Creating a Successful Data and Analytics Program - Execute (Part 5 of 6)

This is part 5 in a 6-part series about creating a successful Data and Analytics program. To read parts 1, 2, 3, or 4, please follow the links below.

• Part 1 - Introduction
• Part 2 - Spell out your ambitions
• Part 3 - Select the organizational structure
• Part 4 - Perform a proof-of-concept

And now on to part 5...

Step 4:  Execute

Step 4.1:  Assemble the necessities

Assuming the items identified in the prior parts of this series have been satisfied, there are 3 necessities when implementing a Data and Analytics program.

First and foremost is the data. Typically, having data is not an issue for most companies. In fact, there is often too much, and that is one of the factors that instigates the need for a Data and Analytics initiative in the first place. However, it is very important to find and understand the data and the business processes behind its creation. A common term you will hear in conjunction with this step is, “source of truth”. This refers to the fact that every single piece of data can be tracked back to the system in which it was originally created. The source of truth should be found for as much data as possible which should then be collected into a central repository in its completely raw, unmodified form. This is the foundation for all downstream analytics, reporting, and other tasks, and gathering the data in this manner ensures the ultimate flexibility in its use.

The second necessity is the platform. To house and work with the data that is collected both initially and on a continuing basis requires a powerful, flexible system. It must be able to scale up and down in terms of both storage and computing power as needs arise while serving multiple use cases as outlined previously in part 4 of this series.

The third and final necessity is the expertise. Without expertise, either in-house or external, the initiative will never take off. There are very specific skillsets, like that of a Data Scientist, Data Engineer, or Data Analyst, that are required to both implement and utilize the platform. Whenever possible, core business and data knowledge should be learned by in-house resources while tasks like configuring and maintaining the underlying infrastructure can be outsourced with minimal difficulty. Due to the popularity of Data and Analytics, skilled local, near-shore, and off-shore external resources are readily available and should be utilized appropriately to ensure the most cost-effective implementation in combination with in-house colleagues.

Step 4.2:  Horizontally expand analytics capabilities

Note:  In this section, we are assuming the implementation of a Center of Excellence (CoE) model as it is generally the model that I prefer and could work well in a majority of situations. Horizontal expansion applies in all cases but would proceed somewhat differently for each. For instance, step 1 below would be important for a fully centralized, but 2 and 3 would be less so. Though, it is still always important to have the users who will benefit from the final product highly involved in its creation. Much like getting kids to eat a proper dinner, people are more likely to use something they had a hand in creating.

When we think about horizontal expansion, perhaps one might picture their waistline around the holidays. In this instance, though, horizontal expansion refers to the introduction and adoption of analytical skills and data literacy across a wide range of the business – well beyond the confines of a CoE. This expansion of knowledge and skills involves members of IT and most facets of the business with everything done under the guidance and supervision of the CoE.

There are three levels of Data and Analytics knowledge and use that apply to this expansion. I've listed them here in order of increasing complexity.

1. Consume and take action – This capability applies to just about every colleague from senior leaders to font-line workers. All the data, analytics, reporting, forecasting models, and prescriptive suggestions are meaningful only if colleagues crafting the business plans and executing the day-to-day tasks learn to consume the reports, dashboards, and forecasts, trust them, and use them to influence their decisions.
2. Discover and share – A smaller number of colleagues will have access to modern tools like Microsoft Power BI, Tableau, etc. to discover data, analyze and create visualizations with that data, and share those visualizations with their colleagues. The CoE ensures that the tools, clean and organized data, best practices, and training are available to empower colleagues to help themselves via this top tier of self-service. The CoE can also certify what has been built to validate that they are using trusted sources of data and appropriate business logic as part of a larger Data Governance initiative.
3. Strategize and wrangle – The volume and complexity of data available to an average business today is generally more than could ever be handled by a single team. Thus, it is critical that a set of Data Champions spread throughout the organization have tools and abilities to contribute to this effort. “Strategize” refers to the act of coming up with the list of key data sources, associating those with business processes, and building and prioritizing the use cases that will best serve their area of the business. These Data Champions take ownership for the planning and delivery of these use cases even if they are not responsible for execution. The work may fall back to the CoE, an external vendor, or other colleagues, but the business must at least own the requirements and the eventual usage. “Wrangling” is a somewhat recent term in the Data and Analytics realm coined to describe the act of gathering, understanding, cleaning, and organizing data. Modern tools like Informatica, Trifacta, Paxata, etc. provide interfaces for Excel-jockeys to work in a familiar environment but with potentially terabytes or petabytes of data behind the scenes. It is important to remember that wrangling is a significant undertaking that can consume 50% of all time spent on any given use case. This effort also involves the Data Champions taking responsibility as stewards of the data that falls in their area of the business. Similar to bullet 2 above, the output of wrangling can be certified and operationalized by the CoE as part of a Data Governance initiative.