Technical Selection
With a plan to more than double the number of new stores built each year, Maverick Country Stores hired a second real estate representative six months ago, and armed him with a technology tool to help assess new territory.
"We've grown at a rate of seven or eight stores per year, and now we are looking to construct 15 more stores a year," said Dan Murray, director of store development at Maverik Country Stores, based in North Salt Lake, Utah and operating 169 stores in Utah, Idaho and northern Arizona.
"I've been working in real estate for 22 years at Maverick, and we had in-house development software and used outside consultants, but had mixed success with those approaches," explained Murray. "Then we found Prediction Analytics Inc.'s (PAI) Retail Performance Modeling platform almost a year ago."
The platform is customized to meet Maverick's needs and is a Web-based program accessed with a username and password via the Internet. When assessing a particular location, a street address is typed into the system and Maverick's customized 55-question form is filled out.
"The questions deal with what we are proposing to build," noted Murray. "Questions like what area we are building the store in and how the surrounding property is developed."
The system then pulls information about the surrounding competition from its database.
"It pull data within a half mile and mile of the spot we identify on the map and calculates the population and employment information for the area as well as how many competitors," said Murray.
The program provides two sales forecasts for each site -- a general forecast and another report based on the influence of different location variables, with the ability to manipulate the data to run different scenarios, according to Murray.
"For example, we can adjust the influence of residential growth on the forecast of sales," he explained. "It's one thing to tell the model there is residential growth, but does that mean 50 new homes or 100? And also, we can adjust employment information as well as traffic factors."
In addition to the sales forecasts, the system analyzes the risks of the store being underperforming using a percentage range. A low-risk store would be less than a five percent probability, according to PAI.
"We wanted to get a second opinion to avoid opening underperforming stores," said Murray. "With a new, and less experienced real estate rep., this provides him with a tool to understand the relationship between the variables to forecast a site."
Maverik pays an annual fee for the program, which is based on the number of times the company runs the model, and with almost a year of use, Murray is pleased with the results so far.
"The results are quite encouraging," he explained. "We have dropped locations wee were looking at because the model forecasted it to low. I think we are getting better results compared to our old way of doing it."
"We've grown at a rate of seven or eight stores per year, and now we are looking to construct 15 more stores a year," said Dan Murray, director of store development at Maverik Country Stores, based in North Salt Lake, Utah and operating 169 stores in Utah, Idaho and northern Arizona.
"I've been working in real estate for 22 years at Maverick, and we had in-house development software and used outside consultants, but had mixed success with those approaches," explained Murray. "Then we found Prediction Analytics Inc.'s (PAI) Retail Performance Modeling platform almost a year ago."
The platform is customized to meet Maverick's needs and is a Web-based program accessed with a username and password via the Internet. When assessing a particular location, a street address is typed into the system and Maverick's customized 55-question form is filled out.
"The questions deal with what we are proposing to build," noted Murray. "Questions like what area we are building the store in and how the surrounding property is developed."
The system then pulls information about the surrounding competition from its database.
"It pull data within a half mile and mile of the spot we identify on the map and calculates the population and employment information for the area as well as how many competitors," said Murray.
The program provides two sales forecasts for each site -- a general forecast and another report based on the influence of different location variables, with the ability to manipulate the data to run different scenarios, according to Murray.
"For example, we can adjust the influence of residential growth on the forecast of sales," he explained. "It's one thing to tell the model there is residential growth, but does that mean 50 new homes or 100? And also, we can adjust employment information as well as traffic factors."
In addition to the sales forecasts, the system analyzes the risks of the store being underperforming using a percentage range. A low-risk store would be less than a five percent probability, according to PAI.
"We wanted to get a second opinion to avoid opening underperforming stores," said Murray. "With a new, and less experienced real estate rep., this provides him with a tool to understand the relationship between the variables to forecast a site."
Maverik pays an annual fee for the program, which is based on the number of times the company runs the model, and with almost a year of use, Murray is pleased with the results so far.
"The results are quite encouraging," he explained. "We have dropped locations wee were looking at because the model forecasted it to low. I think we are getting better results compared to our old way of doing it."