Getting your hands on data with mobile analytics

Year after year, organizations cite business intelligence and analytics as top investment priorities, but they’re seldom satisfied with the results. In part that’s because traditional solutions require a handoff to experts who know how to query warehouses full of data — and dutifully return static reports that reveal somewhat less than anticipated.

That’s why we’ve seen a new wave of tools to help business users visualize data and interact with it in real time. In this week’s New Tech Forum, Donald Farmer, vice president of innovation and design at the data discovery firm Qlik, explains how mobile analytics opens a whole new way of interacting with critical business data. — Paul Venezia

Mobile analytics is not reporting

When you consider mobile business intelligence, whether for sales teams on the road, executives visiting remote offices, or retail managers walking around their own store, bear in mind one caveat: Mobile reporting alone is not good enough. You need analytic tools, too.

Why? And why particularly for mobile BI?

On the move we encounter new problems and scenarios that don’t easily fit the static view of the world that a report addresses. After all, you’re probably mobile in the first place to learn what you can’t see from the static view back at the office.

Mobile analytics requires a lot more than traditional BI squeezed into a skinny form factor. The whole experience of mobile analysis can be quite different, in fact, and may even change the way you work with data comfortably behind the desk.

Let’s look at three basic aspects of mobile analytics and their rather surprising implications:

1. To touch is to explore

For any application, the touch interface stands out as a striking difference between mobile and desktop environments. In the past, vendors often attempted to crowd more and more information into an app with additional details on mouse-over hints and right-click context menus. These options simply don’t exist in the same way on a mobile device. Add the need to make interactive areas large enough to touch (Apple recommends 44 by 44 pixels at least), and mobile design demands a cleaner, less crowded experience.

You will also find subtle but important differences between touching your data right there on screen and pointing and clicking with a mouse. I often say seeing is believing, but touching is trusting. Our user research shows that customers — in the real world, not in the lab — spend considerably longer browsing data with a touch device than in the same application on a laptop.

In one case, a firm selling retail mortgages found that locations using tablets increased their conversion rate (of consumers entering their store, then signing for a mortgage in that same visit) by 30 percent compared to stores running the same loan-modelling tool on a laptop. Customers spent longer with the touch app and felt more comfortable with the offers they found and explored.

A touch interface challenges more than the user experience. It encourages a new attitude to the data: exploration, not just passive reading. You need a platform that offers more than merely static reporting.

2. Responsive design

Mobile analytics also demands responsive interface design — your analysis must work across many devices. However, responsive layout and coding for analytics have a surprising twist: You need a new approach not only to interface design but to information design, too.

You may be familiar already with the advice of visualization experts: how to choose color ranges to reveal patterns, which chart type works best for data of particular characteristics, and (often repeated) that you should avoid pie charts like the work of the devil!

Much of this advice, excellent as it is, is predicated on a static representation of data in a predetermined format on-screen or even printed out.

A responsive interface, by its nature, supports many form factors. The carefully chosen scale for your line chart in landscape mode can look deceptive if the user simply turns their device on its side. A legible and useful visualization on a desktop or tablet may turn into an unusable, eye-straining puzzle on a phone. A widget designed for the phone can look absurdly simplistic enlarged on a tablet.

You can solve this with painstaking design and coding, but you shouldn’t need to. Your BI platform, if well-designed for mobile, ought to handle these scenarios generically. No dashboard designer should be forced to handle the complexities of multiformat design.

For example, a mobile BI platform should intelligently scale charts appropriately at any size. A scatterplot may show many data points at a high resolution. At a low resolution, the platform could algorithmically determine the optimal key points to show, still revealing the pattern and range of the data. However, at any size, the user should be able to zoom and select and interact with the chart to see their information from a different “viewpoint” to gain a yet more complete understanding.

Working with a responsive and exploratory design, old certainties about information design start to change. Personally, I find it painful to see a traditional pie chart with many segments and complex labels. What a horrible way to show useful data! However, if you can enlarge or shrink the pie to suit your format, and spin the chart around its center to bring different segments to the top, all the while showing only the labels most relevant to the current view … now the pie chart feels more usable and interesting and informative.

3. Speedy discovery

With a responsive interface you’re on your way to a better mobile analytic experience. Yet some developers worry that a JavaScript platform, for example, will not perform well enough to be “one client to rule them all.” To be sure, users moving beyond static reporting to an exploratory experience need a fast client. If the client does not feel fluid and fast, users can lose interest or lose the thread of their thought when exploring data. Then they lose the value of an analytic application compared to a static reporting solution.

That’s why a mobile BI platform demands great performance on the client. At least use CSS media queries to restrict the visualization and design features that will be appropriate and available for the client in use. And look to complement that responsive approach by building with a framework such as AngularJS. The two-way data binding makes it easy to listen for screen resizes to determine what kind of elements you would like to load efficiently. If you find that this complicates working with the code, use a CSS pre-compressor (such as LESS.js) and compile everything to single file.

There’s another facet of user behavior that may help to squeeze performance from an analytic client. Users following “information scent” mostly make selections that narrow the scope of the data they are looking at, only occasionally moving to a new or broader scope. Looking at U.S. sales, you most likely “drill down” to consider sales in, say, Texas rather than jumping to look at European sales. So caching data on the client can be quite practical, especially if you use JSONPatch to update the array when needed.

The same tendency to start from a higher level and drill down can speed up visualizations using paging. You rarely need to plot 1 million data points, even when rendering a complex data set. The algorithmic approach that enables you to scale visualizations can be used to optimize performance. You may show only key points at first and page in more data as needed when the user selects to explore further.

A natural approach to analytics

As you can see, mobile BI requires that you think beyond mobility. A good mobile analytic platform is fast and facilitates an interactive experience that is responsive and touch-enabled.

These practices not only enable, but encourage exploration. We are compelled by natural curiosity. As we follow it, we make discoveries and form new insights and new understanding.

This is only natural. In business we can only spend so much time at our desks. That’s why we get out into the real world, we go on the road, we walk round the store or warehouse or factory floor in order to see what’s new. Mobile analytics should go with us — and work with us the way we need it to work.

New Tech Forum provides a venue to explore and discuss emerging enterprise technology in unprecedented depth and breadth. The selection is subjective, based on our pick of the technologies we believe to be important and of greatest interest to InfoWorld readers. InfoWorld does not accept marketing collateral for publication and reserves the right to edit all contributed content. Send all inquiries to newtechforum@infoworld.com.

This article, “Getting your hands on data with mobile analytics,” was originally published at InfoWorld.com. For the latest business technology news, follow InfoWorld.com on Twitter.