Support Deflection Estimator
Estimate cost savings from self-service ticket deