New contributor rates take effect January 2027. Review the contribution guidelines for full details.

Signaling Theory

Why costly proof beats cheap talk

Signaling Theory
Idea In Short

Replace claims a weaker competitor could copy with proof that would cost you to break. Picture the weakest plausible version of your company and ask whether it could afford to send the same signal. If it could, the signal is decoration. Offer the guarantee that would sting most if you underdelivered, such as month-to-month terms, a public record of past outages, references buyers can call without you, or fees tied to results measured in the client's own numbers. Buyers should ask the same question in reverse: who pays if the claim is false? Costly signals earn their keep when doubt is high, at a first purchase or after a public failure, and they are wasted on buyers who already trust you.

What is signaling theory?

It is the economic study of how a party with private information can convince a less-informed party that its claim is true. A promise carries weight when breaking it would cost the sender something.

Why does a signal have to be costly?

A cheap signal can be copied by anyone, so it tells the buyer nothing. A signal separates strong senders from weak ones only when it costs the weak sender more to send.

How can I test whether my own claim is credible?

Picture the weakest plausible version of your company and ask whether that version could afford to send the same signal. If it could, the signal is decoration, however polished it looks.

Why did Hyundai's warranty work as a signal?

Honoring a 10-year, 100,000-mile powertrain warranty would have ruined a carmaker whose engines failed early. Buyers could therefore read the warranty as a claim about the engines.

Is an expensive signal the same as a costly one?

No. The two overlap only when the expense hurts a weak firm more than a strong one. FTX paid for an arena naming deal that a firm spending its customers' deposits could afford as easily as a sound one.

Why can review averages mislead buyers?

Sellers can buy reviews. Researchers who followed about 1,500 Amazon products found that ratings rose while fake review campaigns ran and sagged when they ended, so the trend and the bad reviews carry the useful information.

What should I look at when I am the buyer?

Sort by lowest rating first and read the company's replies. A business that answers its one-star reviews in public, with specifics about what went wrong and what it changed, is paying a small, steady cost.

Which signals can a seller offer?

Examples include a public status page with every past outage, month-to-month terms or a paid pilot, references the buyer can call without you on the line, and fees partly tied to a result measured in the client's own numbers.

Do values statements count as signals?

They count only when something checkable sits behind them. Attach a pledge to a figure you publish every year, such as emissions or refund rates, and it starts doing real work.

When is a costly signal not worth buying?

When doubt is already low. A firm with 15 years of renewals and a steady referral pipeline pays for reassurance its buyers stopped needing long ago.

Two agencies pitch the same marketing director in the same week. Both decks promise measurable growth and a senior team, and each has the obligatory slide of client logos. Near the end, one of them adds a line the other doesn't, letting the client leave at the end of any month with no notice and no penalty.

That clause is the only part of either pitch the director can't shrug off. Any agency can say it gets results, but one that expects to be fired in month two can't afford to make firing it free.

Economists study this under the name signaling theory, and it explains a lot about which claims get believed. A promise carries weight when breaking it would cost the sender something. It carries more when that cost would land harder on a weaker competitor.

The cost has to land unevenly

Michael Spence worked out the mechanics in 1973, in a paper about why employers care so much about degrees.1 Finishing a degree is harder for people who'd be weaker at the job, he argued, so the credential sorts candidates even when the coursework teaches little they'll use. That paper became part of the work that won him a share of the 2001 Nobel Prize in economics with George Akerlof and Joseph Stiglitz. The cost is what makes the degree informative. If everyone found it equally easy, it would tell an employer nothing. The Nobel committee's own summary calls this a fundamental insight: signaling cannot succeed unless the cost differs enough among the senders.

The requirement is stricter than it first sounds. A signal does not need to be expensive in absolute terms, and it does not need to be impressive. It needs to cost a weak sender more than a strong one, a condition the economics literature describes as differential cost.2 A degree meets the test because the same four years are harder for the applicant who would struggle at the job.

Hyundai ran the same logic with cars. The company was in trouble: By 1998, its U.S. sales were heading for their lowest level ever, roughly 90,000 vehicles, and the brand had become a punch line on late-night talk shows. For its 1999 models the company offered a 10-year, 100,000-mile powertrain warranty, at a time when Lexus covered powertrains for six years. Honoring that promise would have ruined a carmaker whose engines failed early, so buyers could read the warranty as a claim about the engines. In 2004 Hyundai was selling more than 400,000 cars a year in America, though the company had spent years fixing quality too, so the warranty can't take all the credit. Hyundai's launch material paired the powertrain term with a five-year, 60,000-mile bumper-to-bumper warranty, against four years and 50,000 miles at Lexus, so the whole package signaled, not only the headline number.

The cost doesn't have to be money. Exposure works too, because a claim anyone can check is expensive to make falsely. A Race Against Blindness, which funds childhood blindness research through vehicle sweepstakes, publishes an explainer that sends readers to the Internal Revenue Service [IRS] database, its state charity registrations and three years of annual reports to check for themselves. Copying that page's wording would be easy for an organization with something to hide, but the public records behind it would expose the copy within an afternoon of checking.

