Lesson 8 of 12 · Module 8 · Share something new
Publish responsible original research
What it is
Choose a concrete question your data can answer. One store's customers do not represent the whole market automatically. 1. Define the question/audience, such as which capacity range appears in this store's quarterly orders. 2. Inventory sources, permissions and privacy requirements. Do not put customer-level exports in public tools. 3. Define dates, eligible records, units, duplicates and missing data. Orders, products and customers are different denominators. 4. Percentage = relevant count / eligible total × 100. Zero eligible records means unavailable, not a fabricated percentage. 5. Scope wording carefully: in an illustrative fictional sample, 504 of 1,200 eligible orders contained travel-size items (42%). This is not 42% of all shoppers and is not real client data. 6. Publish dates, counts, methods, accessible tables and limitations. Observational correlations are not causal proof. 7. Removing names does not guarantee anonymity: small groups, unusual purchases and dates may identify people. Minimize, restrict and aggregate data; seek privacy review. 8. Pitch useful findings without requiring coverage or links; retain methodology and set an update schedule. Brief: Question | Audience | Source/permission | Sample | Dates | Calculation | Finding | Limitations | Review owner.
Why it matters
Helpful resources • Free: private Google Sheets for aggregate calculations/charts. • Internal: Backlinks and Outreach to review coverage and follow-ups. Sign-in/selected client required; live checks may cost confirmed credits. Outreach uses your own mail app, not platform sending. • Source: Shopify privacy settings.
Free exercise workbooks
This lesson: Research Brief.
PDF: print and write by hand. Spreadsheet: edit in Excel or import into Google Sheets. Keep completed copies private; answers are not uploaded.
✓ Good example
A fictional report clearly states store-only dates, eligible count and exclusions, not industry-wide preferences.
✗ Common mistake
Publishing individual orders with names removed, inventing research or claiming correlation proves causation.
You pass when: You have a question, permission/source inventory, reproducible calculation or honest no-data decision, methods and privacy-reviewed limitations.