost of each non-conformance.

t        - is the response time taken to address each of these ranging from theoretically ideal ‘0’ response time to the longest period taken from the start of an event requiring a response until an action is taken and completed to address it.

 

η PNRC – is the number of possible instances of a non-conformance occurring throughout all the steps of a process for which a response capability does not exist at all. This response time is taken to be an extremely large number ד reflecting the time between the start of an event requiring a response that will not be forthcoming for a long time, and when eventual action is taken and completed. If the event remains latent and is never discovered before the cost of the non-conformance reaches some maximum cost ~C, the response time becomes infinitely long and represents the “Absolute Anti-Response’ – a term being defined for the first time here in this paper for the purpose of this discussion.

 

ד             - is taken to be an extremely large number ד,  representing the time between the start of an event requiring a response that will not be forthcoming for a long time (because nobody knows that action should be taken or that action has not been taken), and the time that it takes to discover the non-conformance providing the discovery is made before the maximum conformance cost ~C is reached.

Text Box: t=0 ∫ t t=ד dt         t= ∞ Text Box: ד=t ∫ t t= ∞ dt         t= ∞

 Therefore - t response = η PRC      +  η PNRC

 

A method for an empirical experiment can be proposed using two hypotheses:

 

1.       If an open system – the organization - achieves visibility and a real time response capability several cost factors associated with non-conformance will be seen to be minimized as visibility and real time response capability increases.

 

2.      If an open system – the organization – does not achieve visibility and a real time response capability several cost factors associated with non-conformance will increase as latency and invisibility increase until ד (time when a very late and costly discovery is made) is reached or the non-conformance remains latent and undiscovered past the point where visibility and action will have any impact on cost.

 

Cost Factors: *DL – Direct Labor; *IDL – Indirect Labor; *%FC – Percent Fixed Cost; *MC – Material Cost; OC – Opportunity Cost Lost; PRC – Public Relations Cost

 

 

Simulation Experiment

Introduction

The component of non-conformance cost that is due to material and resource waste (i.e. the cost of materials, direct labor, indirect labor, and percent fixed cost that is irretrievably lost with the non-conformance) is for the purpose of this paper, being considered to be at most equal to and at least negligible to the size of these same costs in the presence of undiscovered incidence of non-conformance. Furthermore we theorize that these non-conformances become aggravated and multiply with increasing time and distance away from the event(s) that gave rise to them increasing the overall non-conformance cost. The purpose of this experiment is to empirically verify that this is in fact true, and to test if the relationship as stated in this paper, between visibility and a real time response capability is an inverse proportionality or has some exponential components. Because of the improbability of successfully configuring this experiment in a ‘real’ enterprise a simulation study is being proposed to carry out a virtual experiment.

 

                                                            ~C α (1/v + 1/ų) where ų = 1/ t response

 

Method

The approach is not a rigorous one and requires further analysis and thought. But within the scope of this paper it should be enough to lay a framework for certain conclusions to be made. Our approach is a basic 3-point test against both the proposed and control hypotheses. A point represents any occurrence at varying distances from the entry boundary (being defined as the point at which a product / service begins to cost the enterprise something) and the outer boundary (being defined as the point beyond which a product / service ceases to cost the enterprise anything). An internal point is any point in between these two boundaries that we want to test at increasing distances away from the entry boundary. Finally to lend coherency and create a real enough virtual model it is proposed that a list of non-conforming events be compiled. A list based on the manufacturing example used in this paper is a good place to start. Establish a method and system of reporting findings as follows:

 

                                       

 

Resources will need to be defined along with the process or activity performed by each and their associated processing times. Each resource will need to be assigned a $Resource Cost per unit of time, and a cost attribute – DL, IDL, MC, or %FC - will need to be associated with the resource cost. A suitable product / service entity will need to be selected for the experiment preferably one for which an arrival model has been successfully validated.

 

Variables

ρ= probability density of a non-conformance occurring at some D~T (Distance-Time).

A suitable random occurrence generator for non-conformances will need to be found utilizing roe.

Each non-conformance will need to have associated with it an irretrievable dump cost factor that accurately reflects the cost impact on each of the four cost factors – DL, IDL, MC, or %FC - occurring at intervals of D~T. Use the best expert guess available based on historical trends. Configure accumulation variables to measure the non-conformance cost at varying internal points. Finally the statistics that will be collected will need to be configured. Run the simulation.

 

 

 

**** Copy Rights and Rights of Authorship *****

All of the definitions and special terms used in the Appendix and at various places in this paper are the original work of the author. Any copying or attempt to reuse these terms without all suitable citations and credits to the author of this paper will be considered a breach of its intellectual property and rights of authorship and legal recourse will ensue against all such plagiarisms and breaches.