amazon highlights: Reaching the Goal / J.A. Ricketts / 2008

Reaching The Goal: How Managers Improve a Services Business Using Goldratt's Theory of Constraints (IBM Press) by John Arthur Ricketts

 

1 Introduction

Services that seem straightforward can be immensely complicated beneath the surface. the further services depart from industry, the harder it is to apply industrial methods.

Once the Industrial Revolution began in the U.S., it took nearly 100 years for services employment to exceed industrial employment. In contrast, China, India, and other developing economies are already there.

A constraint limits what can be produced by the factory or office as a whole. Everything else is a nonconstraint.

One fundamental difference between industry and services is that services cannot start until the customer arrives.   

Another difference concerns inventory. In industry, inventory issues are pervasive. In contrast, there can be no inventory of completed services because as soon as there's something to ship, it's more like a product than a service.

A third difference between industry and some services is less often discussed: customization.

 

2 Services On Demand

Goods are tangible and can be consumed now or later, while services are intangible and cannot be produced in advance. Of course, many purchases consist of a mix of goods and services. services on demand simply means getting service when, where, and how you want it.

PSTS (Professional Scientific and Technical Services), however, sells expertise. That is, to win business, PSTS enterprises must more often differentiate themselves via unique capabilities than any other sector. The main reason for this emphasis on expertise in PSTS is that its services generally require extensive customization for specific customers. In particular, the PSTS sector is unique even among services sectors because it's the only sector where:

• The primary output is services, but sales are made on expertise.
• Workers are typically assigned to serve specific clients.
• The degree of customization for specific clients is extremely high.
• Reliance on intellectual capital is quite high.
• Repeatability of processes is relatively low.

An on-demand enterprise has several defining attributes:

• Responsive—The enterprise can sense and respond because it has an integrated view of its customers, employees, suppliers, business partners, and competitors.
• Variable—The enterprise can alter its productivity, costs, marketing, and finances as conditions change because it has flexible processes and resources.
• Focused—The enterprise concentrates on its core competencies while using strategic business partners to handle its noncore tasks.
• Resilient—The enterprise readily handles changes and deals with threats because it has a flexible operating environment.

often recommended by TOC, is to become more responsive to customer needs by changing distribution rather than production. That is, most changes are made downstream in warehouses, transportation, and the backroom of retail stores rather than in the factory itself. The TOC application that accomplishes this change is known as Replenishment. TOC can be the means to become an on-demand manufacturer. TOC can also be the means to become an on-demand service provider.

Replenishment allows a factory to produce large batches efficiently while becoming more responsive to customers by distributing small batches. No compromise is required. a handful of services conflicts:

Resource Management—Efficiency requires high utilization, but responsiveness requires resources on the bench, available for immediate assignment.

• Project Management—Clients value projects that start and finish on time and within budget, but unique projects are inherently unpredictable.
• Process Management—Ongoing business processes are regulated by service-level agreements, but the service provider has little or no control over demand.
• Finance and Accounting—Expense control optimizes individual contracts, but the enterprise will sacrifice future performance if it fails to invest sufficiently.
• Marketing and Sales—Taking on riskier contracts increases revenue, but it also may increase troubled projects, which lead to higher costs and lower client satisfaction.
• Strategy and Change—Experts who advise others on strategy and change can have blind spots when it comes to their own enterprise's strategy and change.
• Implementation and Technology—One of the biggest obstacles to implementing new technology is the large installed base of old technology.

 

3 Theory of Constraints

All TOC applications spring from a common premise: If an enterprise is viewed as a chain, the enterprise as a whole can produce only as much as its weakest link will allow. That weakest link is, of course, the constraint. But the chain analogy has another implication: Pulling a chain is a lot more effective than pushing it. So switching enterprises from push to pull is a key ingredient in every TOC application.

A likely place to see WIP is ahead of the constraint, because it can produce less than any other step, by definition. Therefore, that WIP is sometimes mistaken for the buffer, but the drum buffer is actually all work scheduled on the constraint, even if it's currently at an earlier step. That is, the buffer is measured in time, not physical WIP units.

final element of DBR is the ropes, which govern when gating events occur. The shipping rope governs work on the constraint needed to meet market demand and keep the shipping buffer green. The constraint rope governs the release of raw materials to start new jobs that should keep the drum buffer green.

Like the buffer, the length of ropes is measured in time, and the ropes are actually contained in an information system.   

Placement of the drum, however, should be strategic, not accidental. That is, when DBR is first implemented, the constraint location may not be correctly aligned with respect to profitable market opportunities. If so, rather than implement DBR around this previously unseen constraint, it generally makes more sense to adjust capacity so that the control point represented by the drum is relocated to a position where the factory will be better able to produce goods that meet market demand profitably.

