Most people in your company are not database experts. They are interested in something other than databases and ELT programs. They want information that supports their work needs. Human beings are visual creatures, and interactive dashboards are visual data expressions.
Dashboards enable everyone to see, understand and interact with data in your database, converting that data into usable information. The dashboards created during your proof-of-concept sprint are the most critical deliverable. It is the first time you will distribute content to a broader audience, and the first time the audience can provide feedback. The dashboard designs are 90% complete. The input received from your target audience will complete the final 10%. This tweaking requires much less time and effort.
Expect some negative feedback. In most cases, this will be the first time many people reviewing the content will pay close attention because they will have a genuine, tangible product to review. Requested changes are ordinarily easy to implement. When conducting dashboard reviews with clients, I immediately implement the changes while they sit with me. Improving designs provides them with a sense of the power of the software. If the revisions are extensive, schedule a follow-up review.
After each sprint, more content will become available to your user base. Executives, managers and staff will use some or all of the content. The results should yield measurable value. Each sprint completed yields more content. More content produces greater engagement and excitement. However, achieving your project goals only marks the completion of the first phase of your system’s evolution.
BI Systems as Data Workflows
BI systems should evolve, grow and adapt as your business needs change. Does your company ever stop growing, changing or adapting? Will you be doing the same things you are doing today in five years? Of course not. Businesses must respond to competitive challenges. Your BI system must adapt as your business needs change. When releasing your first set of dashboards from your POC, your project should be:
- Demonstrating the software works well.
- Training everyone on how to use the software.
- Learning and applying best practices (maintenance/design).
- Explaining data governance and security models.
- Processing issues and opportunities quickly.
You will have three broad categories of user groups:
- Analysts will not only want to consume other people’s dashboards but also want the ability to create their own. The data source may include formally approved data sources, as well as data sources that are outside of your vetting process.
- Heavy users don’t want to spend time finding, cleaning and transforming new data sources. Still, they will want to query data within existing data sources by selecting different measures and dimensions in the published dashboards.
- Information consumers will be your largest constituency. They include managers and executives who need more time or inclination to build their reporting and analysis. They will rely on other people for that. Their needs will be constant for performance indicators, trending, outlier reporting and commentary.
Accessing data can be accomplished with methods besides dashboards. Search-based data analysis is an alternative that is becoming increasingly popular. Using this method provides benefits for distributed teams with more targeted questions. The combination of dashboards and search-based data consumption is an alternative that may be an effective and economical way to support them.
Training User Groups
I’ll cover training in more detail in my next post. Still, analysts, heavy users and casual information consumers should receive nuanced training sensitive to their needs.
Don’t shortchange your roll-out by trying to save money on training. Get experts to do this task. A good designer will create information designs like the dashboard in a BMW—scrolling to see your speed above 55 mph will be verboten! A good consulting partner will demonstrate the best design practices to your team as part of your POC deliverables. A great data architect will know how to use advanced features in the software. The experience and knowledge these experts bring will advance your team’s knowledge and provide a long-term resource they can tap in the future if necessary.
The next post I’ll discuss how to provide training.