Hospitality Industry Case Study
Over the past several months, the basic building blocks for development of key performance indicator (KPI) targets and threshold levels have been introduced and discussed. Both KPI x-bar charts and KPI range charts were profiled in detail (see the May and June issues of DM Review, respectively). This knowledge will now be leveraged to illustrate a real world case study in the hospitality industry - the challenge to effectively track and manage reservation wait times, an important success ingredient for retaining existing customers and acquiring first-time customers. Keep in mind that the Six Sigma control chart techniques presented here could also be extended to similar venues such as call centers, help desks and service repair. In fact, any KPI metric that emanates from a process with normally distributed observations is a candidate for this control chart approach.
Hospitality Industry Overview
During the last decade, the reduction in business travel and the cyclical nature of the hospitality industry has challenged management to develop creative branding approaches (nap hotels, executive apartments, HHonors Rewards/frequent guest programs), superior accommodations (Heavenly Beds/Grand Beds, digital HDTV, high speed Internet) and optimized operations (Energy Star compliancy, weekly commission payments). Throughout the hospitality industry, the mantra of "excellence" is consistently mentioned as the key ingredient to success. All the major players have imbued their vision statements and core values with references to pride, spirit, integrity, quality and consistency - all imperative inputs to the excellence equation. The value statement for Starwood Hotels and Resorts Worldwide, Inc. specifically mentions a "passion for excellence" and "encouraged innovation." In the search for performance management excellence, the Hilton Hotels Corporation has implemented a balanced scorecard that incorporates revenue maximization, operational effectiveness and brand management. The culture at Marriott International, Inc. prides itself on its reputation for superior customer service - "people serving people."
Over the years, standard metrics such as RevPar (revenue per available room), ADR (average daily rate) and occupancy (rooms sold divided by rooms available) have been developed to capture financial and operational excellence. While these metrics are useful in capturing the current and past "state of the business," they fail to forecast business trends for the coming weeks or months. Within the balanced scorecard framework, these types of KPIs are referred to as lagging indicators. In the ideal world, the balanced scorecard would also contain several leading indicators that predict customer growth and future revenues. While these indicators represent the holy grail of KPI metrics, their definitions and constructs often remain elusive. In the following case study, we will explore one such KPI that meets the criteria for a leading indicator.
In an industry where customer interactions occur on an hourly basis, each customer touchpoint is critical for building personalized service creditability and developing customer affinity. One of the most important touchpoints occurs early in the customer experience cycle during the hotel reservation process. An initial negative or positive customer reservation experience can significantly influence future impressions and expectations. A miscue at this stage, such as excessive wait time, can lead to a lost reservation and potentially a lost customer. In the next section, we will explore how control chart techniques facilitate the development of effective (and statistically valid) wait time KPI targets and KPI thresholds that can assist hospitality management in controlling and reducing reservation wait times.
Control Chart Implementation
In this series of columns - "KPIs: Avoiding the Threshold McGuffins" - many new terms and concepts have been introduced relevant to control chart methodology. Six Sigma terms such as x double bar, control limits and adjustment factors have peppered our recent discussions. For those of you experiencing a Six Sigma overload, I have created a glossary (see Figure 1) that consolidates all the terminology required for the control chart implementation process that follows. More extensive discussions on development of the x-bar and range charts can also be found in the earlier columns of this series.
Figure 1: Control Chart and KPI Glossary
Before embarking on the specifics of the hospitality case study, a quick review of the relationship between KPIs and control charts is appropriate. The objective in this case study is to develop effective and statistically valid KPI targets and KPI thresholds for reservation wait times. This is accomplished by sampling daily wait times and using these observations to calculate x-bar and range control charts. The x-bar becomes the KPI target, and the XUCL (upper control limit) and XLCL (lower control limit) become the upper and lower KPI threshold levels, respectively. The KPI targets and thresholds are thus driven by historical data rather than gut feel or Kentucky windage.
The comprehensive control chart development process consists of eight important steps. All the relevant data collection forms, formulas and charting tools are illustrated below.
Step 1: Select process to chart - define the variable to be measured.
- wait time required by customers to make contact with a reservation agent
Step 2: Determine process sampling plan - describe number of samples and frequency of sampling.
- eight subgroup samples collected starting at 6 a.m. and ending at 10 p.m. EST
- each subgroup sample will consist of four observations collected over two-hour time periods
Step 3: Calculate data from process - capture and organize observations from sampling plan.
- see sampling plan data in Figure 2 for results from the eight subgroup samples
Figure 2: Sampling Plan Data
Step 4: Calculate control chart specific statistics - develop fundamental statistics.
- see Figure 2 for calculations of x double bar and R bar
- x double bar = 2.97 seconds (average of x-bars of subgroups)
- R-bar = 1.50 seconds (average of ranges of subgroups)
Figure 3: Control and Warning Limit Calculations
Step 5: Calculate control and warning limits - calculate formulas and factors to calculate control limits.
- see formulas and calculations in Figure 3 (Control and Warning Limit Calculations)
- because n = 4 observations per sample group, use following adjustment factors: A2 = 0.73, D3 = 0 and D4 = 2.28 (see Figure 4: Control Limit Factors)
Figure 4: Control Limit Factors
Step 6: Construct control chart - construct both the x-bar chart and range chart.
- incorporate averages, control limits and warning limits calculated in Steps 4 and 5
- wait time values are graphed on the y-axis and the subgroup sample number are reflected on the x-axis (e.g., for subgroup 8, the red circle indicates the value of 4.2 seconds on Figure 5, while the range value of 1.3 seconds is shown on Figure 6).
Figure 5: x-bar Control Chart
Figure 6: Range Control Chart
Step 7: Interpret control chart results - examine the x-bar and range control chart for anomalies and trends.
- examine range chart first to ensure stability of KPI metric process: because no ranges are outside the RLCL or RUCL, the wait time process is in control (see Figure 6)
- examine x-bar chart (see Figure 5) to identify statistical patterns and data outliers (see the June 2005 issue of DM Review for a portfolio of pattern detection rules)
- Rule violation 1 - four consecutive data points in a row trending up or down merits a caution. This occurs in subgroups 1 to 4 where the x-bar starts at 2.4 seconds and edges up to 2.5, 2.7 and finally 2.9 seconds. This violation will alert management to potential problem areas.
- Rule violation 2 - one individual data point outside a control limit is a serious problem. This occurs for subgroup 8 where the value of 4.2 seconds exceeds the Upper Control Limit of 2.97 seconds. This violation must be addressed immediately to identify the root cause and correct the situation.
Step 8: Incorporate KPI thresholds into triggers and visuals - design visual elements and triggers.
- pattern detection rules need to be integrated as business rules and triggers into the delivery and presentation system
- pattern detection rules can then be linked to visuals such as stoplights, beacons and flags
- Rule violation 1 - four consecutive data points in a row trending up would trigger a yellow beacon indicator
- Rule violation 2 - one individual data point outside the control limit would trigger a red stoplight or flag
Keep in mind that development of the reservation wait time control charts is only the first phase in a continuing process to meet the customer service excellence mantra. Potential opportunities exist to further decrease reservation wait time by understanding the business and process drivers. Additional Six Sigma techniques such as y-to-x treeing, input/output analysis, cause/effect matrix, fishbone diagrams and SIPOC analysis would be effective as follow-on activities.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access