Adastra pitches for better use of data

The American Sarbanes-Oxley legislation, the new Basel Capital Accord and the subsequent fallout may be a worry for most companies, but it proves the value of BI and data warehousing. At least that’s the take by Gary Saarenvirta, chief operating officer, Adastra Corp. of Markham, Ont.

“If you read between the lines, [the legislation] says that good systems plus good data equals good information equals knowledge and that equals corporate stability and profit,” he argues. “There is a regulatory business case that says data warehousing and business intelligence is not just good, it is mandatory.”

Adastra is an international consultancy based in Toronto with a core focus of helping customers bring data together and use that data effectively. It offers solutions relating to customer intelligence, business intelligence, intelligence infrastructure, data integration and analytics. Although it is vendor agnostic, key partners are IBM, Microsoft, Oracle and Siebel. It also has staff skilled in NCR Teradata solutions.

“We support whatever tools our customers use,” says Saarenvirta. “Sometimes we help customers formulate a process to select a vendor to work with. All major vendors have credible solutions and none are orders of magnitude better than another. We offer the methodology and processes on how to build those systems regardless of the technology.”

What really matters, he says, is the quality of ETL (extract, transform, load) data. “Whether NCR, IBM or Oracle, they all have the same issue. Someone somewhere at the end of the day has to make sure that if you apply the business logic rule that says ‘is this account delinquent,’ what does that mean? Did we apply the definition of delinquent correctly? Did we merge the two right customers? Is J. Smith in the call centre database the same as John Smith in the transaction system?”

Vendors are selling what he calls application islands that are not getting good information and good data. “It is not the applications’ fault specifically but in the process of implementing those solutions, they ignored this critical component: data quality. We help bridge those islands and help make sure the technology that requires sometimes huge investments works.”

Stressing that the opportunity is there beyond the largest companies to get involved in data warehousing, Saarenvirta suggests companies begin with looking at what they are trying to achieve. Are they trying to solve a particular business problem? Then, build a pilot to test the business execution that they can later scale before investing big in the technology.

Hidden secret to success

But don’t overlook data quality or you won’t get the returns on investment, he stresses. “It is one of the hidden secrets of data warehousing: getting the data quality right. It is the most critical deliverable. It is the first chain in the link of good systems that turn the data into information. If you get that step wrong and don’t create accurate information, then the company can’t go further down that chain to knowledge, decision-making, execution and profit. A lot of the application vendors focus on the application side and make the assumption that the data coming in is good. That, in many cases, is not the right assumption.”

For the past year, Adastra has worked with IBM as a business partner to help a major retailer build predictive models with a data mining capability to target their catalogues better. “They were using a targeting method built years ago on how to select which customers get those catalogues and not mail to everybody,” he explains. “We thought we could improve the analysis or modelling they do to choose more accurately the customers to get the catalogue. So, they would increase sales or reduce the number of catalogues they send out and maintain the current revenue at a lower cost. Along the way we found that data quality was an issue so we expect they will be getting their returns and a growth in catalogue profit from all this targeting work.”

He finds that retail customers getting involved in advanced analytics are daunted by the costly exercise of bringing information together in a data warehouse. He notes that it is a difficult task to gather the data that resides “all over the place” in a company that has been in business for some time.

“The challenge is that you’re going to have to change some of the business processes of how you do things,” he warns. “If you’re not leveraging information today when you buy merchandise, for instance, and there’s not a lot of historical and predictive analytics being used to do that, then it is going to be hard to sell that concept of collecting data and figuring how to integrate that into the process. If the business users aren’t perhaps as savvy with their capabilities with technology, then the IT department is left to try to build the first generation of data warehouses.”

Finding the value

He advises IT staff in companies getting started with data warehousing to find someone on the business side who believes they need business intelligence and good information to do their job better. The more senior, the better, he finds. In retail, for example, this may be those in charge of merchandise, store operation, category management or inventory.

Those are very large operational processes that have lots of associated costs. He finds one can help drive the data warehousing and BI efforts by getting the business leader to believe the company can save even one or two per cent of their budget.

“The ROI in retail at the very minimum is one per cent of the merchandising cost, one per cent of store operation, inventory, real estate and distribution costs,” he claims. “If you’re a significant retailer, one per cent of those budgets is a huge number and if that is all you achieve, that alone would be worth it.

“Over the course of 18 months from when you start, you should expect to see significant returns, as much as 500 to 1,000 per cent because it is really going after these major costs or driving top line revenue to sell more,” he continues. “The opportunity is big but the effort is not simple.”

The excitement about BI is when it is used strategically to report core business processes like customer management and CRM or inventory management, says Saarenvirta. “But it requires a fundamental assumption that business will start to do things differently. Will they be more scientific about what they do? Will they combine their experience with good information to be better? Once they want to do things more efficiently, then I think BI is a vehicle that can make that happen.”

Adastra is on the Web at

Related Download
The Landscape of Self Service Analytics Sponsor: IBM
The Landscape of Self Service Analytics
Download this report to examine the current state of self-service analytics across all industries and company sizes, and view the technology decisions and analytical performance of organizations that reported high levels of self-service in their analytical use base.
Register Now