Lean Six Sigma Success Story:
Project Summary
- The Challenge: Reducing Cost
- The Discovery: Data Visualization in the Hospitality Industry
- The Improvements: A simple presentation to the laundry service vendor resulted in a refund to the hotel of $250,000.
- The Results:
-
current period (3.30)
-
last period (3.20)
-
same period, previous year (3.87)
-
THE CHALLENGE
Businesses collect lots of data but don’t always translate that data into useful information about critical business processes. Data that companies collect can readily be incorporated into charts and graphs that provide this information. Here are a couple of examples.
THE DISCOVERY
Using a Histogram
A hospitality company experienced an unwelcome earnings surprise at the end of one of its quarters. In fact, it was another disappointing quarterly earnings surprise. The company had incurred several consecutive quarterly earnings disappointments. The CEO understandably demanded that his lieutenants provide an explanation of where earnings were falling short of forecast and why. A group that we were working with was the Business Development group. One of the key functions was to contract new franchise opportunities and plan and open the new hotel. The group was keenly interested in understanding why revenues from its function was falling short as the group had been signing on new franchises at a record pace in the previous 24 months.
To first understand how the group was performing with new hotel openings we gathered the data to baseline how well the group was performing at opening new hotels. The group had exceeded its objective in number of deals closed but the revenue stream of management fees didn’t commence until the properties were actually operational.
The group was keenly interested in understanding why revenues from its function was falling short as the group had been signing on new franchises at a record pace in the previous 24 months.
The group had a New Hotel Opening Process. Upon signing of a contract with a new or existing franchisee, the new property construction and opening was scheduled in project management fashion to a standard project template. Hotels were expected to open 14 months after closing of the contract. As is shown from the data of actual openings, the majority of openings were late. A significant portion of them were very late defined as more than 24 months late. Anticipated revenue from new hotel openings were forecasted from the 14 month “contract to management fee” model. Actual revenue generation from the new properties significantly lagged the forecast model creating one of the discrepancies in the quarterly revenue shortfall.
The company updated the revenue forecast for new properties to reflect actual on-boarding of new properties and launched an aggressive improvement project to reduce cycle time for managing new hotel openings. The team implemented standard work for new hotel openings, visual control to quickly identify problem properties and changed its criteria for assessing the capabilities of its potential new franchisees.
THE IMPROVEMENTS
Using a Times Series or Run Chart
In another example, a manager at a hotel that outsourced its laundry service to a third party vendor noticed a decline in the cost of laundry. Weekly costs seemed lower than what they normally were (3.25 compared to 4.5 per unit). While the cost decrease was a good thing, she was interested in understanding why the cost had gone down. The manager decided to look at the weekly cost per unit from the previous two years.
Viewing the data in a time ordered fashion clearly reflected a change in cost per unit. This visualization of the data over time conveys much more information than the standard management report which typically is a 3-point data management report comparing
current period (3.30)
last period (3.20)
same period, previous year (3.87)
Interpreting the data in the standard management report format, we would note that the cost this week (3.30) is greater than the cost last week (3.20) but compared to the same week, 1 year ago (3.87), the current cost is better. So, we’re not as good as last week but we’re better than last year. Now what?
THE RESULTS
Visualizing Data Helps Uncover the Truth
In actuality, those 3 data points are random points from a continuum and don’t reveal anything about what is really happening in the process. The time series, showing data in time ordered sequence from week to week, allowed the manager to visualize changes in average performance and variability between weeks. The manager was curious as to why the cost per unit suddenly dropped from 4.5 to 3.25. The variability of the unit cost seemed to be the about the same but the average cost per unit had shifted lower.
Upon investigation with her staff, the manager found that her staff in the laundry department were following procedures and had not made changes to the operation. However, one of her staff noted that the third party laundry service recently changed drivers. The only noted change in the process was a new driver. But again the question remained, why the change in cost?
The manager and staff watched the new driver as he collected and weighed the laundry bins. The driver removed the laundry from the bins, weighed the laundry and then returned the laundry to the bins prior to securing the bins on the truck. The staff quickly came up with a hypothesis, “what if the previous driver never removed the laundry from the bin?” Adding the weight of the bin to the weights of the laundry deliveries of the new driver and adjusting for the additional cost, the unit cost would be the same as with the old driver with the average cost returning to the 4.5 unit cost. In other words, the old driver had been adding the weight of the bins to the laundry weight.
A Key to Understanding: Ask “Why?”
Thus, the laundry vendor had been overcharging the hotel for laundry service. A simple presentation to the laundry service vendor resulted in a refund to the hotel of $250,000. It was a good thing the manager wasn’t just satisfied with the observed lower cost. She wanted to understand why.
