The Evolving Data Landscape
Starting with the historical milestones that made today’s data abundance possible, we will appreciate how businesses and decision-making have evolved with technological progress. We will be reminded that every invention started with a need, including data-driven innovation. We will look at data through various lenses. We will define relevant attributes when dealing with large amounts of data and the unique challenges they impose. We will set the stage to understand the challenges and opportunities from the abundance of data available to businesses.
Key Takeaways:
- Technology advancement is responsible for the current data abundance
- Working with large amounts of data requires new thought leadership, mindsets, and tools
- Digital businesses have an inherent edge when it comes to data-driven decisions
- Traditional businesses must adapt and plan to leverage data to avoid disruption within their industry
Business Models for Monetizing Data
There are many ways businesses can monetize the use of data. Once we explore various revenue models, we will distinguish between businesses in which data is the product and businesses that are data-driven. We will examine the goals and maturity models for data-driven businesses. We will conclude by understanding how a business can decide if data monetization is the right strategy.
Key Takeaways:
- Data as a business and a data-driven business are two very different entities
- Data-driven businesses strive for customized solutions for their customers
- There are many business models for data as a business with various pricing strategies
- Businesses, before executing, should have clearly defined data monetization strategies
Business Problem Formulation and Data Collection
Data-driven management starts with identifying the opportunities and problems that are most important to organizations. The next logical consideration is stakeholder identification – both decision makers and those impacted by decisions. Then, an analytical project plan can be created with an understanding of the opportunity, the current approach to decision-making, and what assumptions are in play. An analytical plan consists of questions to explore, conditions to test, and data sources necessary for analysis. We will also contemplate our options in the absence of required data.
Key Takeaways:
- Planning is a time-consuming but important step
- A data-driven decision-making process should be applied to the opportunities and problems most important to the organization
- Understand what opportunities are ripe for data-driven decisions
- Learn to recognize the opportunities that data cannot currently solve
Exploratory Data Analysis
The journey toward modeling, predicting, or leveraging data for making decisions starts with getting familiar with data. In this module, we will learn techniques of data exploration. Specifically, we will summarize data, reveal trends or seasonality, and investigate relationships between variables within different data sets. We will identify events that are outside the normal range based on aggregation. Exploratory data analysis aims to decide if the data is sufficient to answer our stakeholders ' questions and make decisions.
Key Takeaways:
- Eliminate data when it does not apply to a project
- Ensure data is complete and correct before building decision models
- Visualization is an efficient tool to understand data in aggregate
- Accurately predicting with data is only possible after becoming familiar with data
Probability and Statistics with Small Data
We will start building models with our data. We will learn the statistical concepts of likelihood, hypothesis testing, confidence intervals, and regression modeling. We will explore concepts of observational studies and randomized control trials (popularly known as A/B testing) that apply to setting up and conducting an experiment.
Key Takeaways:
- Understand that earlier statistical methods were developed to aid decision-making using limited data.
- Randomized control trials can be expensive, but they are the best evidence to understand the effect of a variable on an outcome.
- With the explosion of mobile and web applications, A/B testing is frequently applied to determine the effectiveness of options.
Machine Learning and Big Data
The field of machine learning, under the umbrella of Artificial Intelligence (AI), is one of the techniques used to explore large amounts of data. Machine learning is leveraged to detect data anomalies, classify data, find associations and patterns, and make reliable decisions. We will examine how companies leverage the above techniques to gain a competitive advantage and build new products.
Key Takeaways:
- Machine learning is a rapidly growing field within AI that trains machines to perform specific functions using large amounts of data.
- Certain problems are better suited to use machine learning successfully.
- Many problems cannot yet be solved by leveraging large amounts of data and technology.
- The excitement about AI is around its potential rather than its present applications.
- Advancements in AI are a significant cause of disruptions to traditional businesses.
Data Visualization and Communication
Analysis to Insight to Action! Data presentations can fail to lead to action if they fail to achieve management buy-in. Audiences buy in if you communicate findings succinctly, reveal unexpected insights, and “money-tize” them; tie them to business outcomes that matter to your audience.
Key Takeaways:
- Analytics is a team sport. Review analysis and presentations with your team before presenting your work to your stakeholders.
- You must explore data before you can communicate what you found and why it matters.
- Data visualizations-graphs and charts-use pre-attentive attributes to turn complex numbers into instantly recognizable and understandable patterns and trends. Numbers as text take time to read, process, and understand.
- Make communication with leaders clear and concise. "Money-tize" the results. Use the data-story technique when appropriate.
Model Creation, Operationalization, and Maintenance
The engineering side of data-driven decisions incorporates all activities required to produce insight from models to drive decisions periodically. We will explore how to gather data from different sources, transform data across sources into a single format, and make it ready for model consumption. Once the model runs, we will consider how to feed it to systems that will take the insight and make it available for users. Lastly, we will explore how to ensure the model is always relevant – employing practices to proactively produce alerts to warn if model results need to be reviewed.
Key Takeaways:
- Data is created at a rapid rate and decays just as quickly
- Model maintenance is expensive but critical to ensuring decisions are made from the most accurate data
- Data-driven decision-making requires both data analysis and data engineering
- The best data-driven companies not only plan for model maintenance but also consider it an extremely important part of analytical projects
Data Governance, Ethics, and Privacy
As a society, we face an intrinsic challenge due to the explosion of data and ease of accessibility. For instance, we protect our medical records even though we recognize society would benefit from an aggregate analysis of medical records. In this session, we will understand ethics as applied to data. We will develop a framework for analyzing concerns as they relate to data. We will also look at specific laws around data privacy. Lastly, we will explore data ownership and the rules around data accessibility and privacy protection.
Key Takeaways:
- Understand data ownership versus data as a public good; transparency and openness versus privacy and security
- The importance of informed consent
- Systematic biases often exist in data-based algorithms
- Develop a code of data ethics
Building and Managing a Data-Driven Team
Starting a data-driven decision strategy in an organization requires a team. We will explore the managerial functions of setting group structure and strategy, identifying roles and functions of team members, and ways to recruit, interview, and retain team members. We will also discuss common hurdles faced daily by managers of data-driven teams.
Key Takeaways:
- Team structure should be based on the organization’s size and needs
- Data engineers and data analysts are unique roles with different skill sets
- The role of a data manager is to ensure the group is communicating with each other and with stakeholders
- The data manager is responsible for and should often educate the organization on data-driven strategies
Program Overview
For an overview of our Mini-MBA: Data-Driven Management program plus program benefits and outcomes, please click here.