Skip to main content

Advertising

How to Hedge Without Losing Your Edge

🧠

Neuromarketing Principle

Personal and higher-likelihood hedges make communicators seem more confident, which can increase persuasion without removing the impression of uncertainty.

πŸ’‘

Application

Discover how small language choices can make recommendations more persuasive in customer reviews, sales conversations, and AI chatbots.

Two people recommend the same product. One says: "It seems like a great choice." The other says: "I think it's a great choice. This is not just a matter of wording. In a recent paper, Effectively communicating uncertainty: The persuasive impact of different types of hedges, Oba, Berger and Boghrati show that the way we express uncertainty can influence how confident and persuasive we appear.

Why some hedges are more persuasive than others

Hedges are words or phrases such as "might," "probably," "I think," and "it seems" that communicate some level of uncertainty. After analysing over 4.7 million Amazon book reviews and conducting six additional studies, the researchers identified two dimensions that influence how persuasive these expressions are.

1. Personal vs. general perspective

Compare "It feels like this is a great option" with "I feel like this is a great option." Adding "I" connects the speaker more directly to the statement. Rather than presenting an opinion as a general impression, the speaker seems personally committed to it. Participants rated speakers using personal hedges as more attached to their statements (4.90 vs. 4.19 on a 7-point scale). This greater sense of personal involvement also made them appear more confident.

2. Suggested likelihood

New to New Neuromarkteing? Join X evidence-based marketers to get the latest insights and improve your marketing results.

Hedges also differ in how likely they make something sound. "Probably," for example, suggests a higher likelihood than "possibly." The researchers describe this as a continuum: every hedge communicates a different degree of uncertainty.

To illustrate: earlier research on intelligence communication, estimated that "likely" communicates roughly 75% probability, while "almost certainly" suggests around 93%. In their own research, three independent coders rated 55 commonly used hedge expressions on a 7-point likelihood scale. Expressions such as "probably," "likely," and "arguably" fell towards the higher end, while "possibly," "might," and "could" fell towards the lower end.

The difference also mattered to consumers. In one experiment, participants interpreted higher-likelihood hedges as expressing an average probability of 73%, compared with 65% for lower-likelihood hedges. These percentages represent average perceptions of the phrases used in each condition, rather than fixed probabilities that apply to every individual hedge.

Who says it matters

The researchers also tested whether personal hedging works differently depending on who delivers the recommendation.

From a friend. A friend saying "We think you'd like this sweater" was more persuasive than saying "You may like this sweater" (4.54 vs. 4.08). The personal wording made the recommendation seem more confident.

From a brand. When the same type of recommendation came directly from a clothing brand, personal wording still increased perceived confidence, but did not significantly increase persuasion. Consumers know that brands want to sell their products, making brand confidence less informative.

From a brand's chatbot. The researchers then tested recommendations delivered through a brand's chatbot. This time, "I think you'd like this" was more persuasive than "You might like this" (3.42 vs. 3.16).

The researchers explain this through anthropomorphism: a human-like chatbot gives the brand a more personal way to communicate. This may allow brands to benefit from the persuasive effects of personal hedging, which are less evident in direct brand communication.

When confidence has credibility

Another important finding concerns the experience of the person giving the recommendation.

In one experiment, participants imagined ordering at a restaurant. An experienced server saying "I feel like this is an excellent choice" was considerably more persuasive than saying "It feels like this is an excellent choice" (5.37 vs. 4.53). For a less experienced server, the difference was much smaller (4.30 vs. 3.97).

This suggests that personal hedging is particularly effective when the communicator has the credibility to support their confidence. The wording can make someone sound more convincing, but it cannot replace actual expertise.

When this works, and when it doesn't

The findings offer several practical opportunities, but there are some limitations to keep in mind.

The effects are subtle. In a controlled experiment, personal hedges received an average persuasion score of 5.22 compared with 4.87 for general hedges, on a 7-point scale. This is a meaningful linguistic difference, but not a guaranteed increase in conversions.

The source matters. Personal hedging was more persuasive in interpersonal recommendations and chatbot communication, but not in direct brand messages. Simply changing "might" to "we think" in your website copy may therefore not have the same effect.

The setting matters. Most experiments used hypothetical scenarios and self-reported purchase intentions rather than actual purchasing behaviour. The Amazon analysis measured review helpfulness, which provides real-world evidence but does not establish causation.

Only English was tested. Whether Dutch equivalents such as "ik denk," "volgens mij," and "misschien" produce the same effects remains to be investigated.

Hedging is insurance

Should you simply remove the hedge and say "You'd like this"? Not necessarily. In their final study, the researchers compared a chatbot using a personal hedge ("I think you'd like this"), a general hedge ("You might like this"), or no hedge at all ("You'd like this"). Initially, the unhedged recommendation was no more persuasive than either hedged version. However, when participants imagined buying the recommended product and not liking it, the difference became clear. The chatbot without a hedge was perceived as considerably more overconfident (5.68 vs. 4.52 and 4.27 for the hedged versions). Attitudes towards the brand also dropped (3.15 vs. 3.43 and 3.54).

Both hedged versions offered similar protection against negative reactions. That is the insurance: hedging did not significantly reduce initial persuasion, but helped protect the brand when a recommendation turned out to be wrong. The researchers describe personal hedging as a communication "sweet spot": it increases perceived confidence compared with general hedging while retaining the protection of communicating uncertainty.

Take-home points

  • Make uncertainty personal. "I think" and "I feel like" can make recommendations sound more confident than general expressions such as "it seems like."

  • Choose your wording carefully. "Probably" communicates a higher likelihood than "possibly." Use wording that matches the evidence you have.

  • Consider who is speaking. Personal hedging can be effective in interpersonal recommendations and through human-like chatbots, but not necessarily in direct brand communication.

  • Confidence needs credibility. Personal hedging has a stronger persuasive effect when the communicator is experienced or competent.

  • Don't underestimate uncertainty. Hedges can help protect your brand from negative reactions when recommendations turn out to be wrong, without significantly reducing initial persuasion.

Β 

Written by Kire Kuil on November 17th 2026

Futher reading

Advertising

The Precision Effect: How Exact Words Make Your Product Feel One-of-a-Kind

You are looking for a new sweater. One label says made in Europe. The other says knitted in Amsterdam. Same fabric, price and look. But something about the second one feels different. It feels more…
🎯

For marketers, by marketers

Ready to enhance your portfolio by starting writing for x hungry marketers? Become an Author.