Sitting at the intersection of business and technology means that I can interact with both groups. As a project manager who specializes in data projects, this often means diving into the analysis of the data, for validation, or understanding requirements, or more importantly to help craft a story, preferably around project success. Companies are becoming more data savvy, hiring Chief Data Officers and data scientists, and focusing on numbers to drive business decisions. I am very supportive of this, but we need to be careful to remember that it’s not just about the numbers.
More important is how we use these numbers to tell our stories. If your customer presents you with a problem, and a set of data that supports their conclusions, you need to look at it end to end to determine how to approach your story. Here’s a few thoughts on how you might approach it.
- What were the assumptions made? In lieu of other information, we start with a core set of assumptions. It’s good to recognize and document these assumptions. It’s also good to challenge the assumptions if you have information not previously considered.
- Evaluate your inputs end to end. Start at the beginning of the process and evaluate each step to understand the decisions made, and how those decisions impacted the outcomes.
- Evaluate the outputs. There are so many opportunities to introduce uncertainty and inaccuracies into human process (data or otherwise). Was the data entry manual? Is there evidence of irregularities? Does the output make sense in relation to the logic of it’s corresponding process? Systematic logic is still at the whims of the people who write it.
- Challenge the conclusion. While it’s important not to be defensive here, you do need to consider the “foregone conclusion” and make sure it withstands probing and prodding. This is where perspective and point of view comes in. Make a conscious determination of the story you are telling based not the facts.
- Communicate the deficiencies. If you go from one extreme to another extreme, swapping the entire frame of reference, it raises questions and credibility issues. As you are diving in and crafting your story, make sure to highlight deficiencies and inefficiencies of all components. Try to be balanced in your analysis, taking into account the work your customer took before it came to you.
There are lots of methodologies being used for diving in to real root causes (i.e. 5 whys, fishbone, etc). In data projects, it’s about being able to understand what the data is telling you and being able to convey that in a coherent and concise manner. Please don’t give me a bunch of facts, without an explanation of how you got there. This is the story. Understanding that data needs explanations, in the form of stories, is why I do so well in managing my data projects.