3: Rethinking Business Processes
There is no single model for reengineering because the "right" process depends on they idiosyncratic mission and strategy of each organization - however, some general observations of typical characteristics can be discussed. Many businesses are similar enough that the most efficient way of accomplishing a given goal will be similar, even across industry types (e.g., the way that a tire manufacturer and a chain of nail salons order supplies and pay invoices is entirely similar).
Also, he industrial model that is common to many firms rests on the basic premise that workers have few skills and little capacity to learn - which invariably leads managers to design tasks such that many people each do a very simple task, with a great deal of process to orchestrate their activities. Even 200 years later, with a considerably more educated and capable workforce, firms are still operating according to the same principles.
Combining Jobs
One common feature of reengineered processes is combining multiple small tasks into a single step. Many firms have broken down tasks too far, to the point where a tasks has been oversimplified, and there is greater waste of time and potential for error in the handoffs between three (or more) workers who are doing what could he handled by one.
An excellent example of combining jobs is seen in improvements in customer service call centers: instead of bouncing a call from one department to the next, the person who answers the phone has access to all information related to a given customer or order and is able to answer basic questions on a number of topics without transferring a call. This dramatically reduces the amount of time it takes to handle an inbound query, which not only saves personnel costs but also delights customers who have become accustomed to long periods waiting on the line.
The same is seen in having general-purpose service technicians who are trained to perform multiple types of repairs: the team that arrives at a site can handle any problem, rather than having to call in a different team when the problem they encounter is outside their expertise or authority to address. The analogy is to a handyman who can do minor carpentry, wiring, and plumbing repairs rather than having to call in specialists, though specialists may still be needed when the repairs are too complicated for a generalist.
In addition to expediting service and reducing head-count, eliminating handoffs does away with miscommunication, delays, rework, and errors. For example, Bell Atlantic reduced the amount of time to install a digital line from thirty days to three, and in some instances the work can be done in hours because one installer can do several jobs (install outlets, run internal wiring, connect to the network, test the circuit, configure a router, etc.) Additionally, a technician who understands the whole system can more efficiently diagnose problems - there are no discussions, no sorting out whose fault or job it is, it just gets done.
Consolidating work also reduces administrative overhead: there are fewer handoffs to coordinate, and employees can quality check and troubleshoot a set of tasks and take greater responsibility for efficiency and accuracy of a broader range of tasks. Additionally, employees who perform more complex tasks are able to discover innovative ways to work more efficiently, and monitoring their performance requires less effort from fewer supervisors.
Empowering Workers
Another common practice is granting more latitude to employees in addressing problems or unusual situation. Traditional job design is based on ideal conditions and often does not provide employees with authority (or strictly forbids them from exercising common-sense) to work through any situation that is not optimal - instead, employees must call a supervisor to resolve a simple problem, who may need to escalate the issue through multiple levels of the managerial hierarchy to get to a position where someone is authorized to make a decision.
Reengineering can identify where decisions should be made closer to the work, enabling employees or low-level managers to quickly authorize a deviation from standard procedure when it is necessary. That is to say, decision-making becomes part of the work, not something that is separated from the work. Not only does this prevent a simple problem from stopping the line, but improves employee morale.
The assumption that the employee is an unintelligent automation, performing a task that could be done by a mechanical device, simply needs to be discarded.
Flexible Workflows
Traditionally, workflows have been designed by a straight-line sequence, with steps being performed one at a time in a specific order. Often, it is discovered that this is entirely unnecessary: the outcome would be the same if the steps were performed in a different order and some parts of the task were done concurrently.
Linear sequences imposes an artificial precedence among tasks, and is the chief cause of waste in process flows: the person who performs step two is idly waiting for someone else to perform step one (and a third person to carry the work from one station to the next), work arrives in large batches that is done hastily, then the worker sets idle again.
A common practice in reengineering is looking at the steps that need to be done, questioning the necessity of each step, and questioning whether there is a necessary order. This, alone, can result in significant improvements in efficiency when unnecessary steps and unnecessary waiting periods are eliminated.
The author reflects on the Kodak example for the efficiency that can be gained, and another example is given if a five-step order fulfillment process - by eliminating arbitrary sequencing, the time to fill an order was reduced by more than 60%.
The term "delinearizing" is used for breaking from strict linear control to a more flexible approach that allows steps to be done in arbitrary order if there is no specific reason to make them linear.
Versioning Processes
A common characteristic of traditional processes is that they treat every situation as a worst-case scenario. The example is given of a man who wanted to build a workshop on his property, and who had to wait six months for a public hearing before a city board to get approval to do so - because his simple hand-drawn plan for a simple workshop had to travel through the same review and approval processes as multimillion dollar office towers.
In a way, this is a byproduct of standardization - developing a standard product for the mass market required there to be only one procedure for production. But when products aren't standard, there can no longer be a single process. But even that metaphor doesn't fit, because processes involve parties that are not under the company's control (vendors, customers, regulators, etc.) whose behavior and needs will vary.
(EN: An interesting point to consider is how much businesses struggle to force these parties to play by its rules. Most of the "policies" that require customers to go through processes are attempts of the firm to control their behavior in a way that is easy, predictable, and standardized for the needs of the business - regardless of how awkward and unsatisfactory it is for the other party.)
Significant efficiencies can be gained by considering the possible scenarios, sorting them into classes, and designing an appropriate process for each. A "simple" credit application can be processed more efficiently than a complicated one; an invoice that matches a purchase order can be paid with far less effort than one that has discrepancies; etc. Processes that have multiple versions of paths have a "triage" step to determine which process should be followed, generally based on some fairly simple logic.
There are some qualifications:
Decentralization
Traditional practices favor centralization, on the premise that a centralized department is more efficient and effective at doing work than having people doing the same job in various parts of the organization. This is sometimes, but not always true - and it requires stepping back from the existing practices to consider where it makes the most sense.
For example, consider that many companies have abandoned having a centralized purchasing department. Rigid control means that the accounting department puts in a requisition for a box of pencils. It must develop a business case to prove to the purchasing department that it really needs them. Then the purchasing department must research vendors, negotiate contracts, place the order, receive and inspect the goods, pay the invoice, and deliver the pencils to the accounting department. It can take many days and lots of labor to make a very simple purchase.
This may sound like an exaggeration, but one of the author's clients monitored this and found it spent an average of $100 in internal costs to purchase any item. At the same client, it was found that 35% of purchase orders were for amounts less than $500. Clearly, this represents a tremendous waste of financial resources and unnecessary frustration to all employees.
Naturally, the answer was to let the accountants buy pencils as needed, and to change the purchasing policies so that departments had a $500 per month discretionary budget for minor purchases, handled through a corporate card system to make the process of making small purchases painless.
In a similar vein, an electronics equipment manufacturer reconsidered its policy that authorized servicemen must perform all repair and maintenance. Instead, it stored a small inventory of parts and supplies at each customer's site and provided phone support for their own maintenance personnel to diagnose issues and, where feasible, to instruct them to perform the task. Service technicians would make site visits only if the problem is complex.
Reducing Checks and Controls
Checking and controlling work is among the most wasteful and unnecessary tasks that can be done. They are generally implemented when a business process is inherently problematic and regular monitoring is needed to handle problems that constantly arise.
In many instances, the cost of checking and re-checking exceeds the value that would have been lost if no checks had been performed. This is fairly easy to identify by examining the report of variances, identifying the dollar-value of the mistakes that would have been prevented and comparing that to the cost of checking and addressing variances.
In other instances, the need to check up is simply a matter of distrust - there is a sense that employees are incompetent or dishonest and require constant monitoring. In many instances, this sense is entirely unfounded. And where there is cause for concern, it is more efficiently and effectively addressed by training and selection.
Reengineered process take a more balanced approach: they design processes that should not require duct tape and consider whether the cost of control exceeds its benefits. Even where control is desirable, it can generally be effected by less intrusive means: control should not create a delay in the workflow, and making corrections should not reduce throughput. Inspection can often be done at the end of the line and process improvements implemented.
Going back to the purchasing scenario, switching to departmental credit cards did no result on employees going on wild spending sprees, selling office equipment on the black market, or any of the ludicrous scenarios the internal auditors may have imagined. However, the system still had a checkpoint in that the budget was modest ($500/month) and departmental manages reviewed the expenditures at the end of each month.
In a similar situation, an automobile insurance company took a similar approach in claims processing: sending an appraiser into the field is expensive and the encounter with the insured is often contentious. By implementing a process change to approve claims when the repair bill was within certain tolerances, the firm decreased the cost of field adjusters, improved customer satisfaction, and reduced lawsuits. By periodic spot-checks, which include comparing the cost and frequency of vendors to one another, the company can prevent serious and ongoing abuse - and while it accepts that it is open to minor and short-term abuse, the cost of prevention exceeds the risk.
Minimizing Reconciliation
Reconciliation is another activity that does not add value to processes, and is undertaken to ensure accuracy. It is likewise prone to overuse - where the cost of reconciliation far exceeds the variances that are detected.
The case-study of Ford considers this: the original process included multiple points of contact at which the supplies were reconciled against various documentation by various personnel. Where there are redundant points of contact, the number of false alarms is multiplied by human error - i.e., the qualities actually matched, but someone made a mistake, and effort is expended in correcting these mistakes. By checking once, two-thirds of the auditing work and all of the work to correct human error was eliminated.
The author speaks to the Wal-Mart model of inventory management - which is to avoid it entirely. Checking freight in and out of distribution centers and monitoring inventory levels requires constant and repetitive reconciliation. For example, Wal-Mart partners with P&G over the management of diapers throughout its supply chain. By sharing logistics data with the vendor, and relying on their expertise in gauging demand for their own products, Wal-Mart could then allow P&G to recommend how much product should be ordered and when orders should be placed.
The system worked out so well that Wal-Mart decided to dispense with recommendations and approvals and have P&G just ship the inventory it felt was needed - thereby eliminating all the internal costs associated to maintaining its inventory. The sole checkpoint became a periodic review of products sold against products billed. (EN: The author doesn't mention this, but Wal-Mart is also known for having an agreement in place by which it doesn't pay for goods until they are sold, which eliminates much of the opportunity for abuse because a vendor can't fake POS data.)
There's a brief mention of positive vendor relations: much of the contentiousness and ill-will between suppliers and their customers arises from the reconciliation process itself - and more often than not, it is because of a human error, such as a miscount, rather than any real discrepancy.
A Single Point of Contact
Another instance in which division of labor created a high level of inefficiency and frustration was dividing call handling among departments - such that a customer would be bounced from one department to the next, each time re-explaining his problem before being forwarded along, and if the problem involved more than one department, the customer was unable to ever speak to a single person who had the authority to solve it.
Customer handling is often reengineered to give the customer a case manager, or a single point of contact who was aware of the problem and its history and authorized to answer questions and solve problems across the multiple business units that might be involved. In most instances, a case manager can resolve a problem - but in the instances were assistance is needed from other departments, the case manager serves as a buffer and coordinator.
Information technology is necessary to support such a position. Prior to computerization, the customer was bounced from one department to the next because the information necessary to investigate an issue was in paper documentation that existed in multiple places - but computerized information systems put all information at the fingertips of one representative.
Hybridized Operations
It is not necessary to choose one side, centralized or decentralized, and do everything by that methodology. Depending on the situation, one or the other may be the most efficient and effective; and within an organization, there may be some elements that are centralized and others that are not.
Information technology, particularly mobile computing, enables more functions to operate with autonomy and, at the same time, have the support they need. For example, a sales representative with a laptop computer and wireless modem can connect directly to information systems at the central office. Orders and inquiries need not pass through several hands in order to be processed.
But on the other hand, this same autonomy often produces chaos. Consider the example of a bank that has multiple autonomous units, such that a customer receives sales calls for several different products during the course of a single afternoon. Decentralization and autonomy work to ensure that each division has no idea of what the other is doing, and that no-one considers the banks' treatment of customers in a holistic manner.
To remedy the situation, some firms are instituting company-wide customer databases that are shared by all operating units. This enables units to share information with one another: what they know about the customer, what interactions they have had, etc. This should enable the units to act independently with a central reference point, without the bureaucratic interference of constant control.