You may think that an expert is an expert. However, not all experts are made equal. If you consider a 5-star Michelin chef, he or she would be an expert at cooking. Eric Grunbaum, the executive creative director behind the ad series that ran between 2006 and 2009 which featured the iconic line “… I’m a mac… and I’m a PC” is an expert in marketing. Although neither of these people are experts in data, they would be better in the data of their related fields than an analyst “off the street”.
What is a SME (Subject Matter Expert)?
A SME is a subject matter expert, as the name implies, this is a person who is an expert in a particular subject. This person knows all about how to run, manage, lead, and/or participate in the execution of the subjects at hand. The chef is an expert in cooking, the preparation steps, cooking times, cooking temperatures, knife skills, seasonings, spices, etc. He or she will know how to prepare the best food imaginable. The marketing executive will know the stories that will engage people, how to arrange the promotions, the mediums through which the promos need to be delivered, etc. These experts and others know their subjects inside and out with creative processes and the right tools to accomplish the end goal.
What is a DME (Data Matters Expert)?
A DME is an expert in data matters. This person could be an analyst or data scientist who has access to data sources (databases, data warehouses, ERP systems, etc.). He or she accesses, analyzes, and presents data. A DME could also be a data team lead who is responsible for leading analysts, understanding, and delivering on data projects. A DME could also be on the technology side leading the development efforts building systems for data to be collected, stored, and accessed. The DME is the person responsible for the data side of a given project or process. This person (or people) know how to turn raw data into actionable insights and decisions.
Why is a SME a better DME?
When the data expert is also a subject expert, this unlocks a world of potential that a data expert independent of the subject would not have access to. The reality of the situation is that when the individual responsible for the data (leader, analyst, scientist, or developer) also knows the subject they understand many of the nuances that are unique to the industry or subject that would otherwise be missed.
Catching the Measurement Steps
With any data project, there are always measurement steps. These are points in the processes within the subject that identify how and what can or should be measured. This is where the developer S/DME really shines and develops systems that are able to capture the important steps in the process that need to be measured. The independent DME would ultimately measure too much or too little. Even if the systems are already established, the independent DME may not identify the points that are measured which need to be highlighted.
Intuiting Expected Results
Any DME that’s working on a data project will present results. Before the project can be considered concluded, the results must be validated for them to be useful. In many cases, one of the stakeholders can intuit results and identify when the results don’t match expectations. The S/DME is able to intuit results without having to connect with any other stakeholder in the project, this saves time and effort speeding up the delivery of the project and making best use of everyone’s time. The independent DME will not be able to intuit these results and therefore will either present erroneous results or take the time of the stakeholders who will ultimately use the tools provided.
Identifying Important Metrics and Measures
One of the most valuable aspects of a data project might be best understood as “data mining”. This aspect is where the DME will review the data and “mine” for insights. These are metrics, measures, calculations, observations, etc. that may or may not be immediately obvious. They may not be established KPI’s but they may be very valuable insights that come from the data. A key stakeholder may ask for specific KPI’s such as dimensional performance (totals, averages, etc.), however, the S/DME will be able to identify measures that are unique to an industry or subject like turn-around-time, in-motion events, or measuring outstanding attributes. Even the independent DME will be able to calculate each of these things, but the S/DME will be the only one who is able to identify that these insights are available from the data and also the best ways to calculate and present them.
Augmenting with Extraneous “Data”
As with any data schema, there are dimensional tables that are available in the existing data sources. These dimensional tables may not capture all of the dimensions needed to bring out all of the insights that are needed by the project. The S/DME will be able to augment the data with extraneous data. Some of this data could be “manufactured”, which are curated based on wrangled data from the existing sources. This manufacutred data could be logical application of categorization labels, on row-level data. This could also be distilling fact-data into unique values and distilling min/max aggregated data for use in extracting a dimension from facts. This could also include augmenting the data with publically or inudstry available data which would be known only to a S/DME.
Whatever the situation may be, remembering that being a SME on a specific subject or area will make a better DME. So, when working on a data project remember the subject matter at hand is important to really understand, even before starting to collect, wrangle, and otherwise handle the data needed for a given project.