How to Write a Bear Case (That Isn't Just "It Could Go Down")
Most retail research has exactly one case: the reason you already want to buy the stock. That's not analysis, it's confirmation. Institutional research forces both sides into the open, built separately and each as strong as its actual proponents would argue it — because the version of the bear case you write yourself, while already convinced of the bull case, is almost never the real bear case. It's a strawman built to be knocked down.
The institutional process
- Build the bull case on its own terms. Steelman it — use the best plausible outcomes and the strongest version of management's own framing, not a wish list.
- Build the bear case independently, with equal rigor. Not "the stock is expensive." A real bear case names the specific mechanism: margin compression from new competition, a moat eroding faster than the multiple assumes, customer concentration risk, a regulatory shift, a cyclical peak mistaken for a structural trend.
- Assess the moat separately from both cases. A moat question is distinct from a valuation question. A company can have a widening moat and still be overpriced. A company can have a shrinking moat and still be cheap enough to matter. Conflating the two is one of the most common analysis mistakes.
- Find the crux. Almost every genuinely contested stock comes down to one specific disagreement — not ten reasons on each side, but one assumption that, if resolved, would collapse most of the argument on one side. Find it and you know what to actually watch.
The moat categories worth checking
| Moat type | What to look for |
|---|---|
| Network effects | Does the product get more valuable as more people use it? Is that flywheel accelerating or plateauing? |
| Switching costs | How expensive — in money, time, or risk — is it for a customer to leave? Is that cost rising or falling as alternatives mature? |
| Cost advantage | Can the company do something structurally cheaper than competitors — scale, location, proprietary process — that isn't easily replicated? |
| Intangible assets | Brand, patents, regulatory licenses, or trust that competitors can't simply build faster with more capital. |
| Efficient scale | Is the market only large enough to profitably support the incumbents already in it, discouraging new entrants? |
Walkthrough: the bull/bear split on a real, contested name
What AlphaLens's Bull vs Bear + Moat run concluded on PLTR
The bull case
Government revenue acceleration and stickiness. Deep embedding across the DoD, NSA, and federal law enforcement creates switching costs few competitors can match, with renewal rates above 95%. Sustained 100%+ US commercial growth is what keeps this case alive.
The bear case
Valuation detached from growth. At 40x+ forward revenue, PLTR trades at 8–15x the EV/revenue multiple of high-growth SaaS peers like Datadog and CrowdStrike. Compression back toward peer multiples (~15x revenue → a ~$60 scenario, versus ~$158 at today's ~40x) does most of the downside work.
Neither side disputes the execution — nine straight beats, and both cases grant it. The framework's moat verdict is what sharpens the disagreement: the government/defense moat is wide and durable, but the commercial moat is narrower and actively eroding as Databricks and the major cloud platforms close the gap — Palantir is, in the report's phrasing, "a government solutions company that aspires to be an enterprise software company — not yet the reverse." So the crux AlphaLens identifies is one assumption: whether a government-services company deserves an enterprise-software multiple. Because most of PLTR's valuation is riding on commercial success specifically, the number to watch is US commercial growth — a deceleration below roughly 80% YoY is the print that tips the multiple toward the bear case.
Do it yourself, then let AlphaLens check your work
Writing both cases out is a discipline you can start with a blank page — no tool required. Once you have a number in mind for either side, this one helps stress-test it:
What's Priced In Calculator
Back into the growth rate today's price already assumes — useful for pressure-testing whichever side of the bull/bear split you're leaning toward.
Open the calculator →Thesis Invalidation Calculator
Once you've picked a side, set the specific print that would prove you wrong — before you're holding the position.
Open the calculator →Common questions
How do you write a good bear case for a stock?
Build it with the same rigor as the bull case — specific, falsifiable claims about what would have to go wrong, not a vague sense the stock is expensive. Name the exact metric or event that would prove it right.
What is an economic moat?
A durable competitive advantage that protects a company's profits from competitors — network effects, switching costs, cost advantages, intangible assets, or efficient scale. What matters most is whether it's widening, stable, or eroding.
Why build the bull and bear case separately instead of just picking a side?
Building both independently, each as strong as its actual proponents would argue it, forces you to engage with the real disagreement instead of a strawman. Most contested stocks come down to one specific crux assumption, not ten scattered reasons per side.
Run this on any stock in seconds
Bull vs Bear + Moat is one of 15 frameworks in AlphaLens — built from live filings and current news, not a frozen snapshot of either side's argument.
Try Free for 7 Days →Related: All 15 frameworks · Fair Value Stress Test · Earnings Quality Analyzer · Learn hub · PLTR full analysis · All free tools