The Soft Side of Data Work.

Kimball and Ross describe the responsibilities of the data warehouse / business intelligence manager as follows:

Understand the business users.

• Understand their job responsibilities, goals, and objectives.
• Determine the decisions that the business users want to make with the help of the DW/BI system.
• Identify the “best” users who make effective, high-impact decisions.
• Find potential new users and make them aware of the DW/BI system’s capabilities.

You complete this step by interviewing people. Interviews are often eye-opening not only for the person asking the questions but for the person answering them as well. Do not allow vagaries to go unchallenged at this stage. Everything that can go wrong cascades from what has been achieved at the end of this step.

You identify the best allies in the organization by their interest, enthusiasm, and curiosity in the project. Regardless of influence, disengaged business people are unlikely to be useful in architecting satisfactory solutions.

The easiest ways to find new people to evangelize your data projects to is to follow the trail of discontent in the organization. The people who express frustration or lack of support in doing their work are the best candidates to involve in data projects as they stand to gain the most, at least spiritually, from a successful outcome.

Deliver high-quality, relevant, and accessible information and analytics to the business users.

• Choose the most robust, actionable data to present in the DW/BI system, carefully selected from the vast universe of possible data sources in your organization.
• Make the user interfaces and applications simple and template-driven, explicitly matched to the users’ cognitive processing profiles.
• Make sure the data is accurate and can be trusted, labeling it consistently across the enterprise.
• Continuously monitor the accuracy of the data and analyses.
• Adapt to changing user profiles, requirements, and business priorities, along with the availability of new data sources.

This is where most data projects fail. Presenting information is easy, making someone care is hard.

The standard of data quality you will attain is a direct function of the skill level of people working on the project. Most visualization tools lack templates and require you to build visualizations ad hoc, from scratch, every single time. Few tools will let you program how to create visualizations. At that point, you’re better off just making your graphs with a programming language.

Ironically, sending people emails with summary numbers and PDF graphs is the lowest bar to get people interested in the organization’s data. If you can’t make someone raise their eyebrows with such a low bar, you have no business architecting dashboards.

Sustain the DW/BI environment.

• Take a portion of the credit for the business decisions made using the DW/BI system, and use these successes to justify staffing and ongoing expenditures.
• Update the DW/BI system on a regular basis.
• Maintain the business users’ trust.
• Keep the business users, executive sponsors, and IT management happy.

Taking credit for things you’ve done well is the greatest social hack to gain recognition. Be relentless against anyone who tries to take credit for your work. It’s likely that your colleagues won’t be technically inclined nor interested in the gritty details of the data work you do. The fastest heuristics they’ll use to judge your achievements are (1) other people’s claims about you (2) your own claims about yourself. Be in control of both.

I find that maintaining trust is easier when you’re dealing with trustworthy, well-meaning people. It’s counterproductive to be trusted by grifters or opportunistic individuals. Trust is achieved by helping people out and delivering on your promises.

References