real SOLUTION U.S. Xpress-Data Quality Drives Millions in Savings three r Bu U.S be Xpress is the thind largest privatcly mpany in the United States. Given the does not get you money, says Rob Karel, principalwhen "The goal is not to get clean data, because beli ecanomic downturn, which has heen very problematic with Forrester Research. "The goal is to fuel youp and decisions in the best way possible r the trucking industry, and the increased facing, the company started to look for ways in which it could control costs and keep its growth plans on track. Although known in the industry as an early adopter of technology U.S. Xpress was one of the first companies to grate satellite communications aboard its flect-management Aspno program using had already this initiative, Leonard launched a data ls from lendor Informatica, , ra had he it r stry as an early adopter of technology had already used them at his previous job with D fully inte- had seen firsthand how powerful those tools were ln ticular, U.S. Xpress implemented two different called Informatica Data Quality, which standardizes o ccurate data about its opera tions would be a major block in the road to better decision mon information elements coming from multiple sure making. For example, cxecutives did not have any accurate and Informatica Data Explorer, which prolesseSI and reliable information on truck idling times, which and maps any data records from any source. Using thel consumed gas but did not have any associated benefits. ter, for example, allowed the company to create a profle f Although seemingly a minor issue, with a large fleet of a legacy database that had been created more than 10 vre trucks those costs added up quite rapidly. But there was not ago, and for the first time really understand how it workel much that could be done about them, since the data needed to take action were just not there Using both tools, Leonard embarked on an aggressive pm gram to normalize and aggregate its disparate data sours into a central master database where they could be bod of The frustration that management voiced was painful for those in IT: After all, one of the main reasons companies cleaned and strictly controlled invest in IT is to have good data that can be the basis for Fast forward four months, and results are starting to making good decisions, isn't it? Although the issue sur show after a lot of hard work. The company used the toolk rounding the need to be able to assess and contain gas costs by Informatica to identify duplicate customer and location and truck idling times became the focal point of contention, information across different trucks and locations it was a symptom of a larger issue related to the heavily balkanized IT environment in which the company was built data-matching rules and integrated geo-coding cape bilities. Searching and matching data moving forwand s operating, largely a result of so many mergers and acquisi-improved by the use of statistical matching algorithms han tions over the years. When Tim Leonard, chief technology ensure accurate results. All common components, such as officer, came to the company from Dell, he found data abbreviations, formats, and so forth are standardized acros data sources and validated along the way. The new, dean separate database, so the company can stored all over the place. Some would be on an AS400 (an IBM hardware product, part of its iSeries server line), and data are stored in a some in other environments. With more than 130 different always go back to the original data in case it needs to e applications and no integrated, master database, putting tract new elements or add something that was not needed together any data for any business purpose was quite com- before. But the two don't mix data would be correct. To add insult to injury, much of ing its mark. Almost the data stored in those files was "dirty, that is, incorrect, create accurate re inconsistent, or incomplete. As a result, IT embarked plicated, and without a high degree of assurance that the The availability of high-quality data soon started malk immediately, U.S. Xpress was able t reports on truck idling times, w on lowed it to drastically reduce vehicle idle time by issuing cleaning up old data and making sure new directives regarding best driving practices. This onec new data would be already clean before entering its systems. Dale Langley, chief information officer for U.S. Xpress, across its flect launched a comprehensive revamping of its IT infrastruc- ture and strategy, with the idling problem becoming the pilot to show the value of this important effort. 218 saves the company $6 million a nnmuly of trucks. "Almost immediately, we cut te aspect alone now percentage of time our trucks stand idle from the high to just over 50 percent. That is saving the company every year-equivalent to a return on investment in e U.S. Xpress board is ecstatic Chapter s other tasks. The greatest 219 To say the imate how well the Data Resource Management bemong other things, truck idle t scorecard a Web. is no lothe bines wilt benefi, however, is in chaneine tation that once prevented the com scen data being used the way it should be. IT has done gect ng on this success, the ng other things, truck idlead tha is no lol operate in the future. Lack of data created a We This ltool that includes a s eb- pany from designing, or even thinking so0 staff who are responsible for man- who Our Chief Operating Office (COO) recently said to ed by about 