One of my colleagues made an offhand comment the other day that stuck with me.
We were talking about dashboards, and he said (paraphrasing here) that dashboards can be bad for users. Given that our employer makes and markets products for creating BI dashboards, his statement bordered on treason.
But then he went on to explain his issue with dashboards: they only show the surface of the data. My colleague, a software engineer, likes to delve deep into data. He maintained that skimming just the surface could invite problems, including the possibility of drawing incorrect conclusions.
Skin, Meat, and Bones
His remarks got me thinking, and then it hit me: BI is like fried chicken. With apologies to the vegans and vegetarians out there, let me explain why.
Fried chicken has three components: skin, meat, and bones. In this analogy, the skin is the front end of a BI system – a dashboard or other UI, for example – and the first thing most people encounter. The meat is the underlying data and the analytic engine, which lay immediately under the skin and which make up the bulk of the system (errr, chicken). The bones are the architecture and infrastructure that hold it all together.
To take the analogy a little farther, the skin is visually appealing and tasty, and people gobble it up. But it’s not the whole meal, or at least shouldn’t be, because the meat is where you find the real value – the nutrition, you might say. However, if it’s not prepared properly, the meat can be dry, tough, and difficult to swallow. The bones, meanwhile, are necessary to keep it all together; they may not be consumed, but the whole package couldn’t exist without them. (Fans of Gary Larson’s The Far Side may remember his comic about boneless chickens.)
Getting Users to the Meat
Most people who develop and manage BI systems share a goal: to get their users past the skin and into the meat. But how do they provide tasty skin to entice users and then deliver enough hearty meat to inform them? And how do they make the meat consumable? These seven strategies and ideas can help satisfy even the pickiest consumers.
- Practice good design. Your front end (the skin) should match your branding and UI, especially if you expose parts of it to your customers. It also has to be intuitive and straightforward to navigate and use, with the right types of charts (not just the best-looking ones) and no chartjunk or distractions.
- Enable drill through/drill down. This is where the skin and meat connect. Users need to move from dashboards and reports to the underlying data with a click. That ability must be aligned with your data security architecture, meaning that drill-down involves everything from the skin to the bone.
- Get analytical. A great way to get users to dig into the meat of data is to give them access to analytics tools. Once the domain of data scientists, analytics algorithms and tools (such as Pareto charts, Decision trees, Holt-Winters forecasting, and K-means clustering) are increasingly being built into BI and analytics systems. In the hands of people with deep subject matter expertise, these tools can deliver real value from business data.
- Make it perform. Busy people will avoid or ignore a slow BI system. Make sure you’ve got the architecture you need – that is, a robust skeleton, a good set of bones. A dedicated, pre-aggregated analytics data source – often a columnar database – helps deliver high performance, especially when using extremely large datasets.
- Embed it. When BI is embedded in applications and portals, it becomes part of users’ everyday work lives. The 2014 Embedded Business Intelligence Market Study from Dresner Advisory Services notes that embedding “falls into the top third of initiatives considered strategic to business intelligence.” Embedding gets to the bones of our fried chicken analogy, because it depends heavily on your BI system’s underlying architecture.
- Move along. If you really want to get people into the meat of your BI system, create apps for mobile phones, tablets, and even wearables. But remember, mobile isn’t a single platform. Charts and reports suitable for a large tablet are useless on a wristwatch screen, so you need to understand both the devices and the use cases, and make sure your BI system’s bones can support them.
- Be useful. Above all else, help your users get the information they want and perform the tasks they need. Their needs may not always overlap with things an enterprise BI system does naturally (especially in an era of multiple business applications and disparate data sources), so it’s important to have a flexible platform with APIs and data connectors that can communicate with the rest of your data ecosystem.
Seeing a BI system like fried chicken makes sense to me. Those who work with BI need to make sure that the skin appeals to users and invites them to consume more, that the meat nourishes them and is easy to digest, and that the bones are strong enough to support it all.
Does this analogy work for you, or would you offer a different one? And what other ideas can you suggest for getting users into the meat of your BI system? Please share.
About the Author: Fred Sandsmark is a Marketing Content Writer employed by Actuate, The BIRT Company. Before joining Actuate, he wrote success stories and marketing materials for Oracle, Symantec, NetApp, Microsoft, and others as a freelance. Read more of his work at blogs.actuate.com, and follow him on Twitter at @FredSandsmark.