cswork / libraryOpen practice workspace

Interview questions · Capital One

Decide whether to invest in R&D

This question evaluates a data scientist's competency in data-driven decision-making, financial modeling, product strategy, and cross-functional leadership by requiring synthesis of customer insight, NPV and cost analysis, cannibalization effects, operational risks, brand positioning, and a staged rollout plan.

Should we invest now in a Vegan burger? Build a decision narrative

Context

Assume you are the owner of a 100‑location fast‑casual burger chain evaluating whether to launch a Vegan burger in the next 12 months. You must make a go/no‑go recommendation supported by analysis and a rollout plan.

Task

Provide a concise, data‑driven decision narrative that covers:

  1. Customer insight
  2. Financials
    • NPV over 5 years at a stated discount rate
    • Fixed vs. variable costs
    • Cannibalization impact on profit
  3. Operational feasibility and risks
  4. Brand positioning and messaging
  5. Staged roadmap with timelines, KPIs, and exit criteria
    • R&D (formulation/sourcing/testing)
    • Pilot (A/B markets, pricing, measurement)
    • Scale or kill (national rollout vs. cancel)

State any assumptions you make and justify them briefly.

Solution

Executive recommendation

Conditional yes: pursue a staged test-and-learn. Fund R&D and a controlled pilot now; scale nationally only if pilot KPIs are met. In a realistic base case (12 units/store/day at 40% cannibalization), a 5‑year NPV is positive (~$0.34M) with material upside if velocity or attach rates exceed expectations. A limited‑time/market pilot de‑risks taste fit, operations, and cannibalization.

1) Customer insight

  • Problem/value prop: There is unmet demand from vegans/vegetarians, flexitarians, and mixed groups that choose restaurants with inclusive menus. The job-to-be-done is “a delicious burger experience without meat,” not “a compromise on taste.”
  • Signals to validate pre‑pilot:
    • Store‑level demand proxies: historical sales of sides/salads without meat, custom orders (no meat), add‑on veggies, and search/social interest in plant‑based.
    • Competitor scan: presence, pricing, and review sentiment of plant‑based burgers in similar concepts.
    • Willingness to pay: quick discrete‑choice conjoint (price points $7.49–$8.99) and sensory tests against beef and leading plant patties.
  • Hypothesis: Moderate new‑to‑brand traffic uplift and higher group conversion (party veto reduction). Expect slightly lower margin vs. beef but incremental party ticket lift via attach (fries/drinks).

2) Financials (NPV with fixed/variable costs and cannibalization)

All amounts in USD; assumptions stated and can be adjusted. Time horizon: 5 years. Discount rate r = 10%.

Assumptions (base case):

  • Footprint: 100 stores.
  • Price: $8.49 per Vegan burger.
  • Variable cost per Vegan unit: $5.49 (patty $3.50, bun/toppings $0.80, packaging $0.40, incremental labor $0.30, waste/shrink $0.49) → Vegan contribution margin CM_v = $3.00.
  • Beef contribution margin CM_b = $3.50 (for cannibalization comparison).
  • Steady‑state velocity after rollout: d = 12 Vegan units/store/day.
  • Cannibalization share c = 40% (share of Vegan orders that would otherwise have been beef burgers).
  • Ongoing fixed launch support F = $200k/year (menu, marketing, QA audits).
  • Upfront fixed costs (Year 0): $1.0M (R&D + supplier qualification + small‑wares/equipment + training content + pilot marketing).
  • Pilot Year 1 footprint: 25 stores for 6 months (≈12.5% of steady‑state year volume).

Formulas:

  • Annual units after full rollout: u = stores × d × 365.
  • Incremental annual gross profit (pre fixed): Profit_gross = u × [ (1 − c) × CM_v + c × (CM_v − CM_b) ]. Interpretation: new customers contribute CM_v; cannibalized customers contribute CM_v minus the margin lost on the cannibalized beef burger.
  • Incremental annual net profit after rollout: Profit_net = Profit_gross − F.
  • NPV: NPV = − CapEx_0 + Σ_{t=1..5} CF_t / (1 + r)^t.

Base‑case calculations:

  • u = 100 × 12 × 365 = 438,000 units/year.
  • Profit_gross = 438,000 × [0.6 × 3.00 + 0.4 × (3.00 − 3.50)] = 438,000 × [1.80 − 0.20] = 438,000 × 1.60 = $700,800.
  • Profit_net (Years 2–5) = $700,800 − $200,000 = $500,800 per year.
  • Year 1 pilot contribution ≈ 12.5% of steady‑state gross = 0.125 × $700,800 = $87,600; Year 1 fixed pilot support ≈ $200,000 → CF_1 = −$112,400.
  • Cash flows:
    • Year 0: −$1,000,000 (upfront fixed)
    • Year 1: −$112,400
    • Years 2–5: +$500,800 each year
  • Discounted cash flows (r = 10%):
    • PV(Year 1) = −$112,400 / 1.10 = −$102,182
    • PV(Year 2) = $500,800 / 1.21 = $414,711
    • PV(Year 3) = $500,800 / 1.331 = $376,069
    • PV(Year 4) = $500,800 / 1.4641 = $342,790
    • PV(Year 5) = $500,800 / 1.61051 = $310,811
  • NPV ≈ −1,000,000 − 102,182 + 414,711 + 376,069 + 342,790 + 310,811 = +$342,199.

Sensitivity (illustrative):

  • If velocity falls to 8 units/store/day (c = 40%): NPV ≈ −$0.36M (do not scale).
  • If cannibalization rises to 60% (d = 12): NPV ≈ −$0.49M (do not scale).
  • If velocity rises to 16 units/store/day (c = 40%): NPV ≈ +$1.13M (strong scale case).

