Anchoring Bias in Finance: A Startup Investor’s Guide

Brian Nichols is the co-founder of Angel Squad, a community where you’ll learn how to angel invest and get a chance to invest as little as $1k into Hustle Fund’s top performing early-stage startups.

A founder says the startup is worth $20 million. The prior round was $16 million. A respected lead is already in. The first two are numerical anchors. The lead’s reputation is a separate social cue.

Anchoring bias explains the risk. For an angel investor, the response starts before the first number lands.

This content is educational and does not constitute investment, legal, tax, accounting, or valuation advice. All startup valuation examples are hypothetical and simplified, not projections. Startup investments are speculative, illiquid, long-term, and can result in total loss. Past results and current marks do not guarantee future results. Outcomes depend on the governing documents, fees, carry, taxes, cap table, valuation policy, and timing. Review the governing documents and consult qualified independent legal, tax, accounting, valuation, and financial advisers.

What is anchoring bias in finance?

Anchoring bias is the disproportionate influence of a numerical reference point on a later estimate or decision. A stock’s purchase price, a home’s list price, an analyst’s old forecast, and a founder’s valuation ask can all serve as anchors. The bias appears when the number receives more weight than its relevance deserves.

Anchoring describes an effect, not one universally accepted mental process. Anchoring-and-adjustment is one account: you start from a reference point and stop adjusting before reaching a fully considered estimate. Research by Nicholas Epley and Thomas Gilovich found evidence for insufficient adjustment especially with self-generated anchors, such as an investor’s own initial estimate or rule of thumb.

Externally supplied anchors need not work the same way. A founder’s ask may influence you without deliberate adjustment. One proposed mechanism, selective accessibility, says that considering an anchor makes anchor-consistent information easier to retrieve. Experimental research supports that account. No single explanation covers every anchoring result.

Amos Tversky and Daniel Kahneman demonstrated the basic effect with a number that was ostensibly generated by chance. A wheel was rigged to stop at 10 or 65. People first judged whether the percentage of African countries in the United Nations was higher or lower than that number, then gave an exact estimate. The median estimates were 25% and 45%, respectively. The irrelevant number moved judgments. The experiment appeared in their foundational 1974 Science paper.

What the evidence does and does not show about startups

Direct evidence that prior-round prices cause anchoring among angel investors is limited. The magnitude of any effect in startup investing is unknown.

There is analogous evidence in other markets. Public-company analysts have produced forecasts that moved too little from an industry median. Prior target-company stock-price peaks have been associated with merger offers and deal acceptance. Professional real estate agents and amateurs have produced estimates influenced by list prices. These studies cover public equities, mergers and acquisitions, and property pricing, all unlike early-stage financing.

The analogy makes reference-point risk plausible. It does not establish how often startup investors anchor, whether prior rounds cause the behavior, or how far an estimate moves. The process below is a safeguard under uncertainty, not a treatment backed by a measured startup effect size.

Why startup deals are vulnerable to reference points

An early-stage valuation is a financing price shaped by supply, demand, dilution, milestones, security rights, and negotiation. It is not a precise reading from a continuous public market. A Simple Agreement for Future Equity (SAFE) valuation cap is also different from a priced-round pre-money valuation.

The ambiguity creates room for shortcuts. Investors may borrow a number from the founder, a prior financing, a comparison company, or a lead investor. Later analysis may cluster around it.

Startup decisions contain many legitimate inputs. In a survey of 885 institutional venture capitalists at 681 firms, the 433 responses to the diligence-time question averaged 118 hours, while the 439 responses to the reference-call question averaged 10 calls. The NBER paper documents several selection and valuation factors. It did not test anchoring among those venture capitalists.

More inputs can challenge an initial number, or they can supply reasons to defend it. Information order and a written update rule help distinguish the two.

Four possible anchors and one adjacent effect

1. The headline valuation or SAFE cap

The first price in a deck can frame the negotiation. A $20 million ask may make $17 million feel cheap even when the business evidence would have led you to a much lower range.

Ask two separate questions. What range does the current evidence support? What ownership, dilution, and return possibilities would the actual security create under its governing terms? A discount from the founder’s ask answers neither one.

