Business Analytics Training, Session 2: Case Pack

Prompt (read aloud)

Step 0: the first question

Ask the room for the first question. Target: "What does each tier look like?" Size the tier first (units and money), then look at its structure (how many players, how they price), then ask what we would have to change to compete there, then check what spills over onto the rest of our business.

Clarifying information

Information to give when asked:

• East Africa: a large share of entry-tier hair dryers is bought by small hair salons for professional use; durability matters to them. Small shops and mobile-money payments dominate retail. Unbranded imports typically last about one year.
• Europe: the ultra-premium tier is dominated by one player with a proprietary motor and patents, sold through salons, department stores, and direct online; buyers are influenced by hair professionals.
• Factory: the APAC factory has spare capacity (this links to Case 2).

Board table A: East Africa, ultra low-end

Analysis A

1. How much of the market can we see today: premium 200k / 4M = 5%. Our current line is blind to 95% of the units.
2. Fair share: if we reached in the entry tier the same 20% we hold in premium, 2.8M x 20% = 560k units. Contribution per unit 8 - 6 = 2. Annual contribution 560k x 2 = 1.12M.
3. Break-even: year one needs (1M + 0.6M) / 2 = 800k units, 29% of the tier. From year two, 0.6M / 2 = 300k units, 11% of the tier. Kenya alone cannot carry this; the five countries together might.
4. Money view: the whole entry tier at wholesale is 2.8M x 8 = 22.4M. Our fair-share revenue would be 560k x 8 = 4.5M. Large units, small money.
5. Price normalization (the Case #8 step): an import at retail 8 lasting one year costs 8 per year; our model at retail 12 lasting three years costs 4 per year. We are cheaper per year of use, but the buyer pays 12 versus 8 at the counter. The barrier is the one-time outlay, not the product's value. The fix is in the channel: small installments through mobile money, shop credit, and small-batch B2B sales to salons.

Structure reading A (p.58)

The entry tier is close to perfect competition; price sits on marginal cost (imports 7 on 5, us 8 on 6). We have no price advantage. The only thing that earns the extra 1 at wholesale is the durability insight (p.58 key driver: consumer insight).

Board table B: Europe, ultra-premium

Analysis B

1. Size: 600k x 245 = 147M wholesale revenue in the tier. One fifth of the East African entry tier in units, more than six times the money.
2. Fair share: at the 15% we hold in premium, 600k x 15% = 90k units. 90k x 165 = 14.85M, about 15M per year of contribution.
3. Break-even: year one 15 - 8 - 12 = -5M. From year two 15 - 8 = +7M per year. Year-two break-even is 8M / 165, about 48k units, 8% of the tier.
4. Trading up: if 10% of our premium buyers (36k units) move up, each unit's contribution goes from 60 to 165, a gain of 36k x 105 = 3.8M. Cannibalization at this end has a positive sign, the opposite of the East African case.

Structure reading B (p.58)

The ultra-premium tier is close to a monopoly; retail 350 against a cost of 80. The incumbent can cut price 20% and still earn a large margin, so a price war is painless for it and fatal for us. The barriers are motor technology, patents, and brand, not distribution.

Side by side

Recommendation (reference answer, leave open for debate)

• East Africa: do not launch for one country. Launch for the five countries together, using the factory's spare capacity, under a sub-brand or without the main logo, and validate first through the salon channel where durability is valued.
• Europe: launch, treating the year-one loss of 5M as marketing spend, on the condition that the product carries a real technical differentiator. Without one, we would be entering inside the incumbent's pricing range.

Risks

• East Africa: price war with imports, brand dilution, distributor payment collection.
• Europe: incumbent price cut, patent exposure, retailer shelf access.

Next steps

• East Africa: pilot with salon distributors in two countries; test installment pricing.
• Europe: confirm the technical differentiator and patent position; line up salon and department-store partners before launch.

Pitfalls to watch

• Sizing in units only or in money only; the two ends give opposite pictures.
• Applying our European share to Africa without checking the tier's structure first.
• Comparing retail prices without normalizing to price per year of use.
• Treating the 12M R&D as a per-unit cost.
• Forgetting that cannibalization has opposite signs at the two ends.

