Computer Modeling in the Services Environment ©

by
Roy Sequeira
Published in the AFSMI Professional Journal, June 2001

Computer modeling is a powerful but underutilized tool in the service manager’s toolbox. Models can help the service manager or executive achieve or surpass objectives with optimal resource deployment. While other corporate functions have used modeling extensively in forecasting and decision support, the service function has been notably slow in adopting this versatile tool. This may stem from the traditional accounting, manufacturing, or marketing views of business that deal with tangibles and hard numbers. Service, on the other hand, with its focus on such intangibles as customer satisfaction and responsiveness to customer needs, appears reluctant to use hard data to make firm business decisions.

Business was slow in recognizing service as a strategic revenue source in its own right. Traditionally, management viewed service as an unavoidable expense, necessary only because customers would not buy product without it. But the majority of today’s management regards service as a valuable, independent line of business. However, this change in perception has been too recent (with some exceptions) for it to work its way fully into mainstream management research and literature.

Enter: modeling. We in the services industry can do much to improve our service to our customers and our contributions to the bottom line by adopting such tools as modeling. Models allow us to test business strategies and tactical approaches before physical and procedural implementation. Models are a low-risk way to control costs. They may not always eliminate the need for a pilot run, but by identifying and quantifying potential problem areas, models allow us to significantly improve our ability to forecast, reduce risk, and iron out potential problems before they become reality.

How does modeling help services management? In many ways. It helps in the new product introduction (NPI) process by creating firm forecasts of costs, revenues, and resource needs that other corporate functions can substantiate and accept. This allows service to influence product design, manufacturing, and testing to improve serviceability. Models have made it possible to obtain service relief for warranty costs. They can help fine-tune proposed monthly maintenance charge (MMC) changes to balance desired revenue gains against lost maintenance contracts, and they facilitate the partnering of the service function with other corporate functions in achieving corporate objectives. In short, models support services management in its decisions and interactions with other corporate functions.

Model Structure and Use
While modeling is useful in forecasting and in decision support, particularly in NPI, to the same extent it can make a significant difference to the effectiveness of every aspect of the services business. A sound approach to a proposed model is to create stand-alone modules for each service function: call center, field engineering, logistics, repair depot, technical support, and so forth. This enables examination of the most-needed modules first, obtaining full benefit from their functionality, and then moving on to the others. Each of these modules (or sub-models) is a model of a separate function. Each module realistically shows how that function performs. Final results or outputs should be in categories such as currency, labor units, and other pertinent data with timeframes shown where appropriate.

Let’s take a look at a generic model to illustrate the principles involved. We will use a model used in the NPI process as it encompasses every function within the services organization and some external functions as well. All modules will align results with the projected marketing ramp. Forecasted items generally include resource requirements, cash flows, capital expenditures, costs, and revenues.

Each function that affects the product’s serviceability will have a separate module. For example, most engineering-development functions have a module that takes design and component data to compute parameters such as unit costs, mean time between failures (MTBF), and mean time between calls (MTBC).

The Repair Depot module takes this data, together with marketing inputs such as projected volumes and ramps and service data such as target turnaround time, to determine the depot parts and labor requirements. A high-failure item on a high-volume product calls for a different repair strategy than a low-failure-rate, low-volume product. Thus, the depot can cost-effectively plan its needs and tailor its outlays to the product.

The Logistics module will then take this data and combine it with service inputs on contractual latency requirements to determine parts consumption and how and where to deploy them. It also will use guidelines on turnover ratios and other pertinent data to determine stocking levels at each stocking location.

The Call Center, Technical Support, and Field Engineering modules will use mean time to repair (MTTR), MTBF, and MTBC, together with other functional data, to forecast the service load in each of these functions. Combined with the marketing ramp data, they will show resource needs by quarter. Resource needs combined with run rates drive labor costs and other budget items. These modules also can include forecasting equipment and other resource needs.

While each module’s design is stand-alone, there is some interdependency. We initially address this interdependency by careful selection of “best-estimate” data that serves until the appropriate module supplies accurate data.

The power of such a model comes from what it enables the service function to do. Since service is generally the last function before releasing a product for general availability (GA), it comes under tremendous pressure to compromise its values to compensate for problems and slippages in the development phases leading up to GA. By allowing the service function to project the effects and cash value of such compromises, models enable rational decisions across the corporation.

Service modeling affects other functions as well. Manufacturing can use the repair depot and logistics model outputs to determine service impact on its schedules. Engineering can get a better idea of the corporate ROI on proposed engineering change orders (ECOs), field change orders (FCOs), and other design changes. The treasury and finance functions get a good forecast of capital investments and cash flows. Marketing can better gauge how various proposals affect the bottom line.

• How many additional people will we need to support this product over the next year?
• Do we need additional stocking locations?
• How much capital investment will this product need?
• How does this product affect my cash flows and P&L statement?

