Sources and References

Primary textbook — Knaflic, Cole Nussbaumer. Storytelling with Data: Let’s Practice! Wiley, 2019. (Chapters 2 and 4 excerpts) Supplementary — Cotgreave, Andy. Various blog posts on Tableau best practices, The Tableau Blog. Esselman, Amy. “Storytelling with Data” blog series. Online resources — EY ARC, “Introduction to Data Visualization” (extract); Tableau Public documentation; Gartner Analytics Maturity Model; Kaplan & Norton, The Balanced Scorecard (Free Press, 1996) for KPI frameworks.

Chapter 1: Foundations of Performance Analytics

What is Performance Analytics?

Performance analytics is the discipline of examining business data systematically to answer three fundamental questions: what happened, why it happened, and what should be done about it. It sits at the intersection of business knowledge, data analysis, and communication — requiring a practitioner to understand the organization, work with available data, and convey findings to decision-makers in a form they can act on.

This differs from general data analysis in an important way: performance analytics is explicitly goal-oriented and audience-aware. Every choice — what data to pull, which chart to build, how to sequence the story — is made with the stakeholder’s decision in mind.

The Analytical Mindset

Adopting an analytical mindset means approaching a business problem with curiosity, skepticism, and structure. It involves:

1. Asking good questions before touching data. What does the stakeholder actually need to know? What decision will this analysis support?
2. Understanding the business model first. A retail analyst who does not know how gross margin is calculated cannot meaningfully diagnose margin compression.
3. Letting data guide, not confirm. Exploratory analysis should be open-ended; the analyst should be willing to be surprised by findings rather than cherry-picking evidence for a predetermined conclusion.
4. Communicating with precision. Analytical conclusions must be expressed clearly, without jargon, at the level of detail appropriate to the audience.

The Performance Diagnostic Framework

The course structures a performance diagnostic around three sequential questions:

A good diagnostic starts at the highest level — overall performance against a target or prior period — and then progressively decomposes that result by segment, product, geography, or time period to identify the root cause of any gap.

Professional Ethics in Analytics

Because analytics shapes decisions, the analyst carries ethical obligations. Misleading visualizations — truncated axes, cherry-picked time windows, inappropriate chart types — distort perception even when the underlying numbers are accurate. Ethical practice requires:

• Presenting data in context (baselines, benchmarks, confidence intervals where relevant)
• Disclosing data limitations and caveats
• Distinguishing correlation from causation in all communications
• Maintaining confidentiality of sensitive business information

The CPA Canada Code of Professional Conduct and the broader literature on data ethics both emphasize the professional’s duty to report honestly and to protect the integrity of the information they handle.

Chapter 2: Business Models, KPIs, and Data

Understanding the Business Before the Data

Before loading a single row of data, an effective performance analyst must understand how the business makes money. A business model describes the value a company delivers to customers, the mechanisms by which revenue is generated, and the cost structure that supports those activities. Key questions include:

• What are the primary revenue streams?
• What are the main cost drivers?
• Who are the target customer segments?
• What operational metrics (KPIs) best reflect health in each segment?

Key Performance Indicators

KPIs can be organized into a hierarchy corresponding to the levels of the organization. Robert Kaplan and David Norton’s Balanced Scorecard framework is a useful organizing lens:

• Financial perspective: Revenue growth rate, gross margin %, return on assets (ROA), earnings before interest and tax (EBIT)
• Customer perspective: Customer satisfaction scores, Net Promoter Score (NPS), customer retention rate, average order value
• Internal process perspective: Order fulfillment cycle time, defect rate, inventory turnover
• Learning and growth perspective: Employee engagement, training hours per employee, technology capability index

Retail Business KPIs

In a retail context, the core revenue drivers are:

and further:

Gross margin in retail is:

Additional retail KPIs include same-store sales growth (which isolates performance from store expansion effects), sell-through rate (units sold ÷ units received), and inventory days on hand.

