Module 4 · Lesson 4 of 4

Calculate first, then write the decision brief

Reconcile a small dataset and separate observed results from explanations to investigate.

Go to the exercise

What you will leave with

A checked calculation table and a concise, source-based decision brief.

Understand

The idea behind the work

Before interpreting data, identify what one row represents, the reporting period, currency, filters and metric definitions. Reconcile retained rows and totals with the source. A duplicate campaign name may represent separate valid rows; do not delete rows merely because labels repeat. Explain any exclusions or transformations.

Compute aggregate rates from totals, use unavailable for a zero denominator and preserve the calculation. Ask AI to explain only the checked evidence. A difference is an observation; an explanation for it is a hypothesis unless further evidence supports causality. Without a comparable prior period, do not claim an increase or decline.

Inspect a worked example

Synthetic example: inspect campaign costs

This is a fictional teaching example, not a client result or a claim about Joseph Carroll's past work.

The values are invented for learning, not a live account export.

The starting input

A: spend 100, clicks 100, conversions 10. B: spend 900, clicks 300, conversions 30.

Work through it

  1. Reconcile spend 1,000, clicks 400 and conversions 40.
  2. Calculate CPC 2.5, click-based conversion rate 10% and cost per conversion 25.
  3. Write that B accounts for most spend; ask for goals and comparable periods before judging effectiveness.

The result

Exact calculations and a bounded next question, with no claim of trend or causality.

A mistake to recognize

Saying performance improved despite having only one period.

How to recover

Request a compatible comparison and keep the current summary descriptive.

Adapt the method

What changes in your business?

The same method can serve different settings. Data, allowed actions and operational support matter more than headcount alone.

Owner-led business or small team

Use the labeled sample before loading an authorized export you understand.

Department or growing team

Agree on the metric dictionary and review every material number against the table.

Large or complex organization

Use governed source definitions and verify ordinary-user access; a dashboard view alone does not establish authorization.

Try it, then inspect the result

Write a checked performance note

Use the synthetic values or the KPI sample. Show totals and one weighted rate, write one observation, one hypothesis or question and the evidence needed next.

Use the CSV KPI workbench

Use a sample or a description you are allowed to share. Your answer stays in this browser when storage is available; export it to keep a separate copy. 0 / 6,000 characters.

Check your reasoning

A dataset has zero conversions and positive spend. What is the cost per recorded conversion?

Inspect your own output

Check each statement after you look at your answer. These evidence checks apply to this visit; your answer and completion status can be saved.

To mark completion, write an answer, check the reasoning scenario and inspect every evidence statement. You can continue to another lesson at any time.

Self-reported completion does not certify competence. Open My AI Plan to review your saved work.

Sources and further practice

The lesson is Carroll Consulting's teaching synthesis. These sources support the underlying capabilities and concepts; check the current product, plan and configuration before applying them.

You can move between lessons freely. Revisit the checks whenever your answer, evidence or business context changes.