Now point the value chain outward. A profit pool (Gadiesh & Gilbert, "Profit Pools: A Fresh Look at Strategy," HBR, May–June 1998) is the total operating profit earned at every link of an industry's chain, mapped. Each link gets its share of chain revenue on the x-axis and its operating margin on the y-axis, so the area of each block is that link's share of the chain's profit. The map is powerful because profit share and revenue share routinely diverge, and wildly. That divergence is the strategy of the whole chain, made visible.
How do you map a profit pool?
Work through five steps, in this order:
- Define the chain end to end, from raw supplier to end customer, including service, financing and aftermarket links. Usually the pool hides in the links everyone forgets.
- Size total chain revenue per year or per unit sold. Per unit is often cleaner: per car, per phone, per paint job.
- Estimate each link's revenue and operating margin. The workhorse trick is to use listed pure-plays as margin proxies for each link. Foxconn stands in for assembly, TSMC for leading-edge fabrication, a listed dealer group for dealerships. Their filings are public, and their margins are the chain's skeleton.
- Compute each link's profit and share of the pool. Then cross-check that the blocks sum to something sane against the end price and known industry profit estimates.
- Explain and extrapolate. Run a compressed five-forces per link, since the pool collects wherever the forces are weakest. Then ask which direction the pool is migrating, and what technology, regulation or integration move is driving it.
Advertisement
Worked example: the smartphone profit pool
The smartphone, worked as a chain. All figures are illustrative, per one $1,199 flagship and industry-wide. Verify them against Counterpoint Research profit-share notes and TSMC/Hon Hai/Qualcomm/Apple filings.
| Link | Revenue per device (approx.) | Op margin (proxy) | Op profit per device | Why the forces put it there |
|---|---|---|---|---|
| Chip fabrication (TSMC) | ~$70 (SoC + others' wafer share) | ~45% | ~$30 | Monopoly at the leading node: no substitute supplier at required performance, so supplier power is maxed |
| Memory (Samsung/SK/Micron) | ~$60 | −20% to +40% (cyclical) | ~$0–20 | Commodity oligopoly with lumpy capacity: capital cycle, not power |
| Display, camera, RF, other components | ~$180 | ~5–20% | ~$20 | Mixed: differentiated parts (camera sensors) earn; commoditized ones don't |
| Chip design & licensing (Qualcomm, ARM) | ~$45 | ~30%+ (licensing far higher) | ~$15 | IP = policy-grade barrier (patents); per-device royalties on others' capital |
| Assembly (Foxconn/Hon Hai) | ~$30 value-add | ~2.5–3% | ~$1 | Fragmented, substitutable, facing one concentrated buyer: the powerless middle |
| Brand/design/OS (Apple) | balance of ~$1,199 wholesale/retail mix | hardware gross ~35–40%; company op ~30%+ | ~$250–300 | Differentiation + ecosystem switching costs + demand-side scale: every force benign for this one firm |
| Distribution (carrier/retail) | margin on device ~2–5% | thin | ~$10–20 | Commodity conduit; earns on the service plan instead |
| Aftermarket (App Store, services) | $ per device-year, recurring | very high | the second pool | 15–30% take rate on a captive installed base: a marketplace bolted onto a device |
Industry-wide, the picture is the same. Global handset industry operating profit has run on the order of $110 to 130bn a year, and Apple has captured roughly 75 to 85% of it on under 20% of unit shipments (Counterpoint estimates; check the current year). Samsung takes most of the remainder, and the median Android OEM oscillates around breakeven. Meanwhile the assembler of most premium devices earns ~2.5% margins on enormous revenue, and the fabricator earns ~45% on enormous capital. The pool is fat at the ends, where the barriers are: leading-edge process technology at one end, brand + ecosystem + demand aggregation at the other. It's starved in the substitutable middle. Acer's Stan Shih coined the name "smiling curve" for this shape, and it recurs across electronics, autos, and apparel. Manufacturing per se, the visible, employment-heavy, politically celebrated link, is often the chain's worst seat.
Never confuse the industry's protagonist with its profit-taker.
How do profit pools move over time?
Pools also move, so learn to read migration. In aviation the pool has moved toward loyalty/cards and the engine aftermarket. In autos (practice set, problem 6) it moved historically from assembly toward financing and aftermarket, and the EV transition is now shaking the component pool. Fewer moving parts mean a smaller aftermarket, and batteries are re-concentrating supplier power. In media it's moved from exhibition toward streaming platforms. At every link, ask yourself whether this link's force profile is improving or deteriorating, and whether anyone is integrating across links to capture a neighbor's pool. Apple designing its own silicon is backward integration into the chain's other fat end. Airlines' co-brand cards are forward integration into payments.