In their pure forms, Enterprise Resource Planning (ERP) assumes infinite capacity and schedules all steps, while DBR assumes finite capacity and schedules just the constraint. Some ERP software can schedule to finite capacity, but it does not have other essential capabilities of DBR software. For instance, ERP prohibits late release of materials, while DBR prohibits early release because it increases WIP. Moreover, ERP drives material requirements all the way through the bill of materials (BOM), regardless of stock on hand, while DBR takes existing stock and buffers into consideration. Thus, ERP and DBR are fundamentally different solutions.

Lean/Just-in-Time (JIT) seeks to optimize individual steps, while DBR optimizes the entire process around the constraint. They are fundamentally similar, but Lean/JIT doesn't work as well in job shops as flow shops because job shops have more diverse and changeable routings.

The best way to break a sales constraint is simply to sell customers what they want, when and where they want it, at a price that corresponds with perceived value.

Replenishment relies on aggregation to smooth demand. The factory warehouse may be owned by a distributor rather than the manufacturer, but the effect is the same.

Because sales occur daily, shipments occur daily, too. And the quantities shipped are just sufficient to replace goods sold. At first glance this might seem to increase shipping costs over what could be achieved by shipping large batches less frequently, but the net effect on total shipping costs is that they usually go down, not up. Stopping the shipment of obsolete goods and reshipment of misallocated goods more than compensates for increased cost created by smaller shipments of saleable goods.

Therefore, as aggregation reduces variability and DBR reduces resupply time, the required buffer size decreases accordingly.

allocation of large overhead expenses on the basis of small labor costs has created distortions.

A second problem with CA (Cost Accounting)  is it can encourage factories to produce excess inventory beyond what is really needed to fulfill customer orders.

A third problem with CA concerns management priorities.

Operating expense tends to be managed closely; revenue tends to be viewed as less controllable; Inventory tends to be a distant third in management priorities because reducing it has an adverse effect on reported income.

TA (Throughput Accounting) is not, however, a substitute for conventional financial reporting because publicly traded companies must comply with generally accepted accounting principles (GAAP). Fortunately, TA can be readily reconciled with GAAP reporting even though TA is a different approach to management accounting.

Under TA, there are no product costs. Instead, there are constraint measures:

• Throughput per Constraint Unit: T/CU = (revenue – totally variable cost) / units
• Constraint Utilization: U = time spent producing / time available to produce
• Change in Net Profit: DNP = DT – DOE
• Payback: PB = DNP / DI

 

To minimize unfavorable deviations from plans, these control measures should be minimized:

• Throughput Dollar Days: TDD = Selling price of late order x days late
• Inventory Dollar Days: IDD = Selling price of excess inventory x days unsold

TDD measures something that should have been done but was not: Ship orders on time. IDD measures something that should not have been done but was: Create unnecessary inventory. The ideal value of these control measures is zero.

In summary, TA is used to identify constraints, monitor performance, control production, and determine the impact of particular decisions. Several TA outcomes are noteworthy:

Financial measures reverse management priorities from OE, T, I to T, I, OE.

Performance measures are not distorted by cost allocations.

Constraint measures eliminate conflict between local measures (machine or worker utilization) and global measures (enterprise performance).

Control measures remove incentive to build excess inventory and establish incentive to deliver on time.

Hence, previous TOC applications turned push into pull. TA tells the enterprise what to pull.

TA is not widely used because:

• services investments can be highly intangible and reusable, while industrial inventories more often are not.
• services in general are less repeatable than industry, and Professional, Scientific, and Technical Services are the most customized of all.
• many services markets are moving away from services as available to services on demand.
• the degrees of freedom in delivering services can be greater than in manufacturing—particularly when the services depend on creativity.
   

4 Resource Management

the central challenge in a service chain is to coordinate the services, not necessarily acquire resources.

Hire-to-deal is based on the assumption that local optimization adds up to global optimization.

Hire-to-plan is based on the assumption that forecasts are accurate and stable enough to produce global optimization. Supply-demand matching is based on the assumption that supply is either fixed or adjustable only in large increments.

In services, the bench, which is composed of resources without current assignments, is a buffer against variability in both demand and supply. This dual nature of the bench, plus the fact that it's composed of people rather than objects, means the bench must be managed differently than physical inventories.

In industry, returned goods are the exception; but in services, resources returning from assignments are the rule.     

In services, however, resupply is driven by incremental demand and attrition.

In services, dependent demand is quite common.

There is seldom much reason for a services enterprise to acquire resources in batches, except during acquisition of other firms or during the college recruiting season.

while physical inventories create operating expense, a carefully managed services bench is an investment.

In services, aggregation can be just as difficult, but for different reasons.

there can be three broad kinds of demand within that market,

• First, normal contracts have a short to moderate sales cycle of six months or less.
• Second, large contracts have a long sales cycle lasting about 6 to 18 months.
• The third and final kind of demand is change authorizations, which change the scope or timing of services.

RS can be applied to any kind of demand, but it is most effective on normal contracts and change authorizations.

