There is no shortage of data in today’s inter-connected world. Companies have recognized this for a while now and have invested significantly in capturing, mining and reporting on this data. But are they satisfied with what they are seeing? The answer is a resounding NO. And why is that? Because of two insidious human elements. One is the quest to identify or extract truly meaningful data and the other is to create an organizational structure that explicitly addresses the need to be data-driven. I will examine both of these in an attempt to suggest more meaningful and methodical changes vis-à-vis data and its vast horsepower.
Data, data everywhere but not a drop of meaning!
Data has the ubiquitous quality of appearing ‘official’ and ‘meaningful’ in whatever context it is presented. Intuitively we know that this is not correct. For example, just because there are data points which are taken at earlier and later points in time, it does not mean we can derive cause and effect relationships between them. In fact, where incorrect or meaningless correlations are made, the result is detrimental to ‘right action at the right time’ and are difficult to challenge. It is my submission, therefore, that experts such as social scientists for social media data/big data, financial specialists for financial data etc be employed to interpret and put intelligence to the data. It is important to note that this requires processes to allow these experts to derive meaning only from data relevant to their area of expertise and prevent them from having a massive lateral view of all data. Yes, data restriction is vital. In order to provide a wide enough view of data to the specialists that allows them to keep the ‘bigger picture’ in mind and narrow enough to not cloud their view, it is necessary to involve them in the design and deployment of all stages of data processing – collection, storage, mining, reporting.
Creating a culture of data
Companies have been oscillating between being numbers driven and gut-feel as they experience the phenomenon of meaningless or misleading statistics/data points and the instinctive need to be objective. In order to promote a culture that is truly acting on intelligent data the following data related roles are recommended:
1. Chief Data Officer: The role must report up to the highest level in order to make deep changes. The person must use a top-down approach starting with strategic measurement areas and mapping these progressively down the food-chain to establish KPIs and reporting needs for each department/function
2. Data Analyst: The role must be a data management expert with knowledge in data modeling, database design, ETL principles, data analytics principles. This role should operate at the division or function level to help decompose the tier 2 KPIs and reporting needs in line with strategic needs defined by the CDO. Data Analysts should report up to the CDO
3. Data Specialist(SME): The role is the expert defined in the section above. They are Subject Matter Experts in a field such as Social Science, Accounting, Marketing, Finance etc. They are NOT data analysts. Rather, they are folks who know their subject area and can interpret data to draw meaningful conclusions and avoid spurious correlations. They are also the 'seekers' of relevant data and determinants of which data is relevant and which data is not given a specific area of analysis. This role should work with Data Analyst to create a full map of source and destination of data (destination could be reports or corrective action). Data Specialists should report up to the functional or departmental heads and have dotted line reporting to CDO’s office in partnership with Data Analyst. Some of these Specialists will be function-specific (such as social scientist for big data) and some will be department focused (such as Marketing).
4. Data Process Specialist: With function and department heads and in collaboration with Data Analysts, this role will define the nature of reporting, processes for corrective/proactive action and police compliance with data-driven communications and reporting throughout the organization. This role can have the role of evangelizing the CDO's mandate of driving processes based on objective interpretation of data.
Finally, a word on a current data-related role in many organizations - the Data Scientist. Their job function varies from one company to another and is a mish-mash of two or more of the roles described above. It is important to note that Analyst, SME, and Process Specialist are three distinct areas of expertise and rolling their functions into one title "Scientist" is short-changing the effort to create a data-driven culture.