Consistent and rational decisions. Most service functions do some form of analysis and planning. Unfortunately, because they do not generally do this in a formalized and standard manner, the results usually are not consistent or repeatable. This means management must spend time and effort in scrutinizing the results to ensure valid results. Oftentimes, other corporate functions that do not always relate to the service environment will dispute the results and the requested action. This frequently leads to decision-making at an emotional or irrational level. By providing all that is concerned with accurate, consistent, and reliable data, a good service model enables rational decisions consistent with objectives.

Acceptability and transparency across the corporation. Publishing the service model with its construction, input data requirements, and output results across the corporation ensures that everyone in the corporation realizes the needs, requirements, and time constraints of the service function. There are no surprises. The model provides all functions with a rational basis, enabling them to work smoothly with the service function.

What-ifs. What-if engineering modified the design to improve MTBF by 30 percent for an increased per-unit cost of six percent. The model easily can compute the resultant savings in service expense and compare them to the engineering and manufacturing cost of the design change. By allowing the development team to immediately see the overall impact to the corporate bottom line, the service model reduces cross-functional wrangling and allows optimal decision-making. If a product design results in an unacceptably high MTBF, the model shows the service function the total costs associated with the design and provides the data it needs to persuade those involved in development to improve the design.

Some Real-Life Examples

Computer-generated spreadsheet models have proven their worth in improving serviceability and profits.

In one instance, modeling the service proposal for a new product showed how a $200,000 investment in logistics could avert $4.9 million in international service costs.

To keep inventory levels down, the manufacturing and engineering departments originally wanted one standard keyboard across the world. When the service model showed this would cost the field service unit $4.9 million for on-site intervention to customize keycaps on replacement keyboards in Europe and was able to substantiate its cost estimates, the service function was able to persuade the others to agree to nine separate keyboards in European markets. The ROI on the $200,000 cost for the keyboards was easy to understand and accept.

In another case, modeling showed how an engineering redesign costing $7,500 could avert a $350,000 field spares expense while improving the first-time fix rate.
The engineering department’s redesign of a CPU motherboard included changing the interface connectors. Unfortunately, the new connector was incompatible with the existing one. When the service models pointed out the high costs resulting from this change to those in the repair depot and logistics, engineering agreed to design a connector that allowed compatibility between existing spares and the new design. Again, the model showed how investing $7,500 had a high ROI.

Models have strategic implications as well.
A customer approached a specialized system vendor requesting significant changes to the standard system architecture. Sales potential ran into the millions and caused intense interest. While the requested changes raised some concern, the development and manufacturing units saw no impediments to the sale. However, modeling the service implications showed that the proposed modified design would severely affect service profitability and had a potentially devastating impact on the modified product’s field usability.

The model clearly showed that:

• While maintenance on the modified systems would bring in only 60 percent of the MMC on unmodified systems, total resource usage and costs stayed the same as that required by standard systems. This would result in heavy service losses.
• Unscheduled downtime would be quite high. Adding field labor could not fully compensate for the increased downtime. Together, these effects would have severely affected market perception of the vendor’s product line.

Modeling helped the service function at a network vendor make a significant contribution to helping a new, low-end product increase market share on a new product. Rapid technological change in this field projected a relatively short useful life before obsolescence. The models revealed an insignificant difference in total service costs between a three-year and a lifetime warranty. The marketing and sales departments seized the opportunity to increase sales and increase market penetration by offering a lifetime return-to-factory warranty.

Creating and Deploying Models Successfully
Modeling is still quite new on the services scene. Services management must clearly understand and communicate expectations and prerequisites communicated within the service functions and in interactions with other corporate functions.

For example, questions such as “What is the hourly field labor rate?” can cause considerable discussion and can be resolved only by a clear understanding of what the model design calls for and how the results contribute to the bottom line. If the service function’s finance support unit understands the model’s intent is forecasting incremental service costs, they may use unburdened labor rates, as fully burdened labor rates will doubly allocate service overhead to the product.

Clear, well-communicated data definitions can assure that the data is consistent and available in a timely manner. Documenting the algorithms and equations used and discussing them with other affected functions will reduce challenge and help ensure acceptance. Management across the company must understand how the model works and accept its validity. Presentations and one-on-one meetings will get the message across.

Spreadsheets are generally the vehicle of choice for most models. They are well-understood and accepted across the corporation and enable faster model design and implementation. Skeptics can probe and test the model to confirm its validity. This helps gain early acceptance of the model.

Summary
As you can see, models can offer many benefits to the services industry. By demanding a clear, concise, and uncompromising investigation and understanding of underlying business basics, they enable better decisions easily communicable to those not familiar with service in a form they can understand. Models can dramatically affect profitability while positioning the service function as a major partner in the corporation’s sales and marketing efforts. By helping resolve differences during the development phase by showing how proposed design or manufacturing changes affect the service function, they can significantly improve cash flows and revenue streams over extended periods.