Deciding how to group resources is the first step in applying RS because aggregation takes place at the group level. Groups should be composed of resources that are as interchangeable as possible. Members of skill groups can be further distinguished by attributes such as specialty, education, geography, and performance. Each resource should be a member of one primary skill group for aggregation purposes, but that does not mean secondary skills and other attributes are unimportant. Note that a skill group is not the same as a resource pool. Skill groups should be reasonably differentiated from each other, while resource pools are relatively undifferentiated. Thus, RS may lead an enterprise from a hierarchical to matrix form of organization.

Buffer sizing is the second step in applying RS. Each skill group requires its own buffer because resources cannot readily be substituted from other skill groups, by definition, and patterns in demand and supply often differ between skill groups. The recommended buffer size in RS substitutes net consumption for total consumption. Net consumption can be computed as new demand minus available supply for a given period or total demand minus total supply for an entire skill group.

While the buffer size determined in the preceding section is the target, the buffer level is usually the actual number of resources on the bench. If the buffer level is negative, however, its absolute value is the number of assignments that could be made if the bench weren't empty. For instance, if the buffer level is –3, three open assignments are waiting for the next available resources.

net consumption and the buffer level move in opposition. When net consumption is persistently high, the buffer level can become negative, thereby indicating that resources are needed but unavailable. When net consumption is persistently negative, the buffer level rises to excessive levels unless steps are taken to move the buffer level toward the intended buffer size.

With this rule of thumb, a normal (green) zone surrounds the buffer size. If net consumption is normally distributed, a normal zone one standard deviation above and below the buffer size covers the buffer level about 68 percent of the time. A shortage (red) zone one standard deviation below the normal zone and an excess (red) zone one standard deviation above will each cover the buffer level about 14 percent of the time. That leaves the buffer depleted about 2 percent of the time or overflowing about 2 percent of the time, but those conditions should be transient if net consumption is not in a significant upturn or downturn.

Recall that in RG, the buffer level is the entire inventory for a given raw material or product, but in RS the buffer level is usually just the bench, not all resources in the skill group. Thus, in this figure the buffer zones are shown floating on top of base capacity and temporary capacity. Base capacity matches the growth trend of the skill group, while temporary capacity matches cycles.

As a practical matter, buffer management under RS typically works fine on a fixed weekly schedule, particularly if most resource assignments start and end on whole weeks. Buffer resizing is often considered at the same time as buffer management, but buffer resizing actually occurs far less often. Likewise, skill regrouping may be considered at the same time as buffer management, but it occurs even less often than buffer resizing. skill regrouping may actually occur only once or twice a year, if at all.

To know how many resources are effectively on the bench, resources who will be returning shortly must be added to those without a current assignment, and those who are already reassigned must be subtracted.

A virtual buffer is a real skill group buffer plus available resources in other skill groups with a matching secondary skill code.

Secondary skill codes capture the entire spectrum of skills possessed by individuals within PSTS enterprises.

For many individuals, secondary skills are just as sharp as their primary skill. However, the proficiency level of resources with a secondary skill code isn't always as high as the primary skill group.

In PSTS, skill groups fall into three categories. First, commodity skills are plentiful, can be resupplied quickly, and never constrain the enterprise, so those skill groups usually have no buffer. Subcontractors to PSTS enterprises may, however, buffer commodity skills. Second, skills that are not commodities, yet are rarely if ever the constraint, should have an appropriate buffer, and managing it is straightforward, because resupply is relatively quick and reliable. The majority of skill groups staffed with employees fall into this category. Finally, skills that usually are a constraint should have an appropriate buffer, but managing that buffer can be anything but straightforward, because resupply is neither quick nor reliable.

RS, on the other hand, does optimize skill groups across the enterprise, and it's responsive to conditions in both the services and job markets. In contrast to prevailing resource management methods, which push resources from practices or departments onto contracts, RS pulls resources onto contracts from relevant skill groups based on actual market demand.

 

5 Project Management

Projects are considered repeatable if they have comparable objectives, work breakdown structures, scopes, schedules, staffing, and deliverables.

Services projects therefore need planning methods that cope with ambiguity and instability in tasks and deliverables.

Services projects thus need measurement methods based on something more flexible than milestones.

Services projects thus need scheduling methods that decouple projects unless they have dependencies based on deliverables.

Services projects thus need communication methods to coordinate constrained resources across enterprises.

Services projects thus need replanning methods to shift selected responsibilities between the client, service provider, and subcontractors.

Services projects thus need quality assurance methods that identify trouble without unduly burdening projects that are trouble-free, diagnose the cause of trouble, and apply a suitable remedy.

Services projects thus need risk-management methods that steer projects based on the larger business context.

Services projects thus need risk-reward sharing methods so that the client and the provider both benefit when the client's needs are well-served.

Services projects thus need assets, and service providers need methods to generate revenue from those assets.  

Services projects thus need professional program and project managers.