Service Business KPIs

Service businesses generate revenue differently — typically through billable time, recurring subscriptions, or project fees. Key metrics include:

• Utilization rate: Billable hours ÷ available hours, the primary driver of professional services revenue
• Average revenue per user (ARPU): Common in subscription and SaaS businesses
• Churn rate: The percentage of customers who cancel in a given period
• Customer lifetime value (CLV): The present value of all future revenue expected from a customer

Data for Analytics

Real business data rarely arrives in a clean, analysis-ready form. Common preparatory steps include:

1. Connecting to data sources: spreadsheets, databases, CRM exports, ERP reports
2. Data profiling: understanding field types, unique values, nulls, and ranges
3. Cleaning: removing duplicates, handling missing values, correcting data types
4. Reshaping: pivoting, joining multiple tables, creating calculated fields
5. Validating: cross-checking totals against known benchmarks

Chapter 3: Exploratory Data Visualization

Principles of Effective Data Visualization

Data visualization serves two distinct purposes: exploration (helping the analyst understand the data) and explanation (helping the audience understand a conclusion). The chart type, color palette, level of detail, and annotation choices should all reflect which purpose is being served.

Common Chart Types and When to Use Them

Tables

Tables are appropriate when the audience needs to look up specific values rather than see an overall trend. They work well for comparing exact numbers across a small set of categories. However, they make it difficult to perceive patterns, trends, or outliers.

Bar Charts

Bar charts are the workhorse of business analytics. They excel at comparing values across discrete categories.

• Vertical (column) bar charts: Best for showing values over time (months, quarters, years).
• Horizontal bar charts: Best for comparing categories where labels are long, or when there are many categories.
• Stacked bar charts: Show the composition of a total, but make it harder to compare individual segments.

A key design rule: always start the axis at zero for bar charts. Truncating the axis dramatically exaggerates differences between bars and misleads the viewer.

Line Charts

Line charts convey continuous change over time. They imply that the data points are connected — that movement between them is meaningful. Do not use a line chart to compare discrete categories that have no natural sequence. Multi-line charts can compare multiple series over time, but readability degrades beyond three or four lines; consider small multiples instead.

Scatter Plots

Scatter plots display the relationship between two continuous variables, with each data point representing an individual observation (e.g., a store, a customer, a product). They are the primary tool for identifying correlations, clusters, and outliers.

A trend line (linear regression line) can be added to make the direction and strength of a relationship visually apparent.

Maps

Geographic visualizations (choropleth maps, filled maps) are powerful when location is a meaningful driver of the variable being displayed. Use maps when the spatial pattern is the insight — for instance, when sales performance varies systematically by region, province, or city.

Tableau Mechanics

Tableau Desktop organizes data into dimensions (qualitative fields like Region, Product Category, Customer Segment) and measures (quantitative fields like Sales, Profit, Quantity). The core workflow is:

1. Connect to a data source (.xlsx, .csv, database connection)
2. Drag dimensions and measures to the Rows, Columns, and Marks shelves
3. Choose the chart type from the “Show Me” panel or by manually configuring the marks
4. Filter to focus on a subset of the data
5. Format for clarity: adjust axis labels, add reference lines, apply appropriate color scales

Calculated Fields

Tableau allows users to define new measures using formulas. Common examples:

• [Profit Margin] = SUM([Profit]) / SUM([Sales])
• [Year-over-Year Growth] = (SUM([Sales]) - LOOKUP(SUM([Sales]), -1)) / ABS(LOOKUP(SUM([Sales]), -1))

Hierarchies, Filters, and Sorting

Hierarchies allow the user to drill down from a high-level summary to a more granular view (e.g., Year > Quarter > Month, or Category > Sub-Category > Product). Filters let users restrict the data displayed in a view without modifying the underlying data source. Sorting controls the order in which bars or rows appear, which significantly affects the narrative the chart tells — descending sort by value makes it easy to identify the top and bottom performers.

Chapter 4: Communicating with Data — Storyboards and Tableau Stories

From Analysis to Communication

Completing the analytical work is only half the job. Translating findings into a coherent narrative that a decision-maker can act on is equally important and equally demanding. This requires thinking like a communicator, not only like an analyst.

