UNIVERSITY PARK, Pa. — Whether they are driven by economic necessity or a desire for greater control over their careers, a growing number of workers are turning to online gig-work platforms to supplement their income with freelance projects like coding or website designing. These digital labor platforms sometimes introduce new systems aimed at helping workers stand out in an increasingly competitive gig economy, but a new study by a Penn State researcher and collaborators suggests that these systems may not always be as helpful as intended.
According to study co-author Chenhui “Julian” Guo, assistant professor of supply chain and information systems at Penn State’s Smeal College of Business, the research is rooted in “signaling theory,” which examines how parties try to resolve information gaps. In the context of hiring, “signals” like reviews or higher education degrees can provide valuable information about a job candidate to an employer. But when labor platforms introduce new, untested signals, like verified statuses or special guarantees, the effects can be unpredictable — and, in some cases, actively detrimental to the potential employees who opt to use them.
The researchers analyzed anonymized proprietary data from a large online labor platform to understand the impacts of a novel worker-offered guarantee on prospective employees who opted to use the new system. The study was recently published ahead of print in the journal Information Systems Research .
In the following Q&A, Guo discussed the study and its findings, as well as the broader value of signals in the growing gig economy and how employers, freelancers and labor platforms should exercise caution when deciding how to use and interpret new signals.
Guo: Hiring someone online can feel like betting on a stranger. Employers often have to decide between many freelancers they have never met, sometimes from different countries, and they cannot directly observe how skilled, reliable or easy to work with those freelancers really are.
That is why signals matter. Ratings, reviews, certifications, work history and platform badges all help employers make a judgment before hiring. Platforms keep adding new signals because they want to make the market feel safer and help good workers stand out. But our study shows a twist: A signal only works if people trust what it means.
Guo: The platform introduced a feature that let freelancers attach a financial guarantee to their bids. In theory, this guarantee is the worker’s way of saying, “I am confident enough in my work that I am willing to put some of my own money on the line.” The platform suggested a deposit equal to 10% of the bid amount. If the worker was not hired, or if the job was completed successfully, the money was returned. But if the worker was hired and failed to complete the job satisfactorily, the money was forfeited and donated to charity — not paid to the employer.
This is important because it is different from a normal money-back guarantee, which protects the buyer directly. This worker-offered guarantee did not compensate the employer if things went wrong. So, from the employer’s perspective, it did not really remove the risk of hiring the wrong person.
The platform intended the guarantee to help workers stand out, but it had the opposite effect. Employers were less likely to hire workers who used it, and later outcomes suggested that employers’ skepticism was not misplaced: Workers who used the guarantee received lower ratings, were rehired less often and took longer to finish their work.
Guo: The simple answer is that employers asked themselves, “Why does this worker need this extra signal?”
In online labor markets, strong workers often already have good ratings, solid reviews and a track record of completed jobs. They may not need to add another layer of proof. But workers with weaker or less established records may have stronger incentives to use a guarantee to get noticed.
So, employers may have interpreted the guarantee very differently from what the platform intended. Instead of thinking, “This worker must be confident,” they may have thought, “This worker is trying unusually hard to convince me.” In a crowded market, that can look less like confidence and more like a warning sign.
Guo: For job seekers, the main lesson is that more proof does not always create more trust. In a competitive market, it is tempting to use every available platform feature to stand out. But if a new signal is not viewed as credible, or if it appears to compensate for weaker existing credentials, it may hurt rather than help.
For employers, the study suggests that new platform signals can provide useful information, but not always in the way the platform intended. Employers should think carefully about who is using a signal and why. A feature designed to communicate confidence may sometimes reveal uncertainty or lower expected quality.
As for the labor platforms themselves, the broader lesson is that trust mechanisms need to be tested in practice. Adding a badge, guarantee or commitment device does not automatically improve trust. Platforms need to evaluate how users actually interpret new features and whether those features improve market outcomes. For example, they could first test a new signal in a limited “sandbox” setting, observe how employers and workers respond, and check whether it actually improves hiring and job outcomes.
The broader lesson is simple: Platforms can design signals, but users decide what those signals mean. Sometimes, a feature meant to build trust can become a red flag.
Co-authors on this study include Qiang Gao, City University of New York; Mingfeng Lin, Georgia Institute of Technology; and Yong Liu, University of Arizona. This work was supported by the Ewing Marion Kauffman Foundation and the Center for International Business Education and Research, Georgia Institute of Technology. This content is the sole responsibility of the authors and does not necessarily reflect the views of the funders.
Information Systems Research
Data/statistical analysis
Not applicable
When “Signals” Boomerang: Employers’ Reactions to a Novel Signaling Mechanism
22-Sep-2026