The capstone combines channel choice, product evidence, catalog controls, margin, service and inventory into a 90-day operating experiment. It is not a revenue forecast. Your goal is a store system that produces trustworthy evidence and protects customers and cash.
Plan Three 30-Day Phases
Days 1–30 — validate: choose one channel and one to three SKUs. Finish truth sheets, samples, unit economics, source-backed listings and a capped demand test. Define privacy, returns, service and stock controls before accepting volume.
Days 31–60 — stabilize: reconcile every order lifecycle, investigate returns and questions, repair listing or packaging defects, and establish the daily control windows. Do not add products while core records disagree.
Days 61–90 — improve one bottleneck: choose the strongest evidence-backed constraint—qualified traffic, conversion clarity, fulfillment, margin, repeat purchase or service. Run one controlled change with a baseline and stop rule.
Build the Evidence Dashboard
Use counts that connect:
qualified visits or marketplace exposure (where available)
→ product-page actions
→ placed orders
→ accepted/shipped orders
→ delivered orders
→ returns/refunds
→ settled contribution
Also track stock accuracy, response/escalation age and top buyer questions. Never combine unlike windows or claim causation from one before/after chart.
Worked Example
A hypothetical Peshawar kitchen-goods seller starts with three storage products on Daraz. Day 30 evidence shows one product receives visits but repeated size questions. Instead of increasing discounts, the seller adds measured dimensions and a scale photograph. Day 60 shows fewer size questions and better fulfilled-order conversion, but contribution is thin. Day 61–90 tests a bundle with a fixed unit cap.
The final report includes source sheets, rejected hypotheses, costs, mistakes and next decision. It does not label the sample as a guaranteed playbook for other stores.
Failure Cases to Diagnose
- Launching too many SKUs to identify what caused problems.
- Changing price, image, ad and description at the same time.
- Reporting gross orders without delivery, return and settlement.
- Scaling before stock and service reconciliation works.
- Presenting sample work or projections as real customer results.
🇵🇰 Pakistan Angle
Write operational contingencies for courier disruption, load-shedding, mobile-data failure and seasonal peaks. Keep PKR cash timing visible. If you handle taxes, regulated products or cross-border trade, get qualified advice rather than asking AI to approve compliance.
Hands-On Exercise
Create a capstone folder containing:
- channel scorecard and research workbook;
- supplier/sample evidence and product truth sheets;
- approved listings and asset map;
- margin model and promotion gate;
- service desk and inventory rules;
- 90-day experiment board and weekly decision log;
- final evidence report with limitations.
Completion Rubric
- Scope is capped to a testable channel and SKU set.
- Customer, margin, stock and privacy controls exist before growth.
- Measures follow the full order lifecycle.
- Each experiment changes one major variable.
- The final report distinguishes fact, inference and next test.
- No income, demand or platform outcome is guaranteed.
Sources
- Shopify Help Center — Reports and analytics
- Daraz Pakistan — Start Selling on Daraz
- Google Sheets Help
- Federal Board of Revenue Pakistan
Key takeaway: A finished ecommerce system is a 90-day evidence loop that earns the right to scale by protecting truth, customers and cash at every stage.