A classic structure for presenting a performance diagnostic is:

1. Situation: Set context — what business are we analyzing, and over what period?
2. Complication: State the performance gap or issue — what happened that merits attention?
3. Resolution: Provide the root cause and recommendation — why did it happen, and now what?

This mirrors the “SCR” (Situation-Complication-Resolution) structure used widely in management consulting.

Storyboarding

A storyboard is a planned sequence of slides or dashboard screens that maps out the narrative before any polished visuals are built. It forces the presenter to decide:

• What is the key message of each screen?
• How does each screen connect to the next?
• Does the sequence answer all three diagnostic questions?

Good storyboards are often sketched on paper or sticky notes before Tableau is opened. The order of the screens should mirror the logic of the argument: from the summary finding to the supporting evidence to the root cause to the implication.

Tableau Stories

In Tableau, a Story is a sequence of Story Points, each of which contains a dashboard or a worksheet with an optional caption. Stories are the primary mechanism for turning individual charts into a connected presentation.

Design guidelines for Tableau Stories:

• Each story point should advance the narrative by one logical step
• Use captions to state the “so what” of each screen — not just a description of what is shown
• Maintain visual consistency: same color palette, same fonts, same axis ranges across related views
• Avoid cluttering a single screen with too many charts; one key insight per screen is usually optimal

The Summary Slide

After a full story, a single summary slide condenses the entire diagnostic into the three answers:

The summary slide tests whether the analysis produced actionable insight. If the “now what” cannot be stated clearly, the diagnostic is incomplete.

Chapter 5: Applying the Diagnostic — Case Studies

Retail Business Diagnostic

A retail performance diagnostic typically begins with a top-level view of total revenue versus a benchmark (prior year, budget, or industry average). If a gap exists, the analyst decomposes revenue by:

• Geography: Are certain regions underperforming?
• Product category: Are specific categories driving the gap?
• Time: Is the gap concentrated in a particular month or season?
• Customer segment: Are the declines in high-value or low-value customers?

Each decomposition narrows the hypothesis space. The analyst should form and test hypotheses iteratively, rather than computing every possible metric and hoping the answer appears.

Service Business Diagnostic

In a service business, the diagnostic must account for the distinct revenue model. Revenue is typically:

The “now what” might involve identifying high-churn segments and recommending retention programs, or identifying high-utilization practice areas where capacity expansion would yield disproportionate revenue growth.

Preparing service business data often involves more complex data transformation than retail, because service billing records, CRM entries, and time-tracking logs may come from separate systems and require joining on client identifiers.

Chapter 6: Synthesis — The Analytics Mindset in Practice

Planning a Diagnostic

Before beginning any analysis, the professional analyst prepares a diagnostic plan that specifies:

1. Business context: The company, its model, and the competitive environment
2. Stakeholder need: What decision is the analysis supporting, and who is the audience?
3. Available data: What data sources exist, what fields are relevant, what are the known limitations?
4. Proposed analyses: Which metrics, chart types, and breakdowns will be used to answer each question?
5. Timeline: When will the analysis be complete and presented?

This planning discipline prevents the common failure mode of “analysis without direction” — spending hours building charts that do not lead to a coherent conclusion.

Iterative Refinement

Analytics is rarely linear. A practitioner will often:

• Build an initial view, notice something unexpected, and pivot to investigate
• Return to the raw data to verify a surprising result
• Revise the narrative as new evidence emerges

This iterative process is normal and healthy. The key discipline is to maintain a clear view of the ultimate purpose — answering the three diagnostic questions for a specific stakeholder — so that exploration does not become an end in itself.

Connecting Analytics to Business Value

The ultimate test of a performance diagnostic is whether it changes a decision. Analytical work that is technically rigorous but fails to reach a decision-maker in an actionable form has not fulfilled its purpose. The communication skills covered in this course — storyboarding, Tableau stories, summary slides — exist precisely to bridge that gap between insight and action.

An effective analyst understands that the chart is not the product. The recommendation is the product. The chart is the evidence that supports the recommendation.