Why I gate Glassdoor signals behind 5+ reviews
Culture scores without review volume are noise. On FeedbackAI we show ratings when enrichment passes a review-count gate — otherwise we show the gap.
I passed on a role last year because the title looked perfect and the Glassdoor signal was basically missing.
Not “bad reviews” — missing data, dressed up as insight on other platforms. That experience shaped how we ship company context on FeedbackAI.
Ratings without volume lie quietly
A 4.2 star average from three reviews is not the same as 4.2 from three thousand. Yet many products render both identically — pretty badge, false confidence.
Candidates make life decisions on that badge. Recruiters cite “great culture score” in outreach. When the underlying reviewCount is tiny, everyone is optimizing on noise.
Our gate: reviewCount ≥ 5
When Glassdoor enrichment passes our gate, company ratings can appear alongside match scores and salary benchmarks (when salary data exists). When data is missing or below threshold, we show the gap — not a placeholder culture score.
That choice trades marketing gloss for trust. Early users asked for “always show something.” We kept the gate anyway because hiring already has enough theater.
How this fits match-first hiring
Job matching should answer:
- Skill fit — explainable gaps vs the role.
- Company context — only when sourced and gated.
- Comp realism — benchmarks when we have them; silence otherwise.
Glassdoor is one input, not the verdict. India tech has employers with thin public review coverage — especially younger Series A shops. A gate prevents us from hallucinating culture where the market has not spoken yet.
For a product-side write-up, see Glassdoor ratings in job matching.
If you are evaluating tools
Ask vendors:
- What is the minimum review count before displaying scores?
- Do you hide or impute when data is missing?
- Can candidates see which fields are empty?
If answers are fuzzy, assume the UI will optimize for demos — not decisions.
If you are hiring
Thin Glassdoor coverage is not a moral failure. It is an invitation to bring proof elsewhere: engineer blogs, manager LinkedIn posts, structured candidate Q&A in loops.
Candidates notice when you replace missing data with spin. They also notice when you are honest about what they should research themselves.
Takeaway
Trust systems should fail visible, not pretty. Gating Glassdoor signals behind review volume is one small discipline — part of the same ethos as explainable match scores instead of Boolean luck.
When culture data is real, show it. When it is not, say so. Candidates have long memories; platforms should too.