800 d coc an alert process that notifics managers of successes, Leonard is now pashing for ckis hissues from ever happening in the first nology. Leonard anticipates that the intnis to b te efforts where the most savings could greatest fian of ITI" says Lecomard. Armedwih eformore than 24hoursstraight. which generation data warehouse also based on informat hing old him he had a parked truck that had important, IT has gone from problcn nand integrated coming from someone who's not the nsible managers are notified in real time enterprise data warehouse may increase dss he respout to go wrong. "We contacted one $9 million and $20 million over the nex days," says I eonard. "He said, 'If I r three days, I'd know about it. Then ad there it was. omethody parked it and left system identified an instance in the past business partner in less thanyear. data are giving managers more and m they can get out of the same truck. Thus, gaining a better ing of how a truck is really used gives managers hat, nks to improved reporting and visibility. managers will point to a piece of data and say that it as wrong not the data itself, but whatever is going on that nhad been running idle for seven days. This is new thoughts about how to manage that asset. Sometimes, happen a er database also allowed operations per- truck maintenance routines. Now, with and clean data, they could better predict results in those reports. And now they have the tools to g out and fix it. SOURCE: Rob Lemos,"Big Data: How a Trucking Firms Drowe Out te larer standards and historical occurrences, instead Big Eron,CiO Megrczine, Fetruary 28, 2011 that the truck might need to Drives Savings of So Million Anaually with Informatica Daua Quality," e loked at, or waiting for it to break down. Leonard esti- Informatica Preu Release, July 12, 2010 An Xcelllent Way of Saving as hat this resulted in additional savings of S1.2 million illia Year: lofrmatice Cesr Stady Tim Lcomard,"Informatics Data Quality, nformarie Management Magesoe, May 1, 2011; Wim r ear, on top of those realized from the reduction of idling Casidly Thecking Gasto the Cleaners,"Jamal ef Cw, May 2 ines This also allowed numerous stafters to focus on 2010, and www.uapress.com, acessed May 22, 2011 QUESTIONS TO CONSIDER I.Oace the technical aspects of data quality are put in 2. Are the benefits outlined in the case the result of better pace, who should be in charge of making decisions technology or improved decision making? Today, is it abkout these issues? Is this a technical responsibility or a possible to clearly separate the two anymore? What are the implications for U.S. Xpress as it decides where to esiness one? What are the advantages and disadvan go next and how to invest in future projects? real SOLUTION U.S. Xpress-Data Quality Drives Millions in Savings three r Bu U.S be Xpress is the thind largest privatcly mpany in the United States. Given the does not get you money, says Rob Karel, principalwhen "The goal is not to get clean data, because beli ecanomic downturn, which has heen very problematic with Forrester Research. "The goal is to fuel youp and decisions in the best way possible r the trucking industry, and the increased facing, the company started to look for ways in which it could control costs and keep its growth plans on track. Although known in the industry as an early adopter of technology U.S. Xpress was one of the first companies to grate satellite communications aboard its flect-management Aspno program using had already this initiative, Leonard launched a data ls from lendor Informatica, , ra had he it r stry as an early adopter of technology had already used them at his previous job with D fully inte- had seen firsthand how powerful those tools were ln ticular, U.S. Xpress implemented two different called Informatica Data Quality, which standardizes o ccurate data about its opera tions would be a major block in the road to better decision mon information elements coming from multiple sure making. For example, cxecutives did not have any accurate and Informatica Data Explorer, which prolesseSI and reliable information on truck idling times, which and maps any data records from any source. Using thel consumed gas but did not have any associated benefits. ter, for example, allowed the company to create a profle f Although seemingly a minor issue, with a large fleet of a legacy database that had been created more than 10 vre trucks those costs added up quite rapidly. But there was not ago, and for the first time really understand how it workel much that could be done about them, since the data needed to take action were just not there Using both tools, Leonard embarked on an aggressive pm gram to normalize and aggregate its disparate data sours into a central master database where they could be bod of The frustration that management voiced was painful for those in IT: After all, one of the main reasons companies cleaned and strictly controlled invest in IT is to have good data that can be the basis for Fast forward four months, and results are starting to making good decisions, isn't it? Although the issue sur show after a lot of hard work. The company used the toolk rounding the need to be able to assess and contain gas costs by Informatica to identify duplicate customer and location and truck idling times became the focal point of contention, information across different trucks and locations it was a symptom of a larger issue related to the heavily balkanized IT environment in