── ── Cognitive bias
Anchoring
When people estimate an unknown quantity they start from a reference point — an anchor — and adjust. The adjustment is almost always insufficient, so the final answer stays closer to the anchor than it should. This pattern holds even when the anchor is explicitly random and subjects are told to ignore it (Tversky & Kahneman 1974). It survives expertise…
Run Anchoring on a real problem
Bring something you're actually deciding — free, in the browser.
How it works
Run the Anchoring Analysis. Detect anchors, judge their informational content, decide deliberately.
1. Identify the anchor. Name the specific number explicitly — it often arrives as casual mention, "data," or expectation, not as a labeled offer. 2. Classify informational content. Genuinely informative → use as evidence. Partial (real signal + proposer's interest) → discount. Pure anchor → treat as zero information; bias still operates, apply structural defense. 3. Run the anchor-free counterfactual. What number would you have produced without hearing the anchor? Write it down first. 4. Consider an opposite-side anchor. Deliberately generate a number in the opposite direction and articulate the reasons — not to use as a counter, but to break the asymmetry. 5. If proposer: open bold but defensible. Bold openers consistently outperform soft ones; absurd anchors lose power and damage the relationship. 6. If respondent: name it, reset with your own independently-generated number, or work with it knowingly — discounting your natural agreement point. 7. Audit hidden anchors in forecasts. Last quarter, board expectations, prior-year baselines are all anchors. Identify which is doing the work. 8. Stop-rule: if the analysis doesn't change the decision, you may have rationalized. Re-run the anchor-free counterfactual.
When to use it
- user says 'what's a fair price / starting number / ballpark'
- counterparty just opened with a number
- user is about to negotiate salary, valuation, or deal terms
- user is making a forecast and a number has already been floated
- user asks about 'first offer', 'framing effect', or 'being anchored'
When not to use it
When the decision is routine and reversible, applying a formal method costs more than it returns.
Worked example
AI Fundraising and Valuation Anchors (2023–2026)
The 2023–2026 generative-AI boom is a live, high-stakes anchoring laboratory. When capital floods into a category faster than any comparable can be built, prices have to be estimated rather than looked up — and estimation under uncertainty is exactly where anchoring dominates. Two anchors did most of the work in this cycle: a small number of headline rounds (OpenAI, Anthropic, and later foundation-model and infrastructure raises) that reset what a "normal" AI valuation looked like, and Nvidia's soaring market capitalization, which became the reference point…
Install this skill (free, MIT)
npx skills add deciqAI/knowledge-skillsUseful? Star the repo — stars help other builders find it.
Related mental models
Authority bias is the automatic tendency to comply with directives from authority figures — often overriding one's own judgment, ethics, or evidence.
The availability heuristic: people estimate probability by how easily examples come to mind.
Maps each business unit on a 2×2 grid of market growth rate vs.
Continuous discovery (Teresa Torres, Continuous Discovery Habits, 2021) replaces one-off research with a weekly cadence of customer touchpoints by the team building the product, structured around…