Scoping a B2B Ordering Portal Under a Fixed Date and Budget
Huaxin Foodservice Supplies (fictional) sells 3,200 SKUs to 420 active restaurant and café accounts. Today 62% of orders arrive by phone or email and are keyed in by eight inside-sales reps; the order-entry error rate is 5.8%, and credit-limit checks are done manually in spreadsheets. Average order value is ¥7,600. The sponsor has approved a four-month, ¥1.2 million project to build a B2B ordering portal and wants an 'Amazon-like' experience. Interviews reveal four pressures: a 30-location chain wants ERP punchout and delivery by 6 a.m.; small cafés want mobile reorder from past orders; reps fear losing commission if accounts order directly; and finance insists no order ships when a credit limit is exceeded. The warehouse cuts off orders at 4 p.m. IT has one integration developer and an ERP with a documented but slow SOAP API. The sponsor has already promised the chain a launch demo in month three.
1. Identify the key stakeholders, their interests and the three most consequential requirement conflicts. What process would you use to resolve them?
Read the worked answer
Use a power/interest grid rather than a flat stakeholder list.
| Stakeholder | Interest | Influence |
|---|---|---|
| Sponsor | Date, demo, visible transformation | High |
| Finance/CFO | Credit risk, receivables, audit | High |
| Sales director and reps | Commission, account control | High |
| 30-location chain | Punchout, 6 a.m. delivery | High (revenue) |
| Small cafés | Mobile reorder from history | Medium |
| Warehouse | 4 p.m. cut-off, pick accuracy | Medium |
| IT integration developer | Feasible ERP integration | Medium but scarce |
Three consequential conflicts
- Direct ordering versus rep compensation. If accounts order online, reps may be disintermediated and will quietly undermine adoption. Resolve with commission neutrality for 12 months, attribution of online orders to the owning rep, and an approval/assist role for exceptions. This is a business-process decision, not a software feature.
- Large-chain punchout versus the four-month date. Punchout (cXML/OCI) with the chain's ERP needs mapping, testing and security review - likely 8–12 weeks with one integration developer. Decouple it: deliver a hosted catalog and order API first, pilot punchout with one chain location, and phase the rest contractually.
- Frictionless ordering versus credit hold. Finance requires hard stops. Design credit check as a synchronous rule at order submission with a clear message and an approval workflow; never let 'pending credit' orders reach the warehouse.
Process
Run a facilitated workshop with named decision rights, keep a conflict log, and record signed-off decisions. Reframe 'Amazon-like' into outcomes: fewer keying errors, faster reorder, credit control. Maintain a traceability matrix from goal to requirement, and escalate only unresolved trade-offs to the sponsor with options, effort and risk.
2. Propose a MoSCoW scope for the four-month, ¥1.2M MVP. Justify what you put in Must and what you defer, using the operational numbers.
Read the worked answer
Must (MVP by month four)
- Account-specific contract pricing and catalog visibility.
- Search, reorder from history and saved lists.
- Cart and order submission with stock visibility (near-real-time, not guaranteed reservation).
- Synchronous credit-limit check and hard hold.
- Order status and history for buyers.
- ERP integration for orders, pricing and stock via scheduled batch plus an exception queue.
- Audit trail and role-based access.
Justification: 62% phone/email orders and a 5.8% error rate mean the core order-entry and pricing path delivers the ROI. Credit hold is non-negotiable for finance and protects receivables. Without ERP integration the portal creates a second source of truth and more errors, so it is Must even if batch-only at first.
Should
- Mobile-responsive PWA, approval workflow for over-limit orders, delivery-window selection, email confirmations.
Could
- Punchout for the top chain, personalized recommendations, live chat.
Won't this release
- Full marketplace, machine-learning recommendations, real-time carrier tracking, open self-registration.
Rationale: one integration developer makes ERP and punchout the critical path. Punchout is high value but affects one account; delaying it risks the chain, so mitigate with an assisted ordering desk and a month-five pilot. Deferring recommendations costs little because B2B buyers mostly reorder known SKUs. The mobile PWA is Should rather than Must because small cafés already order by phone and can use a responsive web form, but it should be built early enough to test with 20 accounts.
3. Write measurable acceptance criteria for the MVP and explain how you would handle the sponsor's late request to add ERP punchout for the 30-location chain without moving the date or budget.
Read the worked answer
Measurable acceptance criteria
- Order-entry error rate below 0.5% within 90 days of launch (baseline 5.8%).
- 70% of repeat orders placed self-service by month six.
- 95% of orders reach ERP within five minutes; failures appear in an exception queue with an owner.
- Zero orders released to the warehouse above credit limit, sampled daily with 100% accuracy.
- P95 page load under 2.5 seconds on 4G; availability 99.5% during 06:00–20:00.
- Order status reflects warehouse state within five minutes.
- Punchout: one chain location live by month five, 20 test orders accepted end-to-end before rollout.
Each criterion is testable and tied to a business outcome, not a screen. Add negative tests: credit-limit boundary, expired contract price, ERP timeout, duplicate submission.
Handling the late punchout request
Do not silently absorb scope. Perform an impact analysis covering effort, integration risk, test effort and the effect on the month-three demo. Present three options:
- Add punchout and descope mobile-responsive work plus recommendations (the sponsor accepts lower adoption for small cafés).
- Keep date and budget, phase punchout to month five, and mitigate the chain with an assisted ordering desk and a written commitment.
- Extend by six weeks with additional integration resource, which changes budget.
Take the options to the change control board; the sponsor decides with the chain contract in view. Record the decision in the decision log, update the traceability matrix, and pilot punchout with one location plus a rollback plan. The key is that the date and budget constraints are made explicit, not that the request is refused.