Fundamentals of Business Analytics

TL;DR

Business analytics helps you make better decisions by using data. It's about turning raw information into valuable insights. You'll learn to ask the right questions, analyze data, and communicate findings effectively.

1. The Mental Model

Think of business analytics as being a detective for your business. You gather clues (data), examine them closely (analyze), and then figure out what happened, why, and what you should do next.

2. The Core Material

Business analytics combines data, technology, analytical methods, and business knowledge to drive better decision-making. It's not just about crunching numbers; it's about understanding the "why" behind the numbers and what they mean for your business.

2.1. Types of Analytics

Photo by Markus Winkler on Pexels

There are generally four types of analytics, each building on the last:

• Descriptive Analytics: What happened? This is about summarizing past data. Think of sales reports, monthly revenue figures, or customer demographics. It provides a picture of the past.
• Diagnostic Analytics: Why did it happen? Once you know "what" happened, diagnostic analytics helps you dig deeper to find the root causes. For instance, if sales dropped, diagnostic analytics might show it was due to a specific marketing campaign ending.
• Predictive Analytics: What will happen? This uses historical data to forecast future outcomes. For example, predicting future sales based on past trends, or forecasting customer churn.
• Prescriptive Analytics: What should we do? This is the most advanced type. It not only predicts what will happen but also suggests actions to optimize outcomes. For instance, recommending which products to promote to maximize profit.

Here's how these types of analytics flow:

graph TD
    A["Descriptive Analytics (What happened?)"] --> B["Diagnostic Analytics (Why did it happen?)"]
    B --> C["Predictive Analytics (What will happen?)"]
    C --> D["Prescriptive Analytics (What should we do?)"]

2.2. The Business Analytics Process

Photo by RDNE Stock project on Pexels

While specific steps can vary, a common process looks like this:

1. Problem Definition: Clearly understand the business question you're trying to answer. What decision needs to be made?
2. Data Collection: Gather all relevant data from various sources (databases, spreadsheets, web, etc.).
3. Data Cleaning & Preparation: Raw data is often messy. You'll need to clean it (handle missing values, correct errors) and prepare it for analysis (format, combine). This step often takes the most time.
4. Data Analysis: Apply analytical techniques (statistics, data mining, machine learning) to extract insights. This is where you find patterns, trends, and relationships.
5. Interpretation & Visualization: Make sense of your analysis. Present findings clearly using charts, graphs, and dashboards to tell a compelling story.
6. Recommendation & Action: Based on your insights, propose actionable recommendations to solve the original business problem.
7. Monitoring & Evaluation: After implementing actions, track their impact and adjust as needed.

3. Worked Example

Let's say a local coffee shop, "The Daily Grind," notices a recent dip in afternoon coffee sales.

1. Problem Definition: Why are afternoon coffee sales declining, and what can we do to reverse this trend?
2. Data Collection: We collect sales data for the last six months, including time of day, product sold, weather, and any marketing promotions run. We also check social media mentions and local event calendars.
3. Data Cleaning & Preparation: We find some missing weather data and correct a few typos in product names. We merge sales data with external weather and event data.
4. Data Analysis:
   • Descriptive: Afternoon sales (2 PM - 5 PM) have indeed dropped by 15% over the last two months.
   • Diagnostic: By comparing sales to local events, we notice a new park opened nearby three months ago, which has a competing coffee cart. We also see a correlation between hotter afternoons and lower coffee sales, possibly suggesting a preference for cold drinks.
   • Predictive: If current trends continue, afternoon sales will drop another 5% next month.
   • Prescriptive: Analysis suggests offering an afternoon "iced coffee special" on hot days and promoting it heavily near the new park, perhaps with flyers or a small discount for park visitors.
5. Interpretation & Visualization: We create a line graph showing the decline in afternoon sales, overlaying it with the park opening date and average daily temperatures. We also make a bar chart comparing sales with and without promotions.
6. Recommendation & Action: Recommend implementing an "Iced Coffee Happy Hour" from 2-4 PM on days above 75°F and launching a "Park-Goer Perk" discount for those showing proof they visited the new park.
7. Monitoring & Evaluation: The shop implements these changes, then monitors sales for the next month to see if the strategies are effective.

4. Key Takeaways

• Business analytics is a structured approach to making data-driven decisions.
• Understanding the four types of analytics (descriptive, diagnostic, predictive, prescriptive) helps you frame your analysis.
• The business analytics process involves defining a problem, collecting, cleaning, analyzing data, and then making recommendations.
• Data cleaning and preparation are crucial steps that often take a significant amount of time.
• Effective visualization is key to communicating your analytical findings clearly to others.
• You're using data not just to know what happened, but to influence future business outcomes.

• Start with a clear business question before diving into data analysis.

• Don't assume raw data is clean or ready for analysis; it rarely is.

• Avoid performing analysis without a clear business objective in mind.
• Don't just present numbers; tell a story with your data and offer actionable insights.
• Don't ignore the "why" behind the data; always seek the root causes.

5. Now Try It

Think about a simple business problem you've encountered recently (e.g., why a certain product isn't selling well, why a website has low traffic). Apply the first three steps of the business analytics process to it:
1. Clearly define the problem.
2. List at least three specific pieces of data you would collect.
3. Describe two potential data cleaning or preparation tasks you might need to do for that data.

Success looks like a clearly articulated problem, a concrete list of relevant data, and plausible data preparation steps, all within 15 minutes.