which the company was built data-matching rules and integrated geo-coding cape bilities. Searching and matching data moving forwand s operating, largely a result of so many mergers and acquisi-improved by the use of statistical matching algorithms han tions over the years. When Tim Leonard, chief technology ensure accurate results. All common components, such as officer, came to the company from Dell, he found data abbreviations, formats, and so forth are standardized acros data sources and validated along the way. The new, dean separate database, so the company can stored all over the place. Some would be on an AS400 (an IBM hardware product, part of its iSeries server line), and data are stored in a some in other environments. With more than 130 different always go back to the original data in case it needs to e applications and no integrated, master database, putting tract new elements or add something that was not needed together any data for any business purpose was quite com- before. But the two don't mix data would be correct. To add insult to injury, much of ing its mark. Almost the data stored in those files was "dirty, that is, incorrect, create accurate re inconsistent, or incomplete. As a result, IT embarked plicated, and without a high degree of assurance that the The availability of high-quality data soon started malk immediately, U.S. Xpress was able t reports on truck idling times, w on lowed it to drastically reduce vehicle idle time by issuing cleaning up old data and making sure new directives regarding best driving practices. This onec new data would be already clean before entering its systems. Dale Langley, chief information officer for U.S. Xpress, across its flect launched a comprehensive revamping of its IT infrastruc- ture and strategy, with the idling problem becoming the pilot to show the value of this important effort. 218 saves the company $6 million a nnmuly of trucks. "Almost immediately, we cut te aspect alone now percentage of time our trucks stand idle from the high to just over 50 percent. That is saving the company every year-equivalent to a return on investment in e U.S. Xpress board is ecstatic Chapter s other tasks. The greatest 219 To say the imate how well the Data Resource Management bemong other things, truck idle t scorecard a Web. is no lothe bines wilt benefi, however, is in chaneine tation that once prevented the com scen data being used the way it should be. IT has done gect ng on this success, the ng other things, truck idlead tha is no lol operate in the future. Lack of data created a We This ltool that includes a s eb- pany from designing, or even thinking so0 staff who are responsible for man- who Our Chief Operating Office (COO) recently said to ed by about 800 d coc an alert process that notifics managers of successes, Leonard is now pashing for ckis hissues from ever happening in the first nology. Leonard anticipates that the intnis to b te efforts where the most savings could greatest fian of ITI" says Lecomard. Armedwih eformore than 24hoursstraight. which generation data warehouse also based on informat hing old him he had a parked truck that had important, IT has gone from problcn nand integrated coming from someone who's not the nsible managers are notified in real time enterprise data warehouse may increase dss he respout to go wrong. "We contacted one $9 million and $20 million over the nex days," says I eonard. "He said, 'If I r three days, I'd know about it. Then ad there it was. omethody parked it and left system identified an instance in the past business partner in less thanyear. data are giving managers more and m they can get out of the same truck. Thus, gaining a better ing of how a truck is really used gives managers hat, nks to improved reporting and visibility. managers will point to a piece of data and say that it as wrong not the data itself, but whatever is going on that nhad been running idle for seven days. This is new thoughts about how to manage that asset. Sometimes, happen a er database also allowed operations per- truck maintenance routines. Now, with and clean data, they could better predict results in those reports. And now they have the tools to g out and fix it. SOURCE: Rob Lemos,"Big Data: How a Trucking Firms Drowe Out te larer standards and historical occurrences, instead Big Eron,CiO Megrczine, Fetruary 28, 2011 that the truck might need to Drives Savings of So Million Anaually with Informatica Daua Quality," e loked at, or waiting for it to break down. Leonard esti- Informatica Preu Release, July 12, 2010 An Xcelllent Way of Saving as hat this resulted in additional savings of S1.2 million illia Year: lofrmatice Cesr Stady Tim Lcomard,"Informatics Data Quality, nformarie Management Magesoe, May 1, 2011; Wim r ear, on top of those realized from the reduction of idling Casidly Thecking Gasto the Cleaners,"Jamal ef Cw, May 2 ines This also allowed numerous stafters to focus on 2010, and www.uapress.com, acessed May 22, 2011 QUESTIONS TO CONSIDER I.Oace the technical aspects of data quality are put in 2. Are the benefits outlined in the case the result of better pace, who should be in charge of making decisions technology or improved decision making? Today, is it abkout these issues? Is this a technical responsibility or a possible to clearly separate the two anymore? What are the implications for U.S. Xpress as it decides where to esiness one? What are the advantages and disadvan go next and how to invest in future projects