TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

4 Steps to BI Self-Service Success

Master these four key areas to successfully create presentations or display findings with self-service BI tools.

• By Julianna Cammarano
• September 27, 2016

Business tools help users pull up data at a moment's notice to create diagrams, maps, and charts that enable stakeholders to pinpoint unexpected risks, recognize untapped opportunities, and visualize trends. According to Gartner, the business intelligence and analytics market will reach $16.9 billion in 2016 as end users seize control of their BI destiny.

For all of these benefits, few BI evangelists mention the elephant in the room: self-service BI does not do your thinking for you. There are at least four key areas users must understand, assess, and address to successfully create presentations or display findings with self-service BI tools.

1. Identifying Problems

When compiling business intelligence for a presentation, it is vital to narrow down the information to highlight a single problem.

Such an issue could be operational, such as a manufacturing plant that may require additional funding or a call center may have a high call volume, queue time, or call time affecting its bottom line. The issue could also be analytical, such as uncovering trends and outliers in a field or using BI to investigate how to change a business's current state.

Some business intelligence is compiled to elicit an executive response, giving company leaders a higher vantage point from which to notice opportunities and obstacles.

2. Highlighting Solutions

Once a problem has been identified, the presentation should combine data and best practices to propose a solution. To make this solution clear, analysts should understand how their company organizes information and determine whether they will use static or real-time data, keeping in mind how real-time data collection impacts the performance of queried applications. Some data is simply noise and BI tools often provide rudimentary operations to combine and filter data sources so only the important data is highlighted.

3. Knowing Your Audience

Presenters should approach BI presentations as if delivering information to a new employee, displaying all data to be easily understood even without prior context or in the presenter's absence. The audience may see the data on a mobile device, so presenters must test how their dashboard displays on various screens.

Reading habits and symbolism also vary by culture and location, so presenters should be conscious of such things as whether their audience reads from left to right and what phrases and images viewers might find offensive.

4. Presentation and Communication

The dashboard's overall design should clearly tie into the problem the analyst is trying to solve. Elegant simplicity is attractive and impactful; information overload must be avoided. Borders, bolding, and lines should be removed whenever possible. Subtle shading is preferable to bright colors, and labels and extraneous text should be kept to a minimum. The table below offers a list of basic formats, but presenters should feel free to move beyond these once confident in their use.

Goal	Format
Compare values between categories of data	Bar chart
Identify a trend or pattern by checking value over time	Line chart
View relationships and distributions among values	Scatterplot
View data related to geography	Geo maps
Compare parts of a whole	Pie charts

Many BI tools feature built-in collaboration, which can be useful for gathering input about a presentation's readability and clarity. Such feedback is excellent for rapid and efficient iteration.

A Final Word

Self-service BI solutions come with a hidden price tag: the cost of recalibrating how you think and work with data and design to successfully communicate your business story. If paid in full and taken seriously, this investment will pay huge dividends over time.