6  Lab - Business Metrics Practice (optional)

Note

If you’re interested to explore further, you can dive into this case study building a dashboard for visualizing business metrics for ad campaigns.

6.1 Business Metrics

The most important question for decision-makers is: What changes can we make in our business processes right now to increase revenue, improve profitability, or reduce risk?

Business metrics provide the foundation for answering that question. By tracking and analyzing key metrics, generally grouped into revenue, profitability, and risk, we can identify where change is needed, measure its potential impact, and guide smarter decisions for immediate and long-term improvement.

• Revenue metrics focus on sales and marketing performance — they’re outward-facing and show how effectively a company generates income.
  • Everything that relates directly or indirectly to selling is a revenue metric.
• Profitability metrics measure how efficiently a company turns resources into profit through production, logistics, and operations.
  • Examples include how often the company is unable to meet urgent customer requests and loses sales because of insufficient production or inventory.
  • These metrics highlight how effectively the business controls costs and minimizes waste to maximize profit.
• Risk metrics assess financial stability and exposure, mainly used by risk managers, creditors, and investors. Such as:
  • Net cash out: how many months can the company survive at the present burn rate?
    
  • Churn rate.
  • Debts.

A good company should track all three types of metrics.

ImportantDynamic Metrics

It’s important to remember that useful business metrics are dynamic — they may change over time therefore they should always be measured and interpreted within a specific time frame.

6.2 Lab: Business Metrics Dashboard

Please download and read the case description:

• Case Study Business Metrics - Ad Analyses.pdf
• ad_table.csv

The goal of this lab is to look at a few ad campaigns, analyze their current performance, identify key business metrics and create a dashboard to report them.

6.2.1 Requirements

From a business perspective, specifically, you are asked to:

• Identify critical business metrics and distinguish them from mere data
• Calculate the metrics that are important and worth reporting
• Design a streamlit dashboard to display these metrics interactively
• Include explanations alongside each metric using text components
  • Ensure users understand what each metric represents and how it reflects overall business performance.
  • In short, don’t just calculate metrics — demonstrate that you can interpret their meaning and implications for decision-making.
  • Explanations can be dynamically generated using an AI-powered feature.
• Clean and well-organized code following the recommended workflow below.

Tip

Developing key business metrics requires domain knowledge — understanding what drives the business. You can use EDA (Exploratory Data Analysis) and AI-powered Q&A to uncover key metrics worth tracking and reporting in the provided data.

6.2.2 Recommended Workflow

A structured workflow helps you move efficiently from exploration to production. It turns messy experimentation into reproducible, professional pipelines for real-world data projects.

Step 1. Rapid EDA in a .ipynb

• Create and Use a eda.ipynb notebook to inspect data quality, clean data, develop metrics, test feature ideas, and prototype formulas and metrics.

Step 2. Refactor into a .py metrics module

• Move code (cleaning, feature engineering, metric formulas) from the notebook into functions in metrics_utils.py. Add docstrings, type hints, and tests.

Step 3. Import the module and design the Streamlit UI

• In app.py, import metrics_utils, design and build a streamlit app that display key metrics that you identified. Choose proper features in the streamlit.

NoteWhy this workflow works (the rationale)?

• Mirrors the real development cycle: notebooks for development and experimentation, .py modules for stable logic ready for release, and Streamlit for production presentation.
• Reusability: Metric functions can be reused across future apps and analyses.
• Testability: Logic in .py is easy to test (if name == “main”:).
• Maintainability: Update formulas once in the module, not in multiple places.

6.2.3 Project Folder Structure

create a adsanalyses folder in your home directory with the files structured below.

# suggested folder structure under your home directory
adsanalyses/
├─ eda.ipynb        # your EDA
├─ metrics_utils.py # functions for load data, and calculating metrics
├─ app.py           # your streamlit app
└─ ad_table.csv

Thank you!