Implications: Set go/no‑go guardrails on velocity and cannibalization; use pilot to estimate these with experiments.

3) Operational feasibility and risks

  • Sourcing: Qualify at least two patty suppliers; verify cost, texture, allergen statements, and shelf life. Contract for lead times and fill‑rate SLA ≥ 98%.
  • Kitchen/process: Cross‑contact controls (dedicated grill or liner, color‑coded utensils), cook time parity with beef, holding time. Validate that speed‑of‑service impact < 10 seconds/order at p95.
  • Inventory: Add SKU to DCs with cold‑chain capacity; forecast demand to limit waste (<1% shrink).
  • Training: Prep, cook, and cross‑contact scripts; allergen/vegan claims compliance.
  • Risks and mitigations:
    • Taste rejection → run blinded sensory tests; iterate formulation; allow mods (sauces).
    • Supply disruption → dual source; safety stock.
    • Ops complexity → simplify toppings parity with beef SKU; SOP checklists.
    • Brand backlash ("not real burger") → position on taste and choice; avoid moralizing.
    • Regulatory/claims → use "plant‑based" unless certifying strict vegan preparation is feasible without cross‑contact.

4) Brand positioning

  • Positioning: "A great‑tasting plant‑based burger—same craveable experience, now with plants." Emphasize taste first, inclusivity second.
  • Naming: Test "Plant Burger" vs. "Vegan Burger"; some flexitarians avoid "vegan" labels. Keep photography consistent with core brand.
  • Pricing: Target parity or modest premium vs. beef only if taste wins; otherwise price at or slightly below beef to drive trial. Test $7.49–$8.99.

5) Staged roadmap with timelines, KPIs, and exit criteria

Phase 0: Pre‑work (2–4 weeks)

  • Tasks: Data pull (POS, modifiers, dietary requests), competitor scan, supplier RFI, initial P&L model.
  • Output: Hypotheses, target costs, pilot design, guardrails.

Phase 1: R&D (8–12 weeks)

  • Activities: Supplier shortlisting, lab/kitchen trials, sensory tests (n≈200), shelf‑life, cookline trials, packaging, training assets.
  • KPIs:
    • Blind taste vs. beef: ≥80% rate "like" or better.
    • Target CM_v ≥ $2.75 at forecasted volumes.
    • Prep+cook time parity within ±15 sec of beef.
    • Supply QA pass; SLA commitments in contract.
  • Exit criteria: Advance to pilot if KPIs met; otherwise iterate twice maximum or stop if CM_v < $2.25 at any viable taste.

Phase 2: Pilot (8–12 weeks in 20–30 stores; stratified by region/demo)

  • Design: Randomized store‑level A/B with holdout markets; price cell tests; marketing on/off cells.
  • Measurement plan:
    • Primary metric: Incremental weekly gross profit/store (with cannibalization adjustment).
    • Methods: Diff‑in‑diff vs. control stores; SKU substitution analysis; A/B price elasticity; attach‑rate uplift modeling.
  • KPIs and thresholds:
    • Velocity: ≥10 units/store/day average in weeks 5–8.
    • Cannibalization: ≤40% of Vegan orders substituting for beef (measured via SKU substitution and customer‑level baskets where available).
    • Margin: CM_v ≥ $2.75; food cost variance ≤ ±0.5 pp.
    • Incremental profit/store/week ≥ $100 (net of fixed pilot support prorated).
    • Ops: p95 ticket time impact < +10 sec; out‑of‑stock rate < 2%; shrink < 1%.
    • Customer: Repeat purchase rate ≥ 25% within 30 days; product CSAT ≥ 4.2/5.
  • Exit criteria at pilot end:
    • Scale: Meets velocity AND cannibalization AND margin thresholds for ≥4 consecutive weeks; positive projected 5‑year NPV.
    • Iterate: Within 10% of thresholds; adjust price/recipe/marketing, extend 4 weeks max.
    • Kill: Velocity < 6 units/store/day OR cannibalization > 60% OR CM_v < $2.00 despite optimization.

Phase 3: Scale or kill (3–6 months)

  • Scale plan: Phased rollout waves (25 → 50 → 100 stores); secure supply; national media; update ops SOPs.
  • KPIs in rollout: Maintain pilot thresholds; monitor cohort retention; protect core beef sales; track brand health.
  • Kill plan: Liquidate inventory, remove menu item, publish learnings; revisit later via LTO or supplier co‑brand if market shifts.

Data and measurement guardrails

  • Instrumentation: POS flags for Vegan SKU; modifier capture; store‑day panel; staff feedback forms.
  • Cannibalization measurement: Estimate with substitution matrix and diff‑in‑diff vs. control markets; validate with customer panel or loyalty where available.
  • Sample size/power: Aim ≥20 stores per cell to detect ±2 units/store/day effect with 80% power (historical SD ≈ 3–4 units/day).
  • Bias controls: Avoid concurrent promos on beef in test stores; normalize for seasonality and day‑of‑week.

Decision logic summary

  • Proceed with R&D and a controlled pilot now.
  • Scale nationally only if pilot hits: velocity ≥10 units/store/day, cannibalization ≤40%, CM_v ≥ $2.75, and projected 5‑year NPV > 0 at r = 10%.
  • Otherwise iterate once; if still below thresholds, stop and consider a limited‑time, seasonal plant‑based offer instead of a permanent menu item.

What could change the decision

  • Positive: Supplier price break (CM_v +$0.50), co‑brand partnership, strong attach rate to sides/drinks, higher than expected group conversion.
  • Negative: Commodity spikes on plant proteins, ops complexity driving speed‑of‑service degradation, brand dilution among core beef‑forward guests.

Community answers

Nothing here yet.

Join in on CSWork →