BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
Good design is crucial on mobile business intelligence applications -- but a user-friendly interface only goes so far. For BI and IT managers, it's also vital to ensure that the back-end side of a mobile BI system can handle the processing load and deliver data to users in a timely manner.
Perhaps the most important piece of the puzzle for business users accessing applications in the field is the mobile network they're using. While bring-your-own-device policies make it difficult to restrict users to a specific network, companies serious about developing mobile BI apps might want to consider buying smartphones or tablet PCs for their employees and using a single carrier.
"Network performance becomes critical as you develop applications," said Michele Pelino, an analyst at Forrester Research Inc. in Cambridge, Mass. In evaluating networks for mobile BI applications, Pelino advised looking at availability, uptime and jitter -- the data-transfer delay caused by network congestion or routing changes as information moves across a network. "You want to know that you have a seamless network to put your applications on," she said.
In addition, she suggested putting applications in different Quality of Service tiers on a network, depending on how important it is to avoid any delays in accessing or updating data.
Another key element in supporting mobile BI applications is the server hardware that powers them on the back end. The good news is that the processing load doesn't necessarily require heavy-duty hardware.
Old, but no longer in the way
For example, food products maker Illes Seasonings & Flavors was able to put some decommissioned servers back into use to run a mobile BI proof-of-concept deployment and then an initial mobile application for its sales force, said Les Howell, the Carrollton, Texas-based company's executive director of IT. Howell set up the application to refresh data only every 30 minutes because of concerns that refreshing it more often would tax the ERP system behind the app. As it turns out, though, he said it takes just five minutes to rerun all of the company's sales data going back six years.
Nonetheless, Howell said he eventually hopes to secure budget money for a purchase of more powerful servers recommended by QlikTech Technologies Inc., his BI software vendor, in order to support a broader rollout of the mobile software.
The required server infrastructure depends on "the size of the organization and where you're going," said Darren Taylor, president of Blue Cross Blue Shield of Kansas City's Cobalt Talon analytics services subsidiary in Kansas City, Mo. Blue KC was able to use existing hardware with an underlying Teradata database and MicroStrategy Inc.'s BI software on top of it to run a mobile application that was developed when Taylor was working as the medical insurer's vice president of enterprise analytics and data management.
The amount of data involved and the complexity of the processing workloads, as well as failover and backup requirements, also affect the design of a systems architecture that can meet an organization's mobile BI needs, according to William McKnight, president of McKnight Consulting Group in Plano, Texas.
Keep it simple -- if your apps aren't complex
Some organizations feed data from source systems to an operational data store, then to a staging system, a data warehouse and an online analytical processing (OLAP) cube before the information finally goes to the end user. "I'm not saying that's uncalled for, but you want to keep it as simple as possible," McKnight said. A streamlined setup will speed up the delivery of data to users, he added.
More on mobile business intelligence applications
Read about a mobile BI project at office furniture maker OFS Brands
Watch a video Q&A with consultant Claudia Imhoff on combining collaboration and mobile BI capabilities
Feeding data to mobile BI users directly from transaction databases isn't always feasible, though. That can work with small amounts of information that doesn't need to be pulled together from multiple source systems, McKnight said. But for larger applications or workloads that require high performance, using a data warehouse or a specialized analytical platform, such as a columnar database, typically is a must.
Keeping mobile BI application performance humming at a fast pace also depends on the work of database administrators (DBAs) and other IT and data management staffers. For example, McKnight said that DBAs can use techniques such as database denormalization, indexing and summarization to help boost back-end processing.
And, Taylor said, organizations should invest in data quality, aggregation and governance initiatives. That isn't the glamorous side of running mobile business intelligence applications, but he said the end result will enhance performance and help ensure that a mobile app deployment meets the needs of business users.
Christine Parizo is a freelance writer who specializes in covering business and technology issues.