Across 52,407 supplier emails at 12 mid-market CPG companies, five supplier-reliability signals emerge — ghost rate, bandwidth tax, price stickiness, disclosure asymmetry, and bid conversion — that ERPs cannot see because they never touch the email layer. Supplier reliability is a leading indicator hidden in communication data. ERPs only surface it once it’s already a trailing indicator in the OTIF report.
Methodology
- 12 customer companies in the $50M–$500M revenue range
- 1,291 suppliers across ingredients, packaging, co-manufacturing, and services
- 52,407 buyer-supplier communications containing 78,247 parsed messages
- 4,916 threads analyzed end-to-end
- 4,942 RFP bids tracked
- 526 supplier-SKU pairs with three or more comparable quotes
Analysis restricted to suppliers with five or more threads or bids. Price analysis restricted to identical units and currency. Identities stripped or archetype-labeled. Verbatim quotations preserve original language. Study period: Q4 2025 through Q1 2026.
Signal 1 — Ghost rate
45.9% of the 4,916 threads ended with the supplier ghosting for 14+ days. At the supplier level, the median vendor ghosts 52% of threads. Only 3% of suppliers never leave threads unanswered; 21% of suppliers ghost 75%+ of interactions.
A ghost is more informative than a “no.” It means the supplier has low enough internal priority on your account that no one owns closing the loop. Chronic ghosting is a structural capacity signal, not a one-off miss.
Signal 2 — Bandwidth tax
1.2% of bids required 25+ emails to resolve. One bid consumed 212 emails. 8.6% of active suppliers average more than eight emails per bid. Most quotes resolve efficiently — 46% in one to three emails, 39% digitally.
The long tail is the cost. A supplier that reliably delivers but takes 25 emails per bid silently consumes procurement capacity that should be going into competitive sourcing.
Signal 3 — Price stickiness
Among 526 supplier-SKU pairs with repeat quotes:
- 49% swung at least 5% between quotes.
- 25% moved 25%+.
- 11% more than doubled.
Suppliers used “subject to change” 335 times in dataset communications — contractual camouflage for non-committal pricing. High stickiness variance combined with hedging language is a reliability signal, not just a negotiating tactic.
Signal 4 — Disclosure asymmetry
Suppliers initiate messaging on unfavorable topics far less often than buyers do:
| Topic | Share initiated by supplier |
|---|---|
| Tariffs | 91% |
| Quote expiration | 84% |
| Stock-outs | 78% |
| Logistics delays | 73% |
Buyers initiate accountability conversations:
| Topic | Share initiated by buyer |
|---|---|
| Volume commitments | 58% |
| Negotiation | 64% |
| Pricing | 69% |
A supplier that proactively raises problems is structurally different from one that waits to be asked. That difference is invisible to an ERP but obvious in email.
Signal 5 — Bid conversion
Only 32.5% of formal RFP invitations end with the supplier submitting a price. Two-thirds of invitations yield no quote — even when the supplier engaged (asked questions, requested specs, promised follow-up). An invitation isn’t a bid. A pipeline counted in invitations overstates real competitive coverage by roughly 3x.
Five supplier archetypes
- The Vanishing Vendor — early engagement followed by disappearance between sample and PO.
- The Bandwidth Tax — eventually delivers, but costs excessive email threads along the way.
- The Silent Inviter — accepts RFPs without submitting quotes; the ~two-thirds pattern.
- The Weather Vane — inconsistent pricing with contractual camouflage language.
- The Reliable Voice — low ghosting, efficient quotes, stable pricing, proactive problem-flagging. Under 5% of suppliers.
Why ERPs can’t see any of this
Email was always the system of record. LLMs finally made it readable. Traditional ERP systems capture trailing indicators — OTIF, quality defects, aging receivables. Email contains leading indicators of supplier reliability: how often they reply, how many rounds a bid takes, whether they proactively flag stock-outs, how their pricing moves quote-over-quote.
Until recently, structuring this data at scale wasn’t possible. It is now.
What to do with this benchmark
Rank your supplier list by ghost rate first. Then by bandwidth tax. The lists rarely match the supplier list you’d build from ERP data alone — and that mismatch is where the risk you haven’t measured lives.
Source: Waystation AI analysis of 52,407 buyer-supplier communications at 12 mid-market food, beverage, supplement, and pet food companies ($50M–$500M revenue range), Q4 2025–Q1 2026. The full dataset is published at waystationai.com.