 

6 Process Management

Interface constraints are, however, neither external nor internal. They exist literally at the interfaces between the service provider and its clients, its subcontractors, its service partners, and its recipients if they are separate from clients.   

An interface constraint applies when something at an interface prevents the provider from delivering more service, and thereby prevents the client or recipients from consuming more service than they otherwise would.  

Regardless of whether they are implicit or explicit, service level constraints govern how, when, and where services are delivered. They are, therefore, policy constraints—but a different kind of policy constraint from those that TOCG typically addresses. 

Service level constraints define acceptable ranges for speed, cost, quality, and volume. Service optimization depends on adaptability more than predictability. Service flow, not capacity, is balanced. Resource utilization is maximized at constraints, but not necessarily everywhere else. Service constraints are strategically located rather than allowed to form anywhere. Finally, work-in-process accumulates mainly ahead of the constraint, not just anywhere.  

In DBRS, the buffer contains the total amount of work ahead of the constraint, which is more than just the work queued immediately ahead of the constraint. That buffer is bidirectional: Either red zone may trigger a change in capacity. The constraint rope governs changes in capacity at the constraint. The service level rope governs work on the constraint. The buffer is often measured in items, the constraint rope is measured in items per period (flow rate), and the service level rope is measured according to the SLA.

DBRS can thus be implemented as follows:

• The buffer can be monitored by counting incoming SR, XA, and CF and then subtracting items closed on or before the constraint, as well as items diverted around the constraint.
• Changes in the flow rate needed to achieve service levels are translated into resource levels during capacity management.
• Work on the constraint is not necessarily done first come, first served (FCFS). Priorities and due dates stipulated by the SLA may affect the order in which items are completed.

In DBRG, capacity is relatively rigid, while in DBRS, capacity is more elastic. Thus, DBRG controls a manufacturing process via buffer management, while DBRS controls a business process via capacity management. when capacity is stable, the backlog is dynamic. when capacity is stable, there's a trade-off between reduced backlog and consistently high utilization. with elastic capacity there is no trade-off between reduced backlog and consistently high utilization: It's possible to have both at once without compromise.

Thus, depending on the SLA, some scenarios might not produce cycle times short enough for services on demand. Moreover, when the constraint shifts between internal and external, elastic capacity outperforms rigid capacity on every measure.

Scheduling covers normal working hours, overtime hours, days off, and on-call hours for existing resources. It thus handles predictable short-term demand patterns.

Adjustment covers demand outside the normal range by adding or cutting time worked by some resources as needed to respond to exploding or vanishing backlogs. It thus handles unpredictable short-term demand patterns.

Deflection covers extraordinary demand that even adjustments cannot handle:

• Channel shifting provides service through different means, such as using a kiosk or IVR system instead of a live agent to give or get routine information.
• Space shifting provides service from an alternative location, such as a remote service center with available capacity instead of a local center that's overloaded.
• Time shifting provides service at another time, such as callbacks that convert current inbound calls into later outbound calls.

Deflection is thus an alternative to queuing.

The DBRG section in Figure 6-8 illustrates two rules of thumb that yield approximately the same result. One rule of thumb is to initialize the buffer at half the original lead time ahead of the constraint because most of it is queue time that will disappear under DBRG. An alternative rule of thumb is to initialize the buffer at three times the service time of activities preceding the constraint. In either case, the resulting buffer size is the length of the constraint rope that governs release of work into production.

Alas, there are no rules of thumb for buffer sizing in DBRS. There are, however, guidelines. Unlike DBRG, where the length of the constraint rope is constant, the length of the constraint rope in DBRS is variable within limits.   

• The first step is to understand how increases in the buffer level affect cycle time.
• The second step is to understand the effect on utilization.
• The third step is to understand the effect on quality.
• The fourth step in buffer sizing is to identify where the service level agreement for cycle time (SLAT) meets the cycle time curve, because that often determines the upper threshold on the buffer.
• The fifth step is to identify where the service level agreement for utilization (SLAU) meets the utilization curve, because that determines the lower threshold on the buffer.
• The sixth and final step is to identify where the service level agreement for quality (SLAQ) meets the quality curve, because that determines the upper threshold if quality dominates cycle time.

These upper and lower thresholds thus define a target buffer range rather than a single buffer size.

The bidirectional buffer of DBRS, however, has two red zones: The low one triggers a decrease in capacity, while the high one triggers an increase. Using buffer thresholds from the preceding section to set DBRS buffer zones directly would not be appropriate because the process would already be missing the SLA by the time the buffer level strayed into a red zone. Instead, buffer zones should be set where they will allow sufficient lead time for capacity changes to be enacted before the buffer level passes either threshold. One way to do that is to inset the red zone boundaries relative to the thresholds by the net change in the buffer level expected during time to change capacity.

The insets do not have to be symmetric.

