**Editor's Note: Got SOA questions? Ask IBM architects and industry experts THIS WEEK at Connectivity in Action! Continue here.
Today's executives are on the move, and they need to be able to run their companies
remotely. Mobile business intelligence (BI) applications allow them to access
critical corporate data from their laptops, smart phones and PDAs. The iPhone
promises to deliver information on an even more compact and versatile device.
But can BI vendors make it work with the information executives depend on?
The answer is yes, but it's important to understand both the limitations and
the architectural requirements of this amazing new platform. Let's take a step
back and consider the iPhone's legacy, beginning with the iPod.
The iPod was successful because it creatively destroyed product boundaries,
hastening the convergence of disk storage, high-resolution video screens, and
music technologies. Apple changed our perceptions about mobile media, creating
a newer, faster, sleeker type of product.
The iPhone is on a similar path of creative destruction -- this time, with
the potential to replace phones, PDAs, and even the iPod itself. It's already
a success in the consumer world. The major challenge facing Apple is to prove
that the iPhone is viable for the business world.
That's where BI comes in.
As business people begin to use mobile phones and laptops almost interchangeably,
and as those phones become capable of displaying rich web content, mobile workers
are seeking more extensive access to corporate data. Apple embedded an advanced
Safari browser in the iPhone to allow people to take advantage of the same web-based
applications that they use on their desktops, eliminating the need for dual
devices for many employees.
If you can write e-mail, watch movies, and browse the web on your phone, why
not check your bills and bank statements as well? How about filing expense reports,
your sales numbers, and verifying the status of an order?
These tantalizing scenarios open doors to an exciting range of possibilities,
assuming your BI software can work properly in the elegant yet constrained iPhone
There are currently two approaches to mobile BI deployments: thick-client and
thin-client. Thick-client deployments run special software on each type of mobile
device, fed by special servers that manage the interactions with those devices.
The client-side software controls how content is displayed, which was an important
factor in the early days of mobile browsers, when each device displayed content
Most BI vendors offer thick-client solutions, with different client software
for the different mobile devices. Their approaches work well for organizations
that have standardized on a relatively small number of mobile devices. However,
thick-client solutions won't work for the iPhone because it is a "locked"
device. Apple does not allow developers to install applications on this platform
-- both for security reasons and because the iPhone uses the Web as a delivery
The iPhone only works with thin-client solutions that use standard Web technologies
to deliver and display information. Thus to gain true analytic capabilities
on the iPhone, you need self-contained BI applications with an active payload
of data that can be delivered as part of a standard HTML page. These "active
reports" can allow users to sort, filter, and query the associated data
through parameterized reports. This gives them complete analytic capabilities,
even when disconnected from the network.
Which brings us to content. Since the memory and processing power of these
mobile devices cannot match that of a laptop, it is critical to deliver only
the most relevant information. Users should be able to select what they need,
then drill down to obtain precise results. When in disconnected mode, they can
drill down into the data contained in the active reports. When connected to
the network, they can select additional information from corporate databases.
But keep in mind, if the information is too difficult to read or the BI application
is too difficult to use, nobody will bother with it. As my former CEO used to
say, unless a report can fit in the palm of your hand, it is not worth reading.
"Use a Post-it note as a template," she would say. In screen geometry,
that translates to about seven rows of data by four or five columns of measures.
BI scorecards fit the bill perfectly.
The iPhone screen is larger than a Post-it note, which makes it an ideal candidate
for a wide variety of scorecard reports, from operational summaries and sales
results to inventory reports and account statements. Its advanced zoom function
makes it perfect for dashboards, too -- as long as the individual dashboard
components, when zoomed, fit into the Post-it template.
By leveraging the type of active reporting technology mentioned above, an "active
dashboard" can pack a lot of information into a very small space -- once
again, with an associated payload of data that enables users to display charts
and tables using the zoom function of the browser. The Safari browser is perfect
for this architecture, since it includes drop-down boxes that make it easy to
interact with multiple pages of information.
Of course, security, usability and whatever legacy equipment you have on hand
will all play significant roles in determining which mobile BI architecture
to choose. But given the resiliency of the Web and the momentum of the iPhone,
I'm betting on thin-client solutions that can work interchangeably in any web
browser, on any device.
In the long run, as enterprise and mobile applications converge, users will
demand consistency, from desktops to cell phones. That's why so many application
vendors are betting on thin-client approaches to carry on the tradition of creative
destruction. Astute BI vendors are doing the same.
About the Author
Dr. Rado Kotorov is a technical director of strategic product management at Information Builders Inc., responsible for emerging reporting, analytic and visualization technologies. Prior to joining Information Builders, he managed the implementation of BI solutions and decision-support systems, data warehouses, and custom applications. He has developed analytic models and applications for the pharmaceutical, retail, CPG, financial, and automotive industries. Rado Kotorov has a PhD in decision and game theory and economics from Bowling Green State University. He has publications on business processes, emerging technologies, CRM, KM, innovation, and entrepreneurship.