Business analysts and business data analysts exist to figure out: What change should we make, right now, to our processes in order to improve [insert important metric]?
Metrics are (circularly) defined as the numbers that the business can impact by changing its processes.
Three categories of metrics:
- Revenue - Important to Sales teams
- Profitability - Important to Operations teams
- Risk - Important to Finance teams & CEOs
Two types of metrics:
- Traditional - standard financial and managerial accounting categories
- Dynamic - metrics that can change within a month or less, and which are sensitive to small changes to business processes
Big Data in business means that information we used to physically write down and then throw away is now stored electronically. Data storage costs have decreased to the point that we don’t need to “throw it away” anymore. Data Analytics is the practice of searching, exploring, and analyzing this accumulating data in order to derive actionable information.
Daniel Egger says “No commercial for-profit company that is in a competitive market can remain profitable or even survive over the next five years without incorporating best practices for business data analytics into their operations.”
The life cycle of a data problem includes the following steps: assembling the data, cleaning it, analyzing it, and communicating to people what it means. Data analytics professionals should be able to navigate this entire process, clearly articulating results and progress to relevant stakeholders, be they programmers, statisticians, business domain experts, or executives. The relevant tools for these stages are SQL for assembling data, Excel for data modeling, and Tableau for understanding and communicating data.
To this end, the courses in this specialization are the following.
- Understand and identify the relevant business metrics,
- Use Excel to do business data analysis,
- Leverage Tableau to create powerful dynamic data visualizations and present business proposals, and
- Retrieve and analyze big data from large industrial-sized relational databases.
- Finally, a final capstone project ties the learning in the first four courses together.
Some of the objectives for this course include:
- describing best practices for data analytics to make organizations more profitable and competitive,
- distinguishing critical business metrics from mere data,
- describe different roles business analysts, business data analysts, and data scientists play, and
- describe the skills required to succeed in each of those high-demand jobs.
The goal of module 2 is to determine which data is irrelevant, and which forms part of an important metric. Once derived, business metrics can be classified into one of three main categories of data: revenue, profitability, or risk. Additionally, metrics can be classified into traditional metric and newer “dynamic” metrics.
Module 3 consists of elaborating what types of data employees are needed in different business types. This will also consist of defining typical job titles, skill requirements, and how the various roles typically interact. Among the job titles are business analysts, business data analysts, data scientists, and software engineers. Module 3 will also classify companies into one of five categories, and discuss how each category is reacting to big data. This will involve a 20 item checklist.
Module 4 will identify current corporate best practices in analyzing business metrics. It will also establish some powerful formulas to extract maximum value from those metrics. In particular, a critical horizontal business area web-based marketing will be examined. And the vertical market of financial services relating to investment management will be analyzed as well. Market sectors will be examined, as defined by a group of metrics.
The overall goal is to set the stage for a successful career as a business analyst or business data analyst, or set the stage for further study on the way to becoming a data scientist. In particular, this course will enable those who complete it to distinguish the signal from the noise when it comes to business data.
Metrics are for Asking the Right Questions
Business analysts and business data analysts exist to figure out the right questions, and then to answer them as well as possible with the resources and time available. The answer should have practical impact, and result in a specific call to action.
The right question is: What change should we make, right now, to our processes in order to:
- increase revenues,
- maximize profitability, or
- reduce risk.
The definition of “right now” can vary from real time, in a computer system, that optimizes for dynamic metrics. In that case, right now means within fractions of a second. “Right now” can also mean just in time, which means that a human being is involved, and takes some action inn respnose to the data. Finally, “right now” can mean as soon as conclusive test results are available, and the change has been processed. The specific context for this would be in A/B testing for a website.
Metrics are defined as numbers that the business can impact by changing its processes. Some things the company cannot change, like interest rates on a loan or the tax rate. Other things, like the percentage of people who viewed an add and then clicked through to the company’s homepage, can be thought of as a metric.
Distinguishing Revenue, Profitability, and Risk Metrics
Who cares about each of these categories of metrics?
|Revenue||Sales and Marketing|
|Profitability||Logistics, Production, and Operations|
|Risk||Risk managers, creditors, outside investors|
Revenue metrics are outward facing. How many units of each product type were sold over a given interval? How does this compare to previous periods? How do these sales vary by product, region, and by existing versus new customers? Key executives monitoring Revenue metrics include VP of Sales and VP of marketing. These metrics could also include information about the “sales funnel,” which describes where potential future customers are in the process of moving towards their purchase. Marketing would be interested in how effective any marketing campaigns might be. How many people have seen a given advertisement? What percentage have taken some action? In summary, everything related to selling falls under the revenue metric category.
Profitability metrics are related to the efficiency with which a company creates and delivers products and services. These metrics are operational in nature. How much cash is tied up in the form of unsold inventory? How much inventory is spoiled/ruined/defective? Cost breakdowns per unit of production: variable costs, raw material, labor. Large, established companies without much ability to increase revenues can improve profitability by increasing operational efficiency
Risk metrics relate to tracking and mitigating dangers a company faces. Net cash out is the most important metric to track. How many months can the company survive at the current “burn rate?” Another metric is the “churn rate,” which is applicable to companies that operate on a subscription model. The churn rate is the rate at which new subscribers cancel within a year. High churn rates result in the risk that there will be fewer unreached consumers in the future, and so historical revenue growth rates many not be able to be maintained. Other risks are specific to the financial industry. These types of metrics include how much exposure a bank may have to potential customer defaults, how many customers are expected to default in the next six months, or how many are currently in default. Other more technical risk metrics include the “volatility of returns” and “maximum historical drawdown from high water mark.” Most risk metrics are related to leverage.
According to Daniel Egger: “Revenue metrics are for optimistic extroverts, profitability metrics for fastidious perfectionists, and risk metrics for informed skeptics.” The best companies will track all three categories.
Distinguishing Traditional and Dynamic Metrics
Traditional business metrics are standard financial and managerial accounting categories. Examples in cldued quarterly net cash flow, balance sheet item changes, and profits and losses. These metrics are acted on, but only slowly, and after deliberation.
Modern data science is more interested in Dynamic business metrics that convey urgency. These metrics share two attributes:
Significant change over month-long, or shorter, time intervals.
Monthly rent on a single store’s 3-year lease is not dynamic. The average monthly rent per square foot for a national retail chain with 1,000 stores that signs seven new three-year leases each week would be very dynamic.
Strong impacts by small business changes.
Dynamic metrics display “twitchy” (small changes in process make a big impact), not “noisy” (impacted by many, many factors), behavior. An an example, the percentage of people who purchase items they put in their online shopping cart is very sensitive to average page load times. If a page takes longer than three seconds to load, 40% of web users will abandon it. This “sales completion” percentage is a metric that is very twitchy, because it is very responsive to the average page load time.