Beruflich Dokumente
Kultur Dokumente
FINAL PROJECT
Sullivan University, Lexington, KY
QNT 550: Advanced Quantitative Methods
Instructor: Dr. Ahmadin
June 7, 2015
TEAM:
Koti Cherukuri
Srinath Meduri
Harith Rojanala
Bharathi Rajana
Vasudha Nukala
Introduction
Ashford Hospitality Trust (NYSE:AHT) is a subsidiary of Ashford (NYSE MKT:AINC).
Ashford is a leading provider of asset management and other services to companies within the
hospitality industry. Currently, Ashford serves as the advisors to two NYSE listed real estate
investment trusts, namely Ashford Hospitality Trust, our Subject Company and Ashford
Hospitality Prime. Combined, Ashford Hospitality Trust and Ashford Hospitality Prime have 126
hotels with more than 28,000 rooms and approximately $6 billion in assets. In particular, Ashford
Hospitality Trust has 135 properties. Ashford being a billion dollar company, there needs to be a
system in place to recognize revenue and forecast future business in order to make key strategic
decisions. This particular project talks about the statistical application to forecast revenue based
on key parameters such as rooms available, rooms sold, revenue per each room, total revenue
generated, gross operational income, etc.
Background
Application of Statistical analysis for revenue projection: Advantages and similar practices
We decided to use Crystal Ball to predict the right forecasting technique for predicting the
Rooms occupied. As we felt the Rooms occupied was a dependent variable for forecasting the
Total Revenue and Room Revenue. Hence we ran Crystal Ball Predictor for each of the 130 odd
properties of Ashford Hospitality for all the 12 months of the year 2014. Based on the right
forecasting method given by crystal ball for each property, we decided to forecast Rooms
occupied for the next 12 months of the year 2015. Based on the forecast of rooms occupied we
could do a regression forecast to determine the Total revenue, using which we decided to arrive
at the Gross Operating Profit. True to our expectations we could use the above forecasting
techniques to arrive at an optimal forecast for Gross Operating Profit, Total revenue and Rooms
Occupied for the Year 2015.
Identification of data
When we received the data we were given monthly data for the year 2014 for five parameters
Rooms Available, Rooms Occupied, Actual Revenue, Room Revenue, and Gross Operating
Profit for each of the 130 odd Ashford properties. We had to make sense of this data to arrive at a
dependable forecast for the Gross Operating Profit and Rooms occupied for the year 2015.
To forecast the Gross Operating Profit for 2015 we had to run a regression using these five
parameters. We had to select relevant data and analyze it to arrive at relevant results. So in effect
we had to sum up the monthly values of all the five individual parameters to come up with the
total Rooms Available, Total Rooms Occupied, Total Actual Revenue, Total Room Revenue, and
Total Gross Operating Profit for the 130 odd properties for the year 2014. After having arrived at
the summed values for each of the five parameters for the 130 odd properties, we decided to use
this data for multiple regression analysis.
To arrive at the forecast for the monthly rooms occupied for the year 2015, we decide to use the
monthly data to run it through Crystal Ball Predictor to arrive at the forecast for the next 12
months of the year for individual property. We are enclosing sample snapshot of the sorted and
cleaned data in the Appendix.
The relationship is
Gross operating Profit = -313736 + 0.225 * Actual Revenue + 0.336 * Room Revenue
To forecast Gross Operating Profit we had to forecast Actual Revenue. Having found a
relationship between Gross Operating profit and Actual revenue, room revenue, we decided to
find the relationship between Actual revenue and the Rooms Occupied and Rooms Available. We
ran multiple regression between Actual revenue as Independent variable and Rooms Occupied
and Rooms Available as dependent. We found the model to be significant with R square of 0.86.
We found that the actual revenue was dependent on only total Rooms occupied with p less than
0.05.
Lastly we had to forecast the rooms occupied using Crystal Ball predictor. We ran crystal Ball
Predictor for each of the property for each month to arrive at the optimal forecasting method and
Forecast for each property.
Ethical implications
Results obtained vs. Research
Conclusion
REFERENCES