Business Case Studies & Business Publications

Supply Disruption at HorizonVue Optical ...

Alizamir, Saed;Sun...

Case

Supply Disruption at HorizonVue Optical Instruments (B)

Alizamir, Saed; Sun, Peng; We, Yehua

QA-0989 | Published August 14, 2026 | 3 Pages Case

Collection: Darden School of Business

Product Details

Early on August 10, 2019, Lin Pinpin, CEO of HorizonVue Optical Instruments (HorizonVue)—a large binocular manufacturer based in Nantong, China—learns that Typhoon Lekima has flooded the plant of the firm’s long-standing eyepiece supplier in Taizhou. The eyepieces are needed for the Artemis 360, a premium binocular; HorizonVue has committed to ship up to 16,000 pairs (32,000 eyepieces), assembled starting October 11, to a major US client. The purchasing team—director Wang Ziteng and analyst Candace Shi, a recent MBA hired to bring quantitative rigor to purchasing—has found a backup supplier in Kunshan that can deliver quickly but at a materially higher price. Lin will not cancel the Taizhou order, both to preserve a valued 15-year relationship and because Taizhou is far cheaper. The decision is therefore how many eyepieces to order from Kunshan, and the answer must be sent by noon the next day. Following the A case, which lays out the business problem and provides a deterministic profit calculator together with low–base–high ranges for the uncertain parameters, this B case supplies richer detail—elicited from the Taizhou owner and an engineer—on the handful of uncertainties that matter, so that students can build a full simulation model, treat the Kunshan order quantity as a decision variable, and produce a defensible recommendation under uncertainty. At the University of Virginia Darden School of Business, this fictional case set is used in the “Decision & Data Analytics” (DDA) elective, in both the executive MBA and MBA programs. It is designed for two 85-minute MBA classes in quantitative analysis or decision analysis, and it is equally suitable for an advanced undergraduate or specialized master’s course (an alternative, single-class teaching plan is available for students already familiar with the simulation and analysis). It exposes students to a realistic decision under uncertainty and works best within a simulation module. The teaching note is written assuming @Risk as the simulation software, but the teaching plan applies equally well with any other package (for example, Crystal Ball) or with a coding language such as R or Python.

By the end of the class session(s), students should be able to do the following: (1) Recognize when Monte Carlo simulation is the right tool for decision-making under uncertainty—namely, when a decision depends on many interacting uncertainties whose combined effect cannot be read off a single-point calculation or a few what-if scenarios. (2) Structure a decision model cleanly as a decision variable, uncertain inputs, a deterministic calculation engine (the cash-flow logic), and a tracked output. (3) Use tornado (single-factor sensitivity) analysis to rank uncertainties by the swing they cause in the key output(s) of interest, and focus modeling effort on the few that matter. (4) Choose and parameterize probability distributions—normal, triangular, uniform, and custom/discrete—from expert judgments and partially qualitative inputs, and model scenario-dependent uncertainties and their dependence. (5) Build and run a simulation model in @Risk, sweep a decision variable with a simulation table, and interpret the resulting risk profile: expected value, median, percentiles, and standard deviation (risk). (6) Practice and reinforce concepts from earlier courses about the trade-off between expected value and risk, and compare alternatives on the basis of their risk profiles. (7) Articulate the flaw of averages—why the expected value of a function is not the function of the expected values, and why ordering to the mean (or median) shortfall is wrong here. (8) Connect the problem to the newsvendor framework from core Operations classes (overage versus underage cost, the critical ratio) and appreciate where HorizonVue is richer than the textbook newsvendor model. (9) [Optional, for more advanced courses] Quantify the expected value of perfect information (EVPI), reason about risk aversion, and see how hedging the currency exposure removes a large, uncompensated risk. (10) Appreciate the managerial context: a family-owned Chinese small or medium-size business professionalizing its decision-making and moving upmarket, and the strategic value of supplier relationships.