If overall process cycle time is short, and the constraint and CCRs are not used in constant ratios, the changes have to be calculated to provide adequate sprint capacity. Moreover, if the process is lengthy, CCRs ahead of the constraint may need their capacity changed well before the workload reaches the constraint, while CCRs behind the constraint may need their capacity changes delayed until the workload changes have passed the constraint.

Thus, the buffer level ahead of the constraint triggers capacity management, but capacity management applies to the entire process, not just the constraint.

Task 3 is the internal constraint because it has the lowest maximum, as indicated by the solid horizontal line. The dashed horizontal line marks the highest minimum, which helps in identifying tasks that would have excess capacity if the flow rate fell below that amount.

Although the limits may be tightest around the constraint, business processes and the resources who perform them can be quite adaptable:

• Generalist/specialist model—If resources can shift between nonconstrained and constrained tasks, even the constraint can sprint.
• Expert/assistant model—If nonexperts can perform some tasks, experts can concentrate on tasks that do require expertise.   
• Master/apprentice model—If resources with lower skills can be substituted for resources with higher skills, more items can be handled.
• Near/remote model—If resources with critical skills are available remotely, and portions of the process can be performed remotely, that can relieve a constraint at the nearest service center.
• Primary/secondary model—If an interface constraint supersedes an internal constraint, the spare resources are available to be used elsewhere.

DBRG has several methods for finding the current constraint. They can be used separately or together.

• First, insider knowledge relies on past engineering and current experience to focus on the likely constraint.
• Second, direct observation uncovers the constraint by noting which activity or resource type is often working feverishly while others don't have enough work to keep every resource busy.
• Finally, picking an activity or resource and implementing DBRG around it will reveal the true constraint, even if the initial choice is wrong, because the real constraint will stand out as DBRG draws off the excess work-in-process.

These methods for finding the active constraint work for DBRS, too.

Having a strategic constraint means its location is intentional, not accidental. In DBRG, a strategic constraint maximizes net profit in both the short and long term. If the constraint is external, this means choosing an assortment of products that appeal to customers. the constraint is internal, it means refusing some orders, and perhaps dropping some products, because others are more profitable.

• Put the strategic constraint where adjustment time is shortest because this fosters adaptability and shorter cycle time leads to higher service recipient satisfaction.
• Locate the strategic constraint where it creates the widest buffer range associated with dynamic stability because this minimizes capacity changes—and their cost.
• Set the strategic constraint where the process yields highest quality because that's the client's overriding concern.   
• Leave the strategic constraint where additional resources are not generally available, but make sure all nonstrategic resources fully support those strategic resources by off-loading all nonessential work.

In a process managed toward balanced capacity, there can be a trade-off between predictability and performance:

The primary impact of DBRS is to make a process more adaptable to unpredictable demand while at the same time reducing its overall cycle time, thereby resolving the conflict.

• Whereas the balanced capacity approach pushes items through processes for utilization, DBRS pulls items through processes to achieve SLAs.
• DBRS is accommodating, which is essential for delivering services on demand.
• DBRS is scalable. Indeed, the more clients and processes the better, because aggregation smoothes demand and shared resources create flexibility.
• DBRS and DBRG are compatible. A single enterprise can use them simultaneously if it does both manufacturing and service. And pure service providers can use DBRG for services as available while at the same time using DBRS for services on demand.

 

7 Finance and Accounting

Capacity versus utilization is a classic services conflict.

Local versus global optimization is a conflict that many enterprises, including service providers, don't even realize they have. Cost versus revenue is another classic services conflict. Investment versus delivery is a conflict often seen in PSTS because the same resources may produce intellectual capital and assets as well as deliver services based on them. Asset-based versus labor-based services is a relatively recent conflict to emerge in the PSTS sector. Troubled projects versus risk-taking is a perennial services conflict. Role conflicts are yet another classic conflict. Finally, services as available versus services on demand is an emerging services conflict.

Standard cost times activity units per service, plus margin and contingency, equals price per service.

Expense control is still important, yet not the top priority, because an enterprise dominated by expense control cannot grow as fast or as far as one pursuing T. Like previous forms, TAS is in no way a substitute for financial reporting according to generally accepted accounting principles (GAAP). It is, however, an alternative to CA and ABC for management decision-making.

• Throughput per hour: T/h = (Revenue – TVC) / Productive hours
• Operating Expense per hour: OE/h = (Direct labor + SG&A) / Available hours

And just as T/h is not a billing rate, OE/h is not a cost rate.

For major decisions, T/CU is an inadequate measure for optimizing the enterprise because it assumes a stable constraint and defined service types. Therefore, the following decision-support measures apply to TAS as well as TAG:

• Change in Net Profit: DNP = DT – DOE
• Payback: PB = DNP / DI

Control measures show whether projects, processes, and resources are deviating from desired results. The following control measures apply in TAS:

• Project or Process Dollars per Day: PDD = NPp / Days
• Resource Dollars per Day: RDD = Excess resources x OE / Day

In general, the higher the value of PDD, the better. Yet PDD can be negative if a project is unprofitable. PDD is based on NPp rather than Tp to encourage delivery within budget as well as on time. Client satisfaction is usually measured too, so gains in PDD will not be achieved by sacrificing quality.

