Find Roles Worth Applying To
Find Roles Worth Applying To
Reputation, explained
Every company's score comes from real activity on this site — what happened after people applied, what they reported back, and how the listings themselves look. There is no manual grading and no per-company overrides. This page reads the exact same rules the score is computed from, so if those rules change, this page changes with them.
Strongest signal
What a company actually does once someone applies moves the score more than anything else. A withdrawn application has no effect either way.
Offer
The strongest positive signal there is.
Interview
The company moved the candidate into a real process.
Timely rejection
A real, closed-loop answer, even if it's a no.
Rejected after an interview
A negative signal, since the candidate already invested real time.
Likely ghosted (no update for 21+ days)
An application that's gone quiet long enough is treated as a negative outcome, not a neutral one.
Direct reports
Ratings and reports on a listing feed into the same company score, weighted a little lighter than a real application outcome.
1–2 star rating
A low rating on a listing.
4–5 star rating
A high rating on a listing.
Reported as expired
Someone flagged the listing as no longer open.
Reported as inaccurate
Someone flagged the listing's details as wrong or misleading.
Reported as spam or duplicate
Someone flagged the listing as spam or a repost.
Lightweight signal
Listing quality is a much smaller, secondary signal. It only kicks in once a company has at least 3 active listings, and only when a real share of them show the same pattern — one polished or one messy listing out of many doesn't move anything.
Pay details shown
Applies once at least 50% of a company's active listings include usable pay information.
Work mode is clear
Applies once at least 50% of active listings clearly say remote, hybrid, or onsite.
Long-running listings (open 90+ days, about 3 months)
Applies once at least 30% of active listings have been open that long.
Same listing reposted
Each extra copy of the same title from the same company beyond the first, capped at 3 extra copies counted.
Combined, these listing-quality signals can move a company's score by at most ±15 points — real outcomes and direct reports are what actually make a score Good or Bad, never listing housekeeping on its own.
How much to trust it
Confidence describes how much real evidence backs a score — not whether the company is good or bad. A company can have a strongly positive score built on very little evidence, which is exactly what Low confidence is for.
High confidence
At least 4 weighted application outcomes.
Medium confidence
At least 1 application outcome, or at least 2 feedback reports with no application outcome yet.
Low confidence
No evidence yet, or only listing-quality heuristics.
That ± number is the confidence interval shown on a company's page — the range the real score could plausibly sit in given how thin or solid the evidence is.
Evidence ages
A signal doesn't disappear as it ages, but it fades. An outcome or report loses value gradually over about 18 months, settling at a floor of 25% of its original weight rather than dropping to zero — so a company's history is deprioritized over time instead of erased. A single offer from years ago counts for less than one from last week, but it's never wiped from the picture entirely.
One more thing
Until a company has built up a few real signals, its displayed score includes a small, fixed cosmetic nudge so the site doesn't show a wall of identical “Moderate” scores before any real evidence exists. That nudge is deterministic (the same company always gets the same nudge) and shrinks to nothing well before real evidence would otherwise take over. It never changes what actually happened at that company — only the very first impression before there's anything real to show.