The Two Review Thresholds That Control AI Citation Eligibility

Earlier this year we documented the star rating quality gate: businesses below 4.0 stars are excluded from ChatGPT's recommendation pool before anything else is evaluated. That post explained an invisible filter most fix plans weren't checking first.

What it didn't fully address: review count operates as a second, independent threshold. Not as a modifier to the star rating filter -- as a parallel gate. A business at 4.7 stars with 11 reviews can pass the quality check and still be effectively invisible in AI-generated recommendations. The two thresholds are not the same problem.

How the two thresholds work

Our 2026-07-06 methodology rec (session 69, `methodology-recs/2026-07-06-review-volume-threshold.md`) formalized a gap in our fix plans: current audits catch the quality threshold (4.0 star floor) but do not specify a volume threshold. The data indicates both operate independently.

The quality threshold is a binary exclusion. Below 4.0 stars on ChatGPT and you don't enter the recommendation candidate pool at all. This is documented in the SOCi 2026 Local Visibility Index across 350,000 locations.

The volume threshold is a gradient with a hard floor. The July 6 rec synthesizes the available data:

- Brands with 80 or more reviews: cited in over 75% of AI-generated answers - Businesses with zero reviews: cited roughly 1% of the time

Source is Duda's 2026 local AEO statistics synthesis. It's a directional finding -- the methodology behind the specific percentages hasn't been independently confirmed. But the direction matches how AI citation averaging works: volume signal weight accumulates. A business with 8 reviews has almost no signal weight even at 4.6 stars. A business with 95 reviews has substantial signal weight even at 4.2 stars.

The July 6 rec is explicit: "A business at 4.6 stars with 8 reviews may pass the quality check but fail the volume check that drives AI citation eligibility." This was a gap our fix plans weren't addressing before this week.

Why fix plans miss the volume problem

The star quality threshold is easy to operationalize: check the current Google average. Below 4.0 means the fix plan starts with review improvement before anything else.

Volume is harder to specify because there's no single confirmed cutoff. The 80-review figure is the practitioner-derived target; our July 6 rec sets a lower "flag" threshold at 20 reviews. The recommended language from that rec:

"With fewer than 20 reviews, your business may pass the quality check but fail to accumulate enough signal weight to compete with businesses that have 80-150+ reviews. Priority action before all infrastructure work: implement a systematic review request process."

The logic behind the 20-review flag isn't that 20 is a meaningful threshold -- it's that below 20 reviews, the gap to the 80-review target is significant enough that infrastructure work is unlikely to produce material citation movement first. Schema and directory optimization return significantly less when volume signal is this weak.

This produces a sequencing rule: quality gate runs first (do you clear the 4.0 floor?), volume gate runs second (do you have enough reviews to compete?). Infrastructure work -- schema, directory presence, NAP cleanup -- runs after both gates are clear.

What the home services data shows

For independent contractors, the volume threshold lands at a particularly difficult spot.

Our 2026-07-06 update to `knowledge/home-services-ai-citation-data.md` (session 69) documents that 87% of independent HVAC and plumbing contractors have effectively zero AI citation share in their own metro. The 5W PR HVAC/Plumbing AI Visibility Index -- 65+ prompts across ChatGPT, Claude, Perplexity, and Google AI Overviews, January-March 2026 -- found a 1.2% ChatGPT citation rate for local contractor locations.

Those contractors are failing on both gates simultaneously. Practitioner data from the same session identified the minimum viable presence for plumbing AI citation: 80+ reviews at 4.6+ average. Volume and quality combine at the competitive end: not just cleared, but above the neighborhood competition.

The 4.6 average figure is notably higher than the ChatGPT exclusion floor (4.0). This gap -- between "not excluded" and "competitive in a real market" -- is where most of the 87% live. They're at 4.1 stars with 22 reviews, and AI engines see them as low-signal relative to businesses with 100+ reviews at 4.5+.

National chains dominate this landscape by volume alone. Roto-Rooter, ARS Rescue Rooter, and Mr. Rooter together account for roughly 19% of all consumer-intent AI citations in HVAC and plumbing -- not because their service pages are better-structured than a local contractor's, but because citation volume compounds. Early AI citation share creates training signal that compounds. An independent contractor building review volume now is doing the only thing that creates a durable foothold against that pattern.

The gradient between 20 and 80

The July 6 rec draws a practical gradient:

Under 20 reviews: signal weight too weak to compete even with correct infrastructure. Flag this as a prerequisite before other fixes.

20-79 reviews: eligible but underperforming. Infrastructure fixes can produce some movement, but the return is lower than it would be at 80+. Running a review request process in parallel with schema and directory work makes sense here.

80+ reviews: volume signal is strong enough that infrastructure quality determines position within the candidate pool. This is where the rest of the fix plan starts to perform as modeled.

The practical consequence: a business at 4.3 stars with 30 reviews should be building review volume at the same time as schema and directory work -- not after. Volume and infrastructure can run in parallel once the quality gate is cleared. But volume can't be skipped under the assumption that better schema will compensate.

What to check before anything else

Before any AI visibility work: check star average and review count across your primary citation platforms -- Google, Yelp, BBB at minimum.

Star average below 4.0 on any of these: quality is Priority 0. Nothing else produces citation movement before that threshold is cleared.

Star average above 4.0, review count below 20: volume is Priority 0. Same conclusion, different mechanism.

Both gates clear -- 4.0+ stars, 20+ reviews: infrastructure work is the right lever, and you're in the range where schema, directory presence, and entity signals produce measurable improvement.

If you're not sure which gate you're failing, Signal Check at sourcepull.ca shows per-platform citation rates across ChatGPT, Perplexity, Gemini, and Claude. When star average looks healthy but citations remain low, the audit identifies which specific infrastructure gaps are the remaining blockers -- and now flags volume against the 80-review target as part of the pre-flight check.