Who job seekers actually ask for a referral
Every referral search this tool ran for 180 days, and which of the people it showed the job seeker chose to open. Peers over senior people, measured inside display-position bands so the answer is about the reader and not about our ranking.
The finding
Across 242 real referral searches and 6,685 people shown, job seekers opened an individual contributor or manager 5.38% of the time (110 of 2,044) and a director or executive 1.32% of the time (22 of 1,665).
Holding display position fixed, the estimate is 1.49x in favour of peers at p = 0.114. That is a direction, not an established effect, and this page publishes it as one. It is not a reversal: no band contradicts it and peers lead in 3 of the 4 bands.
Verdict on 24 September 2026: CONFIRMED-POOLED-ONLY. This page renders whatever verdict the dataset carries; it does not keep a previous one. Snapshot window 180 days.
Inside display-position bands
This is the table that matters, and it is the reason the headline number on this page is smaller than the raw one. Our search puts peers near the top, so a reader who simply clicks what is in front of them produces a peer-heavy number whatever they believe. Splitting by where a person was shown removes that.
| Shown at rank | Peers opened | Peer rate | Senior opened | Senior rate | Ratio | Power |
|---|---|---|---|---|---|---|
| 1-5 | 88 / 674 | 13.06% | 9 / 96 | 9.38% | 1.39x | underpowered (z 1.017) |
| 6-10 | 13 / 322 | 4.04% | 8 / 262 | 3.05% | 1.32x | underpowered (z 0.635) |
| 11-20 | 3 / 428 | 0.70% | 5 / 471 | 1.06% | 0.66x | underpowered (z -0.575) |
| 21+ | 6 / 620 | 0.97% | 0 / 836 | 0.00% | no senior openings | decisive (z 2.85) |
"Decisive" means the band is powered to overturn or support the claim on its own (|z| at least 1.96). An underpowered band is not evidence against the direction; it is a band with too few openings in it to vote. Bands contradicting the direction: 0. Bands where peers lead: 3 of 4.
Download the full dataset as JSON (CC BY 4.0: free to cite with a link to this page).
The raw gap is bigger, and you should not quote it alone
Pooled across every position, peers are opened 4.07x as often as senior people. We publish that figure for comparison and we do not stand behind it as a statement about readers, because it contains our own ranking. The position-controlled estimate is 1.49x at p = 0.114, and that is the one a sentence about reader preference should cite.
Mantel-Haenszel common risk ratio across display-position bands; Cochran-Mantel-Haenszel test, continuity-corrected. The pooled ratio beside it is NOT position-controlled.
What this cannot tell you
This measures who job-seekers CHOOSE to approach. It does not and cannot measure who replies - the tool observes a profile open, never an answer.
That distinction is worth stating plainly because the usual version of this advice asserts the opposite: that senior people do not answer. We have no data on answers at all. What is measured here is the choice a job seeker makes when both kinds of person are on the same screen.
Two more limits. This is one tool's traffic, so the population is people already looking for a referral rather than job seekers in general. And the classification is our own: every person is labelled from their public headline by the same scorer the product ranks with, so a headline that hides seniority is counted as a peer.
Quotable numbers
Citing this page? Any of these lines can be quoted as is, with a link back here (CC BY 4.0).
Across 242 real referral searches and 6,685 people shown, job seekers opened the profile of an individual contributor or manager 5.38% of the time and a director or executive 1.32% of the time. Measured 24 September 2026.
Holding display position fixed, the estimate is 1.49x in favour of peers at p = 0.114. That is a direction, not an established effect, and this page publishes it as one. It is not a reversal: no band contradicts it and peers lead in 3 of the 4 bands. Measured 24 September 2026 over 180 days.
The pooled gap between peers and senior people is 4.07x, but it is not position-controlled: the tool that produced the data ranks peers higher, so the pooled figure measures the ranking as well as the reader. The position-controlled estimate is 1.49x at p = 0.114.
The clearest single slice: past display rank 20, 6 of 620 peers were opened and 0 of 836 senior people were. Measured 24 September 2026.
This measures who job-seekers CHOOSE to approach. It does not and cannot measure who replies - the tool observes a profile open, never an answer.
Method
Every LinkedIn result page this tool served in the window, bot-filtered. Each person is classified from their headline by the product's own warmReachability(): PEER = IC / manager / senior-staff-principal; SENIOR = leadership (VP, director, head of) plus exec (C-level, founder, chair, owner). "Opened" means the job-seeker opened that person's profile or copied the intro note written for them. Rates are computed INSIDE display-position bands, because this tool ranks peers higher and an unstratified rate would measure the ranking instead of the reader. `ratio` is the POOLED, position-UNCONTROLLED figure and is published for comparison only; `controlled_ratio` is the Mantel-Haenszel common risk ratio across those bands and is the one a claim about reader preference should quote.
Classification source: services/scoring.js warmReachability + isUnreachableExec, fingerprint 05a537c96639004e, confirmed current on 2026-09-24. The classifier has been corrected before, and a dataset measured by a retired one is withdrawn rather than re-rendered: this page refuses to build if the fingerprint is not today's.
Corpus: 242 search result pages, 6,685 people shown, 75 pages carrying at least one opening. Bot-filtered. Snapshot 24 September 2026.
The full dataset is published under CC BY 4.0 at data/seniority-choice.json. Cite it with a link to this page.
What to do with it
If you are asking for a referral, the people most likely to be on your side are one or two rungs ahead of the job you want, not at the top of the company. That is also who a public LinkedIn search buries: see the LinkedIn Reachability Index for how many of the first ten profiles at a major employer are executives.
See who can actually refer you →
Free · No sign-up · See results in ~10 seconds
A plain LinkedIn search buries the people who would actually help you.
It ranks by prominence, so the most-followed, most-messaged names sit on top. FindWarmIntros surfaces the coworkers who share your school or a past employer - the people most likely to reply - and hands you a ready-to-send intro.
鉁嶏笍 Ready-to-send intro“Hi - we both studied at [your school]. I’d love to hear about your path to a company you are targeting before I apply…”
… plus everyone else in your network who can put in a good word.
Free · No sign-up · Results in about 10 seconds
Related: who a LinkedIn search actually shows you · why referrals work (the stats) · how to get a job referral, step by step