Do referrals really help? What the hiring research does and does not say

Start with the denominator

When someone says a referral improves your odds, ask: odds of what, among whom? An interview rate among applicants is not an offer rate among interviewees. A hiring rate at one employer is not the share of all jobs filled through networking. Without the denominator, a precise-looking statistic can tell the wrong story.

In its May 2025 report, Ashby analyzed more than 38 million applications across 93,000 jobs from January 2021 through December 2024. Referred applicants advanced to interview at 40%, versus 3% for inbound applicants. Among those interviewed, 16% of referred candidates advanced to offer, versus 6% of inbound candidates. Those are different stages, not interchangeable success rates.

These are observational results from participating employers. Referred candidates may differ in fit, screening and prior relationships before applying. The comparison does not establish that adding a referral to an otherwise identical application causes the full difference. Nor does it measure the success rate of cold LinkedIn outreach.

Myth: the research proves that online applications are pointless

A channel can have a lower conversion rate and still account for many hires because it has far more applicants. More importantly, networking and applying are not mutually exclusive. A conversation can lead to a formal application; a direct applicant can build a relationship while the process is underway.

Claims such as “most jobs are never advertised” need a clearly defined population and a defensible source. They can mix internal moves, informal hiring, vacancies and completed hires. None of the studies discussed here establishes a universal hidden-market percentage. We would not use such a number to decide that you should stop checking actual vacancies.

Prioritize openings that match your experience, learn about the team and follow its application process. Public applications and referrals can both belong in the same search. The data does not justify abandoning either one.

Myth: more strangers means more opportunities

The 2022 Science study by Rajkumar and colleagues used randomized changes to LinkedIn connection recommendations across more than 20 million users. It found an inverted-U relationship between tie strength and job transmission: weaker ties helped up to a point, with differences across industries. This is stronger causal evidence about network formation than a simple comparison of referred and non-referred applicants.

But changing who appears in a connection recommendation is not the same intervention as sending mass messages. The experiment does not validate a particular outreach script, a daily invitation quota, or the claim that close relationships have no value. Our interpretation is that exposure to different circles can be useful; relevance and context still matter.

Reading the older referral studies carefully

The New York Fed staff report by Brown, Setren and Topa, revised in 2013 and later published in the Journal of Labor Economics, studied one U.S. corporation. It found referred applicants were more likely to be hired; an initial wage advantage faded, while longer tenure remained. That is useful evidence about a particular employer and period, not a contemporary market-wide guarantee.

A 2015 Glassdoor analysis of interview reviews examined differences by interview source. Comparing the midpoints of its reported adjusted ranges gives a gap of roughly 18 percentage points between referrals and applying online. That is a derived contrast, not a universal lift. Percentage points are an absolute difference, not an 18% relative improvement. The sample consists of self-reported interview reviews, not everyone who applied, and the analysis is observational.

These distinctions matter when reading the research highlighted on our homepage. Different datasets ask different questions. Their numbers should not be combined into one universal “networking multiplier,” and historical results should retain their dates.

What to change in your search

1. Keep applying to suitable roles. A public application is still a valid route.
2. Use conversations to resolve uncertainty about the role and your fit.
3. Make your actual capabilities easy for another person to describe.
4. Ask for an introduction only when it can be made honestly.
5. Judge your own process by useful learning and real progress, not a borrowed conversion rate.

The evidence supports taking relationships seriously. It does not support blaming job seekers who lack access to well-placed contacts, or treating a referral as proof of merit. Networks are unevenly distributed. A fair search strategy makes room for public applications, accessible communities and transparent introductions rather than pretending everyone starts with the same advantages.

For two related claims, see our examination of the hidden job market and rising application volumes. Neither supplies a universal probability of getting hired.