• PDD can measure an individual contract or a set of contracts.
• PDD can measure contracts over their entire duration or just a specific interval of time.
• PDD can measure completed, active, and planned contracts so long as the numerator and denominator are consistent.

baseline (initial value) and benchmark (target value) can be helpful in judging progress on PDD.

Ideally, RDD should be zero, so the definition of "excess resources" is crucial.

Among all measures comprising TAG, TDD and IDD are arguably the least used because their quantities are not dollars and not days, but the product of dollars times days. The quantities thus grow rapidly with increments in either dollars or days in order to stimulate an urgent response, but the results can become so big so fast that they lose credibility with managers and practitioners. In contrast, the corresponding measures in TAS, PDD and RDD produce quantities that are just dollars. They grow more slowly with increments in either dollars or the number of days over which they are accumulated, yet managers and practitioners often have an intuitive understanding of and reasonable comfort level with what these TAS measures say about progress toward the goal.

The service mix found by TAS using T/FTEC is mathematically optimal, yet it is found without solving an optimization model. Conversely, the service mix found with CA is not optimal because neither service line NP nor margin truly indicates which service types contribute most to total NP when the constraint is internal.

TAS is simpler than CA because it does not require data and procedures for cost allocation.

• TAS accommodates both services as available, which are subject to internal constraints, and services on demand, which are subject to external and interface constraints.
• TAG and TAS are compatible. A single enterprise can use them both if it does manufacturing and services. For that matter, an enterprise can also use TA for Software.
• TAS shapes investment strategies by guiding major decisions that would shift the constraint or forfeit some Throughput on existing services to gain more Throughput on new services.
• TAS also tracks investment performance without distortions from cost allocation.
• TAS helps select the optimal service mix and tune innovative service types.

Whereas TOCS applications covered in previous chapters turn push into pull, TAS tells the enterprise what to pull.

   

8 Marketing and Sales

Pricing services from standard costs, as described in the previous chapter, is the preeminent example of a provider-imposed policy constraint.

Most customers were satisfied with standard products a generation ago, but today most are not. clients increasingly choose services on demand over services as available. Providers therefore have to match their service offerings to their preferred market segments and then tailor them as needed for specific clients.

Most customers were profitable a generation ago, but today most are not. As more segments emerge and each segment includes fewer customers, economic lifetimes of products and services get shorter.

Standard pricing divides any market into three parts: customers who won't buy, those who will buy, and those who would pay more than the standard price.

Getting opportunities into the pipeline is marketing; getting them out is sales.

the client's constraint is a key ingredient in what makes the TOC approach to marketing and sales different.

TOC is not a substitute for marketing and sales skills, but it generally leads to different decisions.

• First, service pricing based on Throughput Accounting (TA) information links the provider's pricing to its own enterprise goal during competitive bidding.
• Second, focusing the proposed services on one or more of the client's core problems ties those services to opportunities to add large amounts of client business value.
• Third, compelling offers not only produce a bigger-than-expected leap in the client's business value, but they also tie the service provider's return to the magnitude of that leap.
• Fourth, market segmentation recognizes that a given service does not necessarily generate the same business value for every client and therefore deserves to be marketed and delivered differently in each segment—yet treating every client as its own microsegment may not be optimal either.
• Finally, effective sales management winnows out the undesirable opportunities and prioritizes the sales effort on the rest.

Under TA, the provider's price is either what the client is willing to pay or something less than what its competitor is willing to bid, whichever is lower. The key to understanding this outcome is to recognize that CA applies standard costs as pricing input, while TA accepts T/h as pricing output. And where CA uses target profit margin to decide whether an individual bid is acceptable, TA uses NP to decide whether the enterprise is being optimized. Thus, CA effectively trims away potential profits both below and above its target, while TA can harvest whatever profits the market currently supports. Because different clients may receive different business value from projects with equivalent scope, what they're willing to pay varies—but TA's ability to harvest the full profits available at any price point means it can generate more NP than CA can in both weak and strong markets.

TOC-based service pricing is based on the client's business value, not the provider's cost. The most direct measure of business value, however, is DNP because it relates directly to the client's goal: to make money now and in the future. the only way to increase Throughput is to help the client get more from its constraint, so helping the client identify and manage its constraint is essential.

 

Nowhere is there any mention of the service provider's cost. The price in every scenario is based on what the client is willing to pay, not what it costs the provider to deliver the services. To get there, however, the client has to get past its policy constraints that say cost reduction is the primary objective and the best deal is the lowest price from competitive bidding. This is where the service provider's marketing and sales efforts come in.

Every TOC-based marketing program is unique because it must solve one or more of the clients' core problems.

The most compelling service offers thus increase the client's T rather than decrease its OE, which is the opposite of what many conventional service offers attempt. Moreover, a compelling service offer ties the service provider's return to the magnitude of the client's increase in business value. Consequently, the service provider following TOC-based marketing and sales in the previous section would not bid a fixed price for any of the scenarios, but would instead make its price contingent on the actual business value the client receives.

Although the provider's price in M&SS is based on the client's business value, that price is not the maximum amount the client might be willing to pay. The price is instead a minor slice of the increase in business value, so the client captures the majority of incremental business value. In this way, the client always feels that the value it receives more than justifies the price it pays.

In a properly segmented market, the price and quantity of goods or services sold in one segment have no effect on what can be sold in another segment.

The TOC market-segmentation principle is to segment the market, not resources, in order for the same resources to then serve more than one segment. If demand across segments is uncorrelated, aggregate demand is nevertheless smoother. And if segments are countercyclical, aggregate demand can be much smoother.

The solution, of course, is to standardize what can be standardized and customize only what remains.

The challenge for PSTS enterprises is to leverage their expertise and intellectual capital so that competitors find it prohibitively expensive to enter or remain in the targeted market segments.

First, for a prospective investment in segmentation, a provider can examine its own potential DNP and payback.   

Second, the service mix decision framework illustrated in Chapter 7, "Finance and Accounting," can be extended to handle segmented markets.

The winner's curse occurs when a low bid wins and the service provider finds the deal unprofitable. Buyer's remorse occurs when the winning provider is unable to meet the client's expectations. They often are opposite sides of the same coin.

• There is no standard TOC application for marketing and sales, because every marketing program is unique, but TOC principles are nevertheless the foundation of M&SS.
• M&SS bases market offers on client value, not provider cost. If the client is an individual rather than a business, personal value is analogous to business value, and M&SS still applies.
• M&SS drives sales via constraint management for the client, which usually requires the service provider to isolate core problems rather than address all the client's pain points at once.
• M&SS relies on DBR, Replenishment, and CC for compelling market offers—plus TA for sales management.

The tipping point in TOC-based Marketing and Sales for Services (M&SS) is reached when the client refocuses on driving up its own business value dramatically rather than driving its service provider's price marginally.

Price is meaningful only in the context of value, and the value of services varies enormously.

 

9 Strategy and Change

Strategy is the specific way an enterprise chooses to pursue its goal, and change is the way it realigns its marketing, sales, and production to carry out that strategy.

some industries are chronically plagued with excess capacity, which means their constraint is external, in the market. In such industries, executives may feel intense pressure to adopt the strategy of competing on price. The consequences are well-known, however. Competitors match price cuts eventually, if not immediately. Employees may feel their jobs are less secure, which affects their productivity and turnover. Suppliers are squeezed to cut their prices, which may reduce their reliability or quality. Finally, customers recognize that price cuts often require a compromise on quality.

Segmenting a market in order to satisfy unmet customer needs is a traditional alternative. So are competing on speed and reliability. A modern alternative is to compete on innovation. A goods producer that leads its supply chain to adopt an innovative product, operation, or business model can change the game in its entire industry.

TOC approach to strategy frequently hinges on policy constraints. When an enterprise truly breaks a policy constraint, it's often hard for competitors to break their comparable constraint.

A policy constraint is a useful strategic constraint if it keeps the enterprise from overreaching. not all policy constraints deserve to be broken at once. When an enterprise's strategy changes, its strategic constraints frequently have to change, too. As strategic constraints shift, investments may have to shift as well to accomplish the restaffing, rescheduling, and reconfiguring.

TOC approach to changing:

• The natural inclination of change agents is to skip the first two phases and rush into presenting the solution. To decision-makers left unprepared by the skipped phases, however, it can look like a solution in search of a problem.  
• Each phase of buy-in addresses a specific issue that can derail change. Thus, no phases can be skipped or completed out of order.
• Adequate preparation for the first three phases can easily take several days because various diagrams must be prepared,   
• If limited time is available for presentation, change agents should do the first two phases well, and then stop. Successfully completing those phases will leave decision-makers wanting more time later to hear the solution and discuss how it can be implemented.

Without TOC, salespeople have the same inclination as change agents: They want to present their product or service as soon as possible. Unfortunately, presenting it too early leads to objections, and the more objections customers raise, the less likely salespeople are to close sales. For their part, customers are frequently disinterested, suspicious, or perhaps even antagonistic because their previous experiences with salespeople have not been altogether positive. So even when customers perceive value in a product or service, they may resist buying it.

• Despite a seller's enthusiasm, buyers don't automatically see the value of products or services. Thus, the seller must build rapport with buyers by using the materials prepared by marketing to confirm and discuss the undesirable effects the buyer feels.
• Buyers seldom believe that sellers really understand their problems. Thus, the seller must set the stage for the offer by using the materials prepared by marketing to confirm and discuss the core problems that lead to the buyers' undesirable effects.
• The higher the value of the product or service, the more anxious sellers are to sell, but the more wary buyers become. Thus, the seller must bring a buyer to want the product or service by showing how it solves the buyer's core problems.
• Buyers are naturally suspicious if an offer seems too good to be true. Thus, the seller must show the buyers what benefits the seller derives.
• When buyers can see why the sale is good for both parties, their resistance diminishes. Then the seller can show how the product or service delivers more value to the buyer.
• Finally, once a buyer is convinced, there may still be obstacles, such as approvals. Thus, the seller must use the list of potential obstacles prepared by marketing and work with the buyer to overcome those that apply.

The TOC approach to sales works only if marketing has created a compelling offer. Of course, even a compelling offer still has to be sold. TOC says that every improvement is a change, but not every change is an improvement.

High-performing enterprises tend to adopt innovative business models, and TOC supports this in several ways.  

First, Throughput Accounting changes the financial model of the enterprise. Second, the TOC approach to strategy employs compelling market offers that differentiate a service provider from its competitors. Third, compelling market offers increase the value of services, so price cutting is no longer a relevant strategy. Fourth, if the market offer is constructed by breaking multiple policy constraints within the service provider's enterprise, it takes longer for competitors to catch up strategically. Finally, if the service provider's strategy is implemented by alleviating interface constraints, its suppliers and clients benefit, too.

 

10 Implementation and Technology

the two primary sources of TOC-compatible systems are software vendors and consultants. Frequently, either software comes with consulting, or consulting comes with software.

TOC has relied on technology from its inception, but TOC has also been at odds with technology. For example, TOCG opposes Enterprise Resource Planning (ERP) if it ignores the manufacturing constraint, as software packages do when they assume infinite capacity. On the other hand, TOCG relies on software to implement Drum-Buffer-Rope (DBR).

If technology doesn't take the enterprise toward its goal, it shouldn't be implemented.

Opinion leaders are vital because each is at the hub of his or her own interpersonal communication network that spreads the word for or against particular innovations. a typical innovation adoption curve. The S-shape of this curve describes how innovations spread, both within a single enterprise and across multiple enterprises, though the time scale is shorter and the slope is steeper for a single enterprise.

• Kicking off the Initiation stage, a small set of adopters (about 2.5 percent) known as innovators buy in to the innovation. Innovators include decision-makers as well as practitioners who do pilot implementations. Their judgment and experience demonstrate the practicality and value of the innovation, but innovators are rarely opinion leaders.
• Wrapping up the Initiation stage, early adopters (about 13.5 percent) put the innovation to work on a small scale.  More importantly, however, early adopters include opinion leaders whose views are broadly influential.
• If the Initiation stage is successful, the early majority (34 percent) adopt the innovation during the Early Growth stage, the late majority (34 percent) adopt it during the Late Growth stage, and laggards (16 percent) may never adopt it unless forced to do so during the Maturity stage.

Innovation adoption thus reaches critical mass—becomes self-sustaining—when 10 to 25 percent of potential adopters accept it. This range is where the S-curve begins sweeping rapidly upward, and it agrees with the TOC rule of thumb that about 15 percent of an enterprise's employees have to buy in to TOC for it to take root. The informal buzz generated by opinion leaders is more persuasive than any mass media or management communication.

 

11 Conclusion

When clients can choose from multiple service providers, those who deliver services on demand can have advantages in speed, quality, and value. leverage points—where a modest change in constraint management can have a major effect on Throughput.

There are, however, significant differences between supply chains and service chains:

• Service chains more often involve clients as active participants in production.
• Production in a service chain occurs more often at the client site.
• The specific nature of services for each client is determined closer to delivery time.
• Service chains manage perishable resources because services cannot be stockpiled.
• Service chains strive for bidirectional optimization, which means doing what's best for clients while at the same time doing what's best for the providers.

Service providers can take over the client's software development and maintenance, data center and network operations, desk-side support, help desks, and other IT functions. Likewise, business functions such as human resources, customer relationship management, finance and administration, and procurement can be outsourced.

Distribution services provides a classic example of insourcing. The service provider becomes seamlessly integrated into the client's core processes because it can perform some activities significantly better than the client can itself. The client continues to run its own enterprise while monitoring the insourcer's service level performance and its impact. For instance, if the on-time delivery rate rises while inventory drops, the insourcer is creating real value for the client and its customers.

Limitations of TOC:

• First, TOCS does not have an application for services delivered without a defined project or process.
• Second, TOCS does not say much about development of new service offerings.
• Third, TOCS does not say much about quality assurance and risk management.
• Fourth, TOCS cannot create absolute stability when clients need services on demand, but it can create the flexibility necessary to deliver services on demand.
• Fifth, TOCS can be difficult to implement for many reasons, including external and interface constraints that refuse to be managed.
• Finally, TOC in general is optimistic regarding change.