Zeros to the Left: BI Best Practices
Zeros to the Left: BI Best Practices
Incorporating elements that could be considered aesthetic or not in line with design trends can bring confidence to your BI data.
Join the DZone community and get the full member experience.Join For Free
Hortonworks Sandbox for HDP and HDF is your chance to get started on learning, developing, testing and trying out new features. Each download comes preconfigured with interactive tutorials, sample data and developments from the Apache community.
It was not long ago when computational systems suppressed databases’ use of the century in the date fields to save space in the “so expensive” storage. The same happened when displaying the dates in fixed size for terminal screens (80 columns × 24 rows) or when printing amounts with zero quantities or zero dollars. It was extremely important and economically savvy to code and design systems for better use of the hard disk space, screen real-estate, and printer toner. Millions of dollars were actually saved for not storing, displaying, and printing zeros.
Having all of those technical barriers economically minimized almost to the point of non-importance, should we still see (or not see) the zeros in reports and screens rendered by BI tools? What are the best practices for BI reports in that sense?
Despite the nature of the reports — either informational, analytical, or comparative — it is a must that reporting platforms in the organization share the same data and some basic presentation rules to achieve a consistent global picture. You might think that not everything should be as straightforward as design, decide, and implement. Basic rules should not apply to all reporting. It is well-known that many area managers think they know what they want in reports and like to define how they need their reports formatted. In doing this, a general consensus is necessary in order to avoid having different report standards throughout the different departments within your organization. Usually, best practices focus on one or more of the following areas:
- Customize the report to the target audience/identify report requirements.
- Identify KPIs and metrics/tell the story clearly.
- Choose the right visualization for each metric/highlight key messages/shading or graphical suggestions.
- Layer the information/drill down.
- Take advantage of online capabilities/create a controlled and auditable environment for business users to adjust or manipulate data for reporting.
Gray areas can be found while defining best practices for reporting within your organization — especially when that definition is purely based either on best practices in other organizations or in the literature that fails to differentiate best practices based on their maturity level of the BI in the organization. If we went into even more detail, we would find suggestions like on how to customize the report:
- Keep it simple and short! Shorter is better and for performance monitoring. A one-page report is ideal.
- Don’t overload the user with too many numbers. Pages of detailed data are not a report — despite some managers wanting all the details in a report.
- Discuss report design with managers/users who request the report.
Observing the above short list of report guidelines, we might find them difficult to implement when the maturity level for BI is at the earliest stage. This usually involves:
- A supply of operational reporting in which the main value is business tracking (since their antique application doesn’t supply enough insight).
- Reports that are based in multiple spreadsheets and SQL output focused on basic data access.
- Reports that are still used to upload information into different systems, making the BI solution work as an ETL tool.
- Frontline users looking for a single version of the truth, creating data aggregations outside the BI tool supplied and comparing against other reports or different data sources.
During an organization’s early stages, despite the suggested best practices for reporting, many users require as much data as possible in the report. This is usually because they do not trust the accuracy of the summarized data. In this stage, you want as much data as possible.
Wrapping It Up
Rather than strictly implementing a series of best practices for BI reporting in an organization, sometimes, it is worth it to incorporate elements that could be considered aesthetic or not in line with design trends. This will allow a certain confidence in the data or process (calculations/data load) — even if that means displaying any leading zeroes to the leftof a number with no effect on the number’s value.
Opinions expressed by DZone contributors are their own.