All three examples give you the same quick test, whether you're making a claim or weighing one. Picture the weakest plausible version of the company behind the claim and ask whether that version could afford to send the same signal. If it could, the signal is decoration, however polished it looks.

Where people misread signals

The most common mistake is treating an expensive signal as a costly one, when the two only overlap if the expense hurts a weak firm more than a strong one. In March 2021, the crypto exchange FTX agreed to a 19-year, $135 million deal to put its name on the Miami Heat's arena. Plenty of retail customers read it as permanence. The price tag was real, but a firm spending its customers' deposits could pay it as easily as a sound one could. FTX filed for bankruptcy in November 2022, and two months later a bankruptcy judge terminated the naming deal and Miami-Dade County began stripping the FTX name off the building.3 At the time, FTX had paid the county about $20 million of the $135 million, so most of the signal was a promise of future payments from a firm that could not keep making them.

Review counts get misread the same way. A product with 2,000 reviews and a 4.6 average feels safer than one with 40, and sellers know it. Researchers from the University of California, Los Angeles [UCLA] and the University of Southern California [USC] followed about 1,500 Amazon products whose sellers had bought fake reviews, and ratings rose while the campaigns ran, then sagged when they ended. Average ratings climbed about 5% at their peak, from 4.3 to 4.5 stars. In the seven weeks after the buying stopped, the share of one-star reviews on those products climbed from about 8.5% to more than 13%. The useful information sat in the trend and in the bad reviews, both of which the star average smoothed away.

Regulators have started to raise the price of this particular fake. In August 2024 the Federal Trade Commission [FTC] announced a final rule that prohibits buying or selling fake reviews and lets the agency seek civil penalties against knowing violators.4 A cost imposed from outside changes who can afford to send the signal, which is why enforcement, not exhortation, restores the information in a rating.

So when you're the buyer, sort by lowest rating first and read the company's replies. A business that answers its one-star reviews in public, with specifics about what went wrong and what it changed, is paying a small, steady cost that a seller propped up by bought reviews rarely bothers with.

Signals also wear out once a market starts chasing them. A trade-show "best new product" ribbon means something the first year. In the fifth year half the exhibitors seem to have one, and there are consultants who'll draft the entry for a fee. At that point the ribbon mostly measures effort spent winning ribbons, which is Goodhart's law applied to reputation. Ask instead what it would cost the holder to lose the badge.

Picking a signal a weaker rival can't copy

Sellers have their own reason to care. When buyers can't tell good providers from bad ones, they pay everyone the average price, the better providers end up underpaid, and some of them leave, which is how adverse selection drags a market's quality down. A well-chosen signal is the way out of being priced as average. It has to be built around what the buyer fears, though, rather than around what the seller is proudest of.

Write down the buyer's biggest fear in a single sentence. For a first-time software customer it's often downtime and lock-in, and for a consulting client it's six months of fees that end in a deck nobody opens. Then list the proofs a mediocre provider would refuse to offer.

  • A public status page that keeps every past outage on record, with the write-up of what went wrong
  • Month-to-month terms, or a paid pilot the client can walk away from after phase one
  • References the buyer can call without you on the line
  • Fees partly tied to a result measured in the client's own numbers, which weak agencies resist, since their reports are often the only place their results look good

Each of these works the way the Hyundai warranty did, costing a good provider almost nothing and a bad one a great deal. Offer the one that would sting most if you underdelivered. Buyers usually notice which guarantee a seller left off the list. Check also how far the commitment reaches: Hyundai's 10-year powertrain term applies to the original owner and drops to five years and 60,000 miles for later owners, which caps the cost the company carries without removing the signal at the point of purchase.

Values statements sit at the other end of the scale. A sustainability pledge or a "customers come first" banner takes an afternoon to write and costs nothing to ignore, and readers have learned to file most of them under virtue signaling unless something checkable sits behind them. Attach the pledge to a figure you publish every year, such as emissions or refund rates, and it starts doing real work.

Costly signals aren't always worth buying, though. They earn their keep when doubt is high, at a first purchase or in the year after a public failure. A firm with 15 years of renewals and a steady referral pipeline is paying for reassurance its buyers stopped needing long ago, and that money would do more good in the product.

Summary

Buyers discount what sellers say about themselves, and they are right to. What moves them is proof that would hurt to fake, like a warranty with real money behind it or records anyone can pull up in an afternoon. From either side of the table, ask who pays if the claim turns out to be false. When the honest answer is nobody, the claim is filler. The test works the same way for a trade-show ribbon, a values statement and a review average: each is worth what it would cost a weaker competitor to copy. Today, open your current proposal or homepage, highlight every claim a weaker competitor could copy word for word, and replace one of them with something a buyer can check or that would cost you to break.

References

    Citation

    Cite this article

    Sridharan, M. A. (2026, November 19). Signaling Theory. Think Insights. https://thinkinsights.net/community/signaling-theory (Accessed [[ACCESS_DATE]])

    Author
    Mithun A. Sridharan Thought Leader
    I'm Mithun A. Sridharan, Founder of this website - Think Insights - on Strategy, Management Consulting, Leadership, Digital Transformation, and Data Literacy. Follow me on social media or connect with me on LinkedIn for updates.