WALTHAM, Mass. -- A thorough understanding of business user requirements and the right consultant can help organizations...
get through a data warehousing project in tough economic times, according to attendees and speakers at the TDWI Boston chapter meeting here yesterday.
The sluggish economy means that even successful firms, like Manhattan Beach, Calif.-based Skechers USA Inc., have to deal with tight budgets, said Barry Galbreath, the popular footwear retailer's manager of business analytics. Galbreath and meeting attendees offered up several tips on how to make sure data warehousing funds are spent efficiently.
After years of steady expansion, Skechers recently undertook a major data warehousing project designed to help the company do a better job of sales forecasting and running business analytics. But it wasn't a super-expensive initiative. Galbreath, who called in to the meeting from Skechers' headquarters to share his story, said getting the project done cost-effectively was a very high priority.
Galbreath told the audience that it's important to spend as many data warehousing dollars as possible without going over budget.
"Buy as much as you can. Get into your budget," he said. "Plan for growth from the beginning and make sure that your systems are scalable."
Skechers ultimately decided to implement a data warehousing appliance from Dayton, Ohio-based Teradata Corp., along with extract, transform and load (ETL) software from Burlington, Mass.-based Expressor Software Corp. Galbreath said the decision was based largely on price/performance considerations. He added that the project was completed over the last year and a half.
The price tag for the entire project -- including the Teradata appliance and the Expressor software -- came in at "far, far less" than $1 million, said Hugo Sheng, Expressor's director of field engineering.
"[People think] it's very curious that Skechers would spend so much on Teradata yet so little money on Expressor because it just seems to me like a huge dichotomy," Sheng said. "But they didn't. They had a shoestring budget and they had to stay within the budget."
Good data warehousing consultants go a long way
The Internet is brimming with corporate and independent consultants that can help organizations make the right data warehousing choices, but be sure to examine credentials closely before bringing a consultant on board, said John Ormond, the director of database technology at UMass Medical School in Worcester, Mass.
UMass Medical School recently completed a major initiative that required Ormond and his team to combine medical industry information, pharmacy data and electronic medical records (EMR) into one "massive" data warehouse while making it easy for end users to build their own business intelligence (BI) reports.
"Try to find independent consultants that are very skillful," Ormond said. "Those consultants that have been doing it a long time will know the right tools to select at the right price point."
Smaller organizations should also be aware that many consulting firms have partnerships with large software vendors and may be incented to send those vendors business. That can be difficult to deal with, especially if the vendor's prices are too high.
"A lot of smaller organizations are priced out of a data warehouse solution because of the cost," Ormond said.
Galbreath said consultants were essential to Skechers' data warehousing initiative.
"We ended up engaging a consulting organization to help us make sure that we had all of our requirements in place and documented properly and take us through the vendor product assessment phase," he said. "That was really money well spent because it enabled us to keep doing our day jobs."
Understand user requirements for the data warehousing project
Organizations should avoid the tendency to go with a data warehousing software or ETL vendor simply because they have a well-known name, said Abhijit Deshmukh, the president of Chelmsford, Mass.-based Abridge Info Systems, a data management services and software firm.
There are plenty of smaller or open source software vendors that can offer data warehousing capabilities at an attractive price point, but many of them are highly specialized. That's why experts say it's important to first gain a solid understanding of end user requirements by talking to business users and pinpointing what it would take to make their professional lives easier.
Questions to consider when attempting to identify requirements include: What are business users trying to accomplish? What information do they need? Where is that information coming from? What types of reports do business users need to run?
A keen understanding of business requirements will also help organizations avoid unnecessary bells and whistles when buying data warehousing software and the BI software that runs on top of it, according to Deshmukh.
"Definitely look around," he said. "Everybody has got their budgets, but within those budgets you need to make sure that you are getting the best service and that product definitely satisfies your needs."