2. The prior-round valuation

“The last round was at $16 million” can sound like a floor. It is a historical transaction under earlier conditions, with a particular cap table, evidence set, investor mix, and security. The company may have made real progress. It may also have missed the milestones that supported the earlier price.

The current 2025 IPEV Guidelines address fair-value reporting, a different job from deciding whether to join a round. Their treatment of recent financing is still useful. A recent orderly transaction price can calibrate valuation inputs and may provide a starting point. It is not automatically fair value or a standalone valuation method. Current performance, market conditions, transaction context, and differences in security rights still matter.

Use the prior round as evidence, never an automatic floor. Rebuild the case from current traction, burn, runway, team, market, financing needs, and the rights attached to the security you would own. Our startup valuation guide explains several lenses for triangulating a range.

The evidence boundary remains important: IPEV tells managers how to approach fair-value estimation. It does not show that startup investors are anchored by prior rounds or quantify such an effect.

3. The lead investor’s price and reputation

Two influences arrive together here. The lead’s price is a possible numerical anchor. The lead’s reputation creates social proof and persuasion pressure, which are related decision risks rather than the same mechanism.

A credible lead can contribute diligence, negotiate terms, and signal commitment. Their decision also reflects their fund size, ownership target, portfolio, relationships, and strategy. Those inputs may differ from yours.

Separate the questions. What evidence came from the lead’s work? Would the price and security fit your own thesis and portfolio if the lead’s name were hidden?

4. Your entry price or latest portfolio mark

After investing, your entry valuation can become the reference point for every update. A higher follow-on price feels like validation. A lower price feels like failure. Neither reaction answers whether a new check is attractive on today’s evidence and terms.

A financing mark can update your records without becoming a guaranteed sale price. In Regulation D private placements, securities may be difficult to resell and may need to be held indefinitely, while disclosure can be limited. The SEC’s private-placement bulletin describes these risks.

Our general partner Eric Bahn frames the question well: “When it comes to markets, you always need to factor in: is something that is true today going to be true tomorrow?”

5. First impressions of the founder are adjacent, not classic anchoring

A charismatic pitch can create an early “exceptional founder” label. A nervous pitch can create the opposite label. Calling this classic numeric anchoring blurs distinct concepts.

Primacy gives early information more influence. A halo effect lets one positive trait color judgments of unrelated traits. Persuasion effects can make delivery feel like evidence. Any of these can shape how you interpret later facts without operating through a numerical anchor.

The practical response is similar. Record the observation instead of the verdict. “Answered the customer-acquisition question with cohort data” is evidence. “Impressive” is a global label. Use the same founder questions and scoring definitions across deals so presentation quality does not quietly change the bar.

A startup anchoring bias example

Hypothetical, simplified valuation example: Assume every figure is a post-money valuation cap on otherwise identical post-money SAFE terms. A founder proposes a $24 million cap and points to an $18 million cap from the prior round. A respected lead has accepted the $24 million cap. The angel negotiates down to $20 million and feels disciplined because the number is below the ask.

Now reverse the information order. Before seeing either cap or the lead’s name, the angel reviews the same company, studies relevant comparisons, and writes an $11 million to $15 million acceptable post-money valuation cap range for that same post-money SAFE. The $20 million “discount” no longer looks like a bargain. The first process stayed close to outside reference points. The second introduced an evidence-linked reference range first.

That independent range can also become an anchor. It could be wrong because comparisons are poor, material facts are missing, or the investor’s assumptions are weak. Its value is diagnostic: the gap forces the investor to name the evidence that justifies moving instead of quietly drifting toward the ask.

A five-stage process for reducing anchoring risk

You cannot stop seeing reference points. You can control their order, expose their influence, and require a reason for each material update.

1. Facts: list reference points and build the outside view

Start your notes with the company’s stage, sector, geography, traction, business model, financing need, next milestone, and material risks. Label each input as an observed fact, founder estimate, third-party evidence, or your assumption.

Also list the possible anchors already in the room: the founder’s ask, prior financing, comparison headlines, your entry price, and your first estimate. Put nonnumeric pressures, such as the lead’s reputation and your first impression, on a separate line.

