Deflection Rate

Definition

Deflection rate is the percentage of inbound support contacts that are resolved by self-service content — a knowledge base, help center, chatbot, or FAQ — before reaching a human agent. It is the primary return-on-investment metric for knowledge base software investment.

Formula: Deflection rate = (Contacts deflected ÷ Total potential contacts) Ã- 100

Where “total potential contacts” = tickets received + estimated self-resolved contacts (the hard part to measure).

What’s a good deflection rate?

Vendor marketing will tell you 60-80%. The real benchmarks from independent sources are significantly more conservative:

Scenario	Realistic range	Source
New KB, month 1-2	8-15%	Supportbench 2026 benchmarks
Established KB, month 6	15-30%	Corebee / supp.support 2026
Mature KB with AI, year 2	30-45%	Supportbench advanced tier benchmarks
Best-in-class (B2B SaaS)	40-55%	Editorial investment + AI + optimisation
Marketing-page “up to”	60-80%	Cherry-picked B2C FAQ volumes

The 60-80% numbers are real but represent trivial use cases: high-volume B2C consumer accounts where the top 5 FAQ answers (password reset, shipping, return policy) cover 80% of contacts. In B2B SaaS, the query distribution is much flatter — there’s no single dominant question, and each answer requires more nuance.

Why vendors inflate their numbers

Vendors have an incentive to report the highest defensible number. Common inflation techniques:

1. Session-based denominator: “Of all users who visited our KB, 65% didn’t submit a ticket.” This includes users who visited and found no answer — they still count as “deflected.”
2. Cherry-picked time period: Report deflection rate during a campaign month when a major FAQ was just published.
3. Channel isolation: Only count deflection from Beacon/chatbot, not from Google-arrived organic visits that ended without a ticket.
4. B2C benchmarks in B2B contexts: Consumer use cases have 10-20x more homogeneous query distributions.

The honest vendor benchmark question: “What’s the median deflection rate for a B2B SaaS customer in their 6th month, using your full feature set, with a dedicated content owner?” Almost no vendor publishes this number.

How to measure it accurately

Step 1: Establish your baseline. Track tickets received per month before launching any KB or self-service content.

Step 2: Instrument your help center. Use a tool that tracks:

• Article views from the contact form pathway (did they read an article before submitting a ticket?)
• Beacon interactions that ended without ticket submission
• AI chat sessions that ended without human escalation

Step 3: Estimate counterfactual contacts. This is the hard part. A proxy: “What would our ticket volume be if it had grown at the same rate as our ARR?” The gap between actual tickets and the counterfactual estimate is your deflection.

Step 4: Report monthly and quarterly. Deflection compounds — month-6 deflection is usually 2-3x month-1 deflection as the content library fills gaps.

Tools ranked by deflection measurement quality

Tool	Deflection measurement	Honest about median?
Document360 Business	Search-failure + Beacon correlation	Partially — dashboard is clear, marketing inflates
Help Scout Plus	Beacon click → no ticket tracking	Yes — reports actual Beacon deflection cleanly
Zendesk Guide Suite Pro	Detailed analytics; agent-KB correlation	Partially — analytics strong, marketing inflates
Guru	Internal query deflection (not customer)	Often misrepresented in marketing
Notion	No deflection tracking	N/A

Related terms

• Ticket deflection — the absolute count of contacts avoided
• Self-service rate — broader metric including phone/email/chat avoidance
• Search-failure rate — most actionable metric for finding KB content gaps
• First contact resolution — adjacent CSAT metric; KB quality feeds this