The divide between IT and business departments is well recognized, but how can an organization get these two sides of the operation to work better together in pursuit of successful business intelligence projects? Speakers at the TDWI BI Executive Summit in San Diego said one of the best practices for business intelligence was to start out focusing on narrow goals.
Tony Rathburn, senior consultant with The Modeling Agency LLC, said analytics initiatives have a tendency to lose sight of the business problem they are trying to solve. Instead of delivering information to sales, marketing or finance departments so workers can do their jobs better, BI projects often deliver dense reports that are unnecessarily granular. Such reports rarely get used because the business people don't really understand what to do with them.
We're here to do business. If our technology doesn't help us improve performance, we've just got some very fancy junk with cool features.
The Modeling Agency
"This is not about technology," Rathburn said. "We are not an academic institution. We're here to do business. If our technology doesn't help us improve performance, we've just got some very fancy junk with cool features."
Rathburn recommended a simplified approach, saying business users don't need reports with hundreds of data points. They just need a simple chart showing, for example, where their sales are and where their sales should be. They don't need to understand every feature about their customers. They do need to know whether customers are large, medium or small purchasers. This is typically enough information to help business people make better decisions on who to market to and where to direct their resources.
Expensive and sophisticated BI tools aren't always necessary to deliver this kind of information. Rathburn said Microsoft Excel has tools to classify data points and show users where individuals rank relative to others. Simplifying the tools gets the business users involved and delivers better results. For example, Rathburn said IT workers at an insurance company may be able to develop a system that identifies claims that are likely fraudulent, but it takes the adjusters to define which fraudulent claims are worth devoting resources toward fighting. It may take an investigator several hours and several thousand dollars of times to challenge a fraudulent claim that only cost the company a couple hundred dollars.
"We spend too much time throwing Ferraris at our business users," Rathburn said. "Let's make them more effective users" by simplifying the tools. Simple tools are more likely to be adopted, bringing IT and business departments closer.
James Taylor, CEO of consulting firm Decision Management Solutions, said another best practice for BI implementations is to know the questions you want the tool to answer from the very beginning. A BI project shouldn't just be about exploring data or experimenting with visualizations. These things can be helpful in mature BI implementations, but they are unlikely to improve business processes at the early stages of a project. Each BI initiative should start with a business decision that needs to be improved. For example, banks may want to find ways to improve basic customer interactions.
Read more on BI best practices
Read this mobile BI guide for more tips
Check out these make-or-break BI best practices
Learn what to look for when buying BI technology
"You'll get more value out of the analytic team, and you'll get more value out of the results," Taylor said.
Organizations often put most of their resources toward high-level, strategic decisions, such as which products to introduce or discontinue, where to open up new branches and which consumers to market toward. But they overlook the potential of bringing analytics to bear on day-to-day operational decisions made by average employees. For example, an insurance company may be able to help their underwriters decide who to insure and at what rates by using BI tools to put better information in front of the underwriter.
"If you don't focus on decisions, you won't be able to improve them," Taylor said, adding that best practices for business intelligence that involve focusing on decisions also introduces clarity and helps bring IT and business teams closer together.
Learn what needs to be in place before embarking on an AI project