Select comparisons using stated criteria before you know which examples support the proposed price. If you have already seen the terms, a range reconstructed from a blank page may still be contaminated. Ask a second reviewer who has not seen the terms to make a blinded estimate, or use the consider-the-opposite test in stage four before comparing ranges.

2. Range: form an estimate and write decision conditions

Use a range instead of one magic number. State what would make you invest, pass, or wait, including acceptable terms, portfolio fit, check size, and unanswered questions. A short investment memo keeps your reasoning separate from the pitch narrative.

This is where our general partner Elizabeth Yin is blunt: “The more disciplined you are in your thought process/rubric, the more you can improve over time.”

Your range is provisional. Write the facts that would move its low end, high end, or both. That update rule guards against clinging to a self-generated anchor.

3. Reveal: compare the outside terms and log the update

Reveal the stated price, prior round, lead, and full terms. Compare them with your range, then record:

• the size and direction of the gap;
• the new evidence that changed your view;
• the assumptions that did not change;
• security rights, fees, carry, taxes, cap-table details, and timing differences that weaken a comparison;
• the event that would make you revisit the decision.

“The lead is excellent” is incomplete. “The lead’s customer references changed my estimate of sales-cycle risk” names evidence and mechanism.

4. Test: consider the opposite and run a pre-mortem

Write the strongest case that each reference point is wrong in both directions. What would make the company worth much less? What evidence could support a higher range? Which fact are you dismissing because it conflicts with the first number or story?

Two experiments found that a structured consider-the-opposite strategy reduced anchoring. Those experiments were not startup-investment trials. A pre-mortem adapts the general tactic to this decision: imagine the investment failed, then name the most plausible cause, its earliest signal, and the diligence question that could expose it now.

5. Decide: make judgment independent before making it social

As a practical safeguard, each investor records a range, key risks, and recommendation before hearing the most senior or confident person. The group reveals views together, discusses the spread, and assigns someone to develop the dissenting case. The leader speaks last. This sequence is a process choice, not an empirically validated intervention for angel groups.

Then record the decision, its evidence, and its revisit trigger. Inside Angel Squad, we use the same practical order: individual views first, peer discussion second. We treat it as a safeguard against borrowed conviction, not proof that a particular sequence eliminates bias.

Anchoring bias versus confirmation bias

Anchoring bias gives a numerical reference point too much influence. Confirmation bias shapes which evidence you seek, trust, or dismiss. The effects can overlap through selective accessibility: considering an anchor can make consistent information easier to retrieve, and that information can then guide later search and weighting. The labels remain distinct even when the same evidence pattern reflects both.

Suppose a founder’s $24 million cap becomes your anchor. You may then search for comparisons near $24 million, give more weight to bullish customer comments, and dismiss a slower-growth comparison as irrelevant. The anchor set the neighborhood. Confirmation bias helped you stay there.

Fight both by fixing comparison criteria in advance, recording disconfirming evidence, and requiring a written reason for every material update.

Frequently asked questions

Can experienced investors avoid anchoring bias?

Experience helps an investor choose more relevant reference points. It does not guarantee immunity. The property-pricing experiment cited earlier found list-price effects among professionals as well as amateurs. That evidence is analogous. It does not measure angel investors or startup valuations.

Is every benchmark an anchor?

Every numerical benchmark can become an anchor. A relevant benchmark remains useful when you choose it before seeing the preferred answer, match it on meaningful characteristics, use a range, and state where the comparison breaks. The problem is unjustified weight, not reference data itself.

Can a checklist eliminate anchoring bias?

No. A checklist can improve information order, consistency, and accountability. It cannot make sparse data complete, settle which psychological mechanism is operating, or turn a startup valuation into objective truth. Use the process to expose judgment, then size the risk for the possibility that your judgment is wrong.

The founder’s ask, the last round, and the lead investor can all inform a decision. None should make the decision before you do. Form your range, reveal outside reference points, test the gap, and move only for a stated reason.

If you want to practice that discipline on real startup opportunities with other operators, apply to Angel Squad.