Casebook page mapping

• p.58: industry structure at both extremes (perfect competition vs monopoly); barriers (distribution and payment vs technology and patents); key drivers (consumer insight on durability; brand equity); trends (rising incomes move entry buyers up the ladder).
• p.59: price times volume in each tier; fixed vs variable (tooling and R&D vs unit cost); price normalized per year of use.
• p.60: new market (East Africa) vs product development (Europe) in the options matrix; source of volume (take share from imports vs from the incumbent); trading down vs trading up; cannibalization.
• p.61: STP and all four Ps change at both ends; the same brand cannot be positioned identically at both.

Closing line

The same hair dryer at the bottom of the ladder fights perfect competition with cost and distribution; at the top it fights a monopoly with technology and brand. We live in the middle, and both attacks come from there.

Prompt (read aloud)

Clarifying information

The source provides no separate clarifying script for this case; all case information is in the board table below. Scenario inputs used in Steps 4 and 5 (10% of flagship buyers switching; the factory already full; volumes at 3M and 5M) are introduced by the interviewer at those steps.

Framework

Two lenses on the same deal. Lens A: we are a product design company, and the factory is a supplier. Lens B: we are a design plus manufacturing company, and the factory's costs are our own fixed and variable lines. The p.58 question "is the industry vertically integrated" decides which line in p.59 is fixed and which is variable.

Data table

Analysis

Step 1, Lens A: we are a product design company

The factory is a supplier; 85 is a purchase price.

• Today: 140 - 85 - 40 = 15 per unit. 4M x 15 = 60M profit.
• Budget version at the same transfer-price formula: 45 variable + 35 fixed allocation = 80. Wholesale 80 minus 80 = 0 per unit, then minus 10M marketing. Loss of 10M. Decision: no.

Step 2, Lens B: we are a design plus manufacturing company

The 85 disappears. Replace it with 50 variable plus 140M fixed.

• Check today: 4M x 140 = 560M revenue; minus 4M x 50 = 200M variable; minus 140M factory fixed; minus 160M HQ fixed. Profit 60M. Same as Lens A.
• Budget version: capacity is enough (4M + 2M = 6M), so factory fixed cost does not change. Incremental contribution 2M x (80 - 45) = 70M. Minus 10M marketing. Net +60M. Profit doubles to 120M. Decision: yes.
• Side observation: at 6M units the factory fixed cost per unit falls from 140 / 4 = 35 to 140 / 6 = about 23. The flagship's "cost" moves from 85 to about 73 without anything changing on the shop floor. Cost allocation is an accounting choice (p.59, cost allocation across multiple products, capacity utilization).

Step 3: why it flips

Lens A treated the factory's fixed cost as if it were variable. The p.58 question "is the industry vertically integrated" is not background; it decides which line in p.59 is fixed and which is variable.

Step 4: what Lens B misses

• Cannibalization (p.60 trading down): if 10% of flagship buyers (400k) switch to the budget version, we lose 400k x (140 - 50) = 36M and gain 400k x (80 - 45) = 14M, net -22M. The deal is still +38M.
• Capacity: if the factory were already full, the 2M budget units would displace flagship units at 90 contribution each. Ask the room what happens then.
• Brand: a budget version on a big retailer's shelf changes how the flagship is perceived (p.61 positioning).

Step 5: the reverse, who hurts more when demand moves

Owning the factory amplifies both directions: more profit in a boom, more loss in a downturn. This is the choice NVIDIA made (fabless, the foundry carries the factory fixed cost) and the choice SharkNinja made differently (owned factories).

Recommendation

As the source gives it: under Lens A (product design company, factory as supplier) the decision is no, a loss of 10M. Under Lens B (design plus manufacturing company) the decision is yes, net +60M, profit doubles to 120M. The correct lens follows from what the company actually is.

Risks

What Lens B misses: cannibalization (net -22M in the 10% switching scenario, deal still +38M); capacity (if the factory were already full, the 2M budget units would displace flagship units at 90 contribution each); brand (a budget version on a big retailer's shelf changes how the flagship is perceived).

Extension

Step 5 above: the reverse case, who hurts more when demand moves. Owning the factory amplifies both directions.

Pitfalls to watch

• Treating the transfer price as a true cost.
• Forgetting the 10M marketing line.
• Skipping the capacity check.
• Allocating fixed cost per unit and then calling the result "the cost".

Casebook page mapping

• p.58: value chain and vertical integration; economies of scale as both a barrier to entry and a burden.
• p.59: fixed vs variable, capacity utilization, cost allocation across products.
• p.60: increased penetration via a new price tier; trading down; cannibalization.
• p.61: price and positioning of the budget version; who it is targeted at.

Closing line

The same deal is a loss or a doubling of profit depending on what we say the company is.

Prompt (read aloud)

Clarifying information

Narrowed prompt (given once framing B survives Part 1): we take 1 million after-sales calls per year, 6 minutes each, human agents cost 1 per minute, so 6M per year. Should we build a voice bot on the 0.05 per minute model, and how?

Framework: Part 1, three framings, two eliminated in one line each

Teaching point: most of industry analysis is deciding what not to analyze. The common mistake is to compute all three.

Data table

Analysis: Part 2, the customer-service framing, in full

Calculation 1: break-even containment rate

Let c be the share of calls the bot fully resolves.

• Cost per call with the bot = 6 x 0.05 + (1 - c) x 6 = 6.3 - 6c.
• Cheaper than 6 when 6c > 0.3, i.e. c > 5%. The bot pays for itself at 5% containment.

Calculation 2: raw model vs platform, with the Tech team's lines added

• Raw model per call: 0.3 API + 0.5 maintenance (500k over 1M calls) = 0.8; year one adds 0.5 integration, so 1.3.
• Platform per call: 6 x 0.15 = 0.9, no integration, no maintenance.
• Year one: platform cheaper by 0.4 per call, 400k. From year two: raw model cheaper by 0.1 per call, 100k per year.
• Ask: is 100k per year worth running our own stack and being tied to one vendor?
• If the vendor raises the price to 0.08: raw model per call 0.48 + 0.5 = 0.98, and the platform at 0.9 is cheaper for good. The headline 0.05 stops being an advantage the moment maintenance is on the board.

Calculation 3: the benefit side is not only labor

• If the bot turns 1% of calls from "return the product" into "problem solved", at 100 per avoided return: 10k x 100 = 1M per year. Ten times the 100k difference in Calculation 2.
• If containment is only 30% and the experience is poor, returns rise 0.5%: 5k x 100 = 500k lost, most of the labor saving gone.
• So the decision hinges on containment rate and troubleshooting success rate, and only the CX team can supply those numbers.

Board table: each team's line

PD and Social Media have no line in this framing.

Recommendation

As the source gives it: only framing B (customer-service voice bot) is worth computing. The bot pays for itself at 5% containment. Between the raw model and the platform, the platform is cheaper by 400k in year one and the raw model cheaper by 100k per year from year two, a difference that disappears if the vendor raises the price to 0.08. The decision hinges on containment rate and troubleshooting success rate, and only the CX team can supply those numbers; the benefit side (1M per year from avoided returns at 1% of calls) is ten times the 100k API-fee difference.

Risks

As surfaced in the source: vendor lock-in and a vendor price rise (at 0.08 the raw-model route loses to the platform for good); Legal and data may veto the raw-model option outright (customer audio going to the vendor; APAC and UK data rules); poor experience at low containment (30%) raises returns by 0.5%, 500k lost, most of the labor saving gone.

Extension questions for a technical audience

• The vendor raises the price to 0.08. Which of the three framings changes? (B survives, the raw-model route loses to the platform.)
• Legal rules that no customer audio may leave the region. What survives? (Platform with regional hosting, or an on-premise model with fixed cost.)
• Peak season doubles call volume for two months. What happens to each option? (Variable options scale, human hiring does not.)

Pitfalls to watch

• Computing all three framings.
• Comparing 0.05 to zero instead of to the alternatives.
• Forgetting maintenance and escalation cost.
• Counting benefits as labor savings only.

Casebook page mapping

• p.58: substitutes, industry structure and buyer power per framing, value chain bottleneck, trends.
• p.59: variable cost replacing variable cost; cost allocation per call; fixed integration vs variable API.
• p.60: diversification quadrant (framing A); switching costs (raw model route).
• p.61: measuring outcomes, the Keep step (retention via service quality).

Closing line

Everyone started by asking whether 0.05 was cheap; three steps later the price is not in the decision. What decided it was which industry we were buying from, where the bottleneck was (troubleshooting success), and whose line was largest (maintenance and returns, not API fees).