The Magic Behind Planning and Executing (Optimal) Service Supply Chains - Part 2

Part 1 of this series established that service supply chains have many planning levels and time horizons, which can be roughly divided into the following: the immediate period around the day of service, and forecasting and planning for the day of service. My post then expanded on the various approaches to tame the challenges on the actual day of service.

The most advanced approach is to use service chain optimization algorithms that balance customer satisfaction and operating costs. But, as in any business environment, the catch is that conflicting forces pull any company’s service scheduling optimization decisions in opposing directions.

ClickSoftware Technologies’ recently published book  “Service Chain Optimization for Dummies” depicts the following usual conflict that seems all too familiar:

  • The Vice President (VP) of Service wants to maximize customer satisfaction

  • The Chief Financial Officer (CFO) wants to reduce operating costs and improve profitability

  • The Business Development Director and Chief Marketing Officer (CMO) want to increase market share and penetrate new markets

  • Service employees want to work reasonable hours and not be overloaded with overtime or excessive travel

  • The Regulatory Body expects the service company to meet the minimum requirements for operating in its (perhaps hazardous) industry; think of the so-called Nuclear Fatigue Rule (NFR) or the mandate in certain regions that a utility crew has to be at the suspected gas leak place within, say, 20 minutes

Like any other company, most service businesses have different operating divisions with separate optimization requirements. Achieving one division’s goal can come at the expense of achieving another. For example, a goal to reduce costs can significantly impact the goal of maintaining high service level agreements (SLAs) and customer retention. Top executives must understand how best to prioritize and determine which holistic goals they cannot compromise.

Workforce size, demand volume and intensity, job types, and rules for scheduling define how flexible the “rubber band” is for a company (and the range of stretching in each division’s direction). Ultimately, the key is to define what the corporate business policy is for prioritizing jobs.

Enter the Service Policy

The service policy is the company’s written definition of the scheduling criteria for its business, which is then configured inside the service chain optimization solution. The policy embodies the overarching goals of the business, taking into account the fact that the abovementioned different stakeholders have different goals.

There is no right or wrong service policy by default, just the one that best suits the particular business. The policy communicates to the dispatch team and the optimization algorithms (within the software package) what they need to achieve, along with the business rules and objectives for getting there.

Needless to say, a computer's intelligence is artificial. All these machines do is practice the instituted service policy in an optimal and consistent way, on a speed and scale that no human can match. The typical result is a significant improvement in the business metrics that the executive team have decided to focus on.

Defining the service policy gives a company an opportunity to rethink how it has been running its business. Trying to reproduce manual working practices, results, and performance (à la “We’ve been doing it this way for ages…” attitude) inside optimization software is risky and can lead to implementation and user acceptance challenges.

The service policy directs the service chain optimization algorithms in terms of knowing what is (or is not) allowed and the overall business goals. In order to be able to resolve many conflicting pressures, the service policy relies on business rules and business objectives.

Ruling the Service Business

The business rules direct the service policy so that dispatchers understand what decisions can be justified in their service business. The rules provide the boundaries of the optimization, since automated and optimized decisions must comply with these rules.

Many business rules are standard and are used across different businesses, and across different industries. Below are several examples of common business rules:

  • Skills -- Only assign a resource who has the necessary skills and qualifications for the job

  • Travel Distance -- Place a maximum distance that a resource can travel from home or between jobs

  • Response Time -- Honor the SLA or appointment time, and if the resource can’t get there in time, automatically assign an alternative

  • Availability -- Only assign resources that are available and working while considering their prior commitments such as doctor’s appointmentss or vacation time

  • Parts and Equipment -- Only assign a resource who has the necessary parts and equipment available, or who can collect these while en-route

  • Sequencing -- Assign jobs that have multiple stages and multiple resources (crews) in the correct order so that they follow the right sequential path

One should not expect perfection all the time even with rules-based optimization. There will always be times when the service manager cannot schedule a job due to unavoidable rule violations. For example, there may simply be nobody available with the right skills at the right time.

Still, the role of dispatch teams is hereby alleviated because of service chain optimization. Namely, instead of building a full schedule, now service managers have time to focus on the more challenging exceptions – i.e., those jobs that cannot be scheduled automatically according to the company’s standard service policy.

The set of “do” and “do not” rules and common decisions within the service policy determine the legitimacy of potential schedules. Service optimization is the process of finding the best schedule from the legitimate options, and one cannot find the best schedule from business rules alone.

Business Objectives Complement Rules

Working alongside the business rules are the strategic business objectives that define the overall goals of the service business and denote those all-too-often conflicting departmental goals with their relative levels of importance. This is where the algorithms come in handy to create the best schedule to meet the business requirements by considering allof the business rules and objectives simultaneously, while also finding a balance between conflicting business goals.

Optimizing using business objectives is all about finding the balance. Any company could certainly reduce response times by having more idle resources waiting for the next job to appear or by making resources drive further. But both of these options have ramifications in that they decrease utilization or raise travel distances. Any astute service manager should know what is best for his/her customers and how to achieve the balance via the service policy’s business rules and objectives.

Business objectives are common across many service businesses and work types, but they are assigned varying levelsof importance based on the company’s service policy. Following is a list of the more common business objectives:

  • Travel -- Minimize the total travel across the entire schedule by reducing travel from home and between jobs

  • Response Times -- Minimize overall response times to customers (or to certain types of customer) by arriving onsite faster

  • Overtime -- Minimize the use of expensive overtime, which services customers late in the day

  • Priorities -- Try to assign higher priority jobs first (although one can assign a lower priority job en-route if it still meets response time commitments for both)

  • Financial viability -- Create the optimal schedule in terms of costs and revenues (i.e., maximize revenue, while minimizing costs)

Despite their usefulness in the background, service chain optimization algorithms lack interpersonal skills, emotions, sensitivity, empathy, intuition, and sense of humor. This is where dispatchers can really shine, and people still have a vital role in making decisions.

In times of crisis, service companies must ensure that their dispatchers are still helpful and friendly to make a great impression on their customers. In fact, companies should solicit help from their dispatch team in defining their service policy, since doing so helps to reduce any staff resistance later on.

Achieving Excellence on the Day of Service via Optimized Schedules

The actual day of service is where everything happens (or where "the rubber meets the road" literally): resources are scheduled to their jobs, most appointments will be honored (but some may be missed), some customers will cancel, traffic jams or road constructions will impede progress, new emergency jobs will pop up, some jobs may take more (or less) time than expected, some resources may call in sick, and many more issues can happen that affect the schedule and how the service business performs that day. Service execution optimization is not just about improving decision-making on the fly, but is also about producing operational effectiveness by reducing unneeded paperwork and repeated voice communications between the dispatch team, resources, and customers, and reducing the overall number of changes in the schedule.

At the end of the day, all of the theoretically available service hours will vanish and the service company’s success will be  governed by how productive its resources were during that time. As said in my “Navigating Between Service Management Scylla & Charybdis” series, one cannot reutilize unproductive time nor save (store) any idle time to use in the future.

Service chain optimization around the day of service helps to achieve this productivity goal by creating and managing an optimized schedule according to the W-6 principle from Part 1 (i.e., Who, does What, with What, When, Where, and for Whom?).  Logically, one should begin by creating an optimized schedule.

To that end, the service jobs scheduled for a certain day did not all necessarily appear that morning or at the same time. Namely, some jobs can be planned installations for new customers and they may have appeared last week. Others will be scheduled periodical maintenance jobs, which the dispatcher knew about quite some time ago. And then there are the emergency repair jobs that arrived yesterday or possibly even today.

Therefore, the initial schedule is typically formed several days before the day of service but in a way that gives the service manager freedom and flexibility to reshuffle as new jobs appear. This so-called morning reshuffling makes further improvements in skills matching, responses times, and travel reductions. The reshuffling effort is all aimed at maximizing the quality of the schedule and squeezing the most productivity out of the company’s available resources.

In the optimized world, shuffling and reshuffling to seek an optimal schedule is way beyond the capabilities of the human brain, even for just a small group of resources. The abovementioned service policy is the business logic that enables the computer to know what is good for the particular service business and the rules by which to schedule.

Booking Appointments Dynamically

As mentioned in my series entitled “Navigating Between Service Management Scylla & Charybdis,” appointment booking is the process of agreeing on a service window with a customer, usually several days before the day of service. Most of us have experienced the morning, afternoon, or all-day windows and the frustrating wait at home that this practice brings.

Most manual processes, and even many service scheduling applications, offer customers appointments based on a number of pre-determined slots for a geographic area, product type, or time window. These slots are arbitrarily allotted to customers as they call, until they are all filled, typically in a simple first-come, first-served (FCFS) approach.

However, this traditional approach ignores important factors such as the location of previous jobs or the true availability of individual resources and their skills. This scheduling vacuum typically leads to inefficiency, ineffectiveness, and poor utilization. Imagine two customers living on the same street who select appointments on the same day but for many hours apart. The time and distance incurred from the excess travel hampers the performance of the service business. If the appointment slots do not consider the varying durations of different types of work, the dispatcher ends up with increases in dreaded late arrivals and resource idle time.

Conversely, when appointment booking is optimized, service businesses identify the criteria that are important to them through their service policy. This may be providing the shortest route between jobs, while still offering a time that is convenient to the customer and that also maximizes filed resource efficiency. Appointment optimization considers the existing workload versus resource capacity in terms of geography, skills, and time.

Moreover, appointment optimization avoids any predefined estimates, which are usually inaccurate travel and labor time “averages” that are based on past experiences (and not on what is really happening now). The appointment schedules are re-optimized throughout the day creating further improvements in resource utilization while maintaining customer commitments and service levels.

In other words, as resource availability changes (customers cancel appointments or new ones appear), dynamic optimization ensures that appointments are offered and delivered in the most efficient way possible. Customers can often be just as pleased with an offered time that is the most cost-efficient for the service provider as long as its commitment is reliable. The service company can always offer higher-cost options if asked for, and sometimes at a premium for convenience.

Reaching Destinations with Street-level Routing

Travel optimization plays a vital role in creating the optimized schedule, regardless of whether this is appointment booking, automatic assignment, rescheduling, emergency jobs taking immediate priority, or anything else. Although no one can drive to any destination in a straight line (at least flying service teams are still prohibitively expensive in commercial sectors), surprisingly, some approaches to scheduling still consider travel “as the crow flies” between jobs based on postal ZIP codes or on pre-defined travel times between service areas.

While this approach may seem acceptable, the benefit is superficial because it ignores vital details that adversely affect field resource arrival times. Just think of possible effects on planned travel times from ignoring bridges, one-way streets, rivers, and other obstacles that get in the way of field resources on their way to customers.

Linear travel estimations cause excessive travel costs and harm customer satisfaction because field resources can arrive late (or not at all in the worse-case scenario). And it takes just one deviation from the planned travel time for a chain reaction (the domino effect) to begin and cascade throughout the schedule.

Conversely, street-level routing takes into account obstacles such as rivers, lakes, mountains, bridges, one-way streets and speed limits, giving a more accurate estimation of travel time. By using detailed data from geographic information systems (GISs), all of the possible geographic obstacles are known, meaning that customers should receive reliable appointment windows and commitments.

Without street-level routing, the dispatcher is likely wasting resource efficiency and pushing up operating costs. Letting field resources determine their own routes is also costly because the company is paying a field resource’s labor rate to do the job of a member of the dispatch team. Needless to say, field resources should not be wasting their time and energy on figuring out best routes at the expense of performing valuable customer service.

Furthermore, the optimization must be performed across the entire workforce and not one route at a time. Street-level routing is not just about which resource should travel from Job A to Job B; the key to productivity improvement is about deciding whom to assign to Job B given everything else in the schedule and not just the prior job. Travel time must be part of the scheduling decision and considered concurrently with all other scheduling criteria.

There is a key distinction between “true SLR” and various kinds of “pseudo-SLR” tools, especially the kind where the SLR is done only after the routes are computed using “as the crow flies” travel times, and the SLR results are then used to fix each individual route This caveat is described in ClickSoftware's white paper “Street-level routing: where the rubber meets the road.”

Keeping Up to Date (in Real Time)

At the start of the day of service, the optimized schedule is only an optimized and most efficient starting point. Namely, the very moment the day begins, the schedule can change and if the service manager does not keep on top of the changes the day can quickly unravel. The day of service is full of surprises, and not always pleasant ones.

Don’t “the best laid plans of mice and men always go awry” (as the proverb says), anyway? Below are just a few examples of why keeping the schedule up-to-date is important:

  • Cancellations and postponements: Customers sometimes cancel because their problem is solved, or they move the appointment because they have other commitments. If the change enters the schedule quickly enough and the resource is notified, then the dispatcher can prevent unnecessary travel and utilize the resource on another job instead.

  • Unexpected problems: Scheduling means one needs to estimate the duration of the job but sometimes this takes longer – a stuck component, inefficient resource, or unforeseeably difficult working conditions can all take their toll so that, e.g., one hour job takes two or three hours instead. If the manager does not re-optimize the schedule swiftly enough then some appointments at the end of the day will be missed (resulting in some unhappy customers).

  • Sickness: Resources unexpectedly call in sick or perhaps fall ill during the day. One must then re-optimize the schedule to reallocate the work to maximize the number of jobs the other resources can complete.

  • Accidents and delays: The geographic nature of field service means that traffic delays due to construction works or accidents are inevitable – these may result in a temporary delay or remove the resource completely. Again, re-optimization manages the impact, since real-time tracking using location-based services, such as the global positioning system (GPS), can have a resource rerouted to a clearer route to save time.

  • The customer is not home: The perennial problem – the resource arrives on time but the customer is not in, resulting in lost time, lost productivity, and the need to revisit on another day.

All of these sudden and unexpected changes need quick attention. Delays in reacting, no matter how short, lose valuable productivity and efficiency. Real-time tracking is self-explanatory: the change occurs, and the schedule should be immediately re-optimized to ensure the maximum possible benefit.

For the unprepared business, unpredictability leads to missed appointments, idle time, increased costs, and panic. One should recognize that achieving total predictability in field service is impossible. Thus, service companies should focus their efforts on being proactive and better managing the unpredictability. Decision support and optimization software for service delivery execution are the cornerstones of many leading service organizations.

Software agents constantly “listen” to the stream of incoming information from the entire system including new emergency jobs, jobs that take longer or shorter than planned, or a technician who is stuck in traffic. Algorithms process that information against a broad set of variables and business rules such as technician skill set, geographic region, tools on the truck, etc. and determine how to keep the schedule continually optimized throughout the day and across the entire enterprise.

Keeping the Lines of Communication Open: A Must

An important part of keeping the schedule up to date is communication, not only with field resources, but also with customers. As also mentioned in my series entitled “Navigating Between Service Management Scylla & Charybdis,” mobile communications technology has made the benefits of service chain optimization even easier to achieve. Gone are the days of “pedestrian” fixed-line telephone calls from dispatchers to resources advising them of the schedule, and of any changes.

And particularly gone are the days when paper job sheets were used. Today’s optimized service business is hi-tech, pushing all job information to mobile devices instantly. The communication also travels both ways. On one hand, filed resources update and close their jobs on their mobile device, and provide updates on their progress so that dispatchers can spot any delays earlier. On the other hand, customers can send location-aware tweets about the equipment problems and give electronic signatures, resources can issue receipts, and the dispatcher can rearrange appointments – all using the field resource’s mobile device.

To receive service the customer usually needs to be present. However, customers’ time consumption is one of the most critical metrics for customer satisfaction, so minimizing it is of the utmost importance. We have all been there – waiting at home for our appointment and then, usually towards the end of the long day, we would receive the dreaded call saying that the service business cannot make it to us today. How frustrating is that? Couldn’t we have been told much earlier in the day, when the problem became apparent?

Today’s optimized service business sends (“pushes”) text messages or voice messages with expected arrival times during the day. If customers want an update, they can go online and check out (“pull”) the Web site. Or they can log in for an update using their mobile phone’s Web browser.

Service chain optimization means regular and timely automated communication with the customer. By automatically updating customers on the expected arrival time of their resource, service companies can provide a commitment time that is wide enough to maximize the efficiency of their service business. Additionally, this minimizes the amount of time that the customers must be available because they know when the technician is due.

The real-time service enterprise involves updating the schedule, re-optimizing, managing the changes, communicating with resources, and communicating with customers. If it misses any of these steps, the service company will not be operating in real time (i.e., it will be hampering possible business benefits and not achieving service excellence).

Nothing Without an Illustration

By integrating real-time field data with automatic decision-making algorithms, service companies can gain tremendous improvements in resource productivity and customer responsiveness. All data from field devices should be analyzed by intelligent optimization processes as they come in, so that optimized decisions can be generated in real-time.

Mobile communication is essential even if the company uses subcontractors. Mobile data, rather than speech, is the most efficient way of communicating. It may sound impersonal, but time will likely be wasted via using paper or voice means of communications.

In his 2007 guest article in Field Technologies Online entitled “The Age of Real-Time Service Enterprise” Dr. Moshe BenBassat, ClickSoftware’s founder, current chairman and CEO, provides a juxtaposing example. The first situation is with a field technician (T1) who is already 40 minutes late leaving the first job of the day at customer (C1), jeopardizing the next job’s on-time arrival. Focusing on the job at hand, he/she forgets to report that he/she is late. By the time he/she does report his/her status, critical opportunity is lost.

Imagine now that the field force of the company is equipped with GPS and handheld mobile data devices. Imagine again that a real-time automatic optimization algorithm is continually watching the location of all technicians and is “listening” to the data streams that arrive from the technicians’ mobile devices. Noticing that T1’s location continues to be at C1 beyond the time the technician was expected to finish, the algorithm will monitor the delay and, at a certain point, will look for a resolution to recover from the chain effect of this delay.

It could very well be that not too far from T1, another technician (T2) reports that he completed his first job earlier than expected. If he/she has the right skills, the algorithm will evaluate the possibility of switching T1’s second job to T2. If need be, the algorithm will also inform a customer service representative to notify customers about late arrival or possibly a postponement until the next day.

As a conclusion of this blog post: the day of service is of major importance, but it cannot be managed in isolation. While a great amount of variation and unpredictability occurs during the day of service requiring real-time management, interestingly, any service business success is also governed by what happens in the lead-up to the day.

The final part of this series will analyze what happens much before the day of service, and not just on that day. In the meantime, please send us your comments, opinions, etc. We would certainly be interested in your experiences with this software category (if you are an existing user) or in your general interest to evaluate these solutions as prospective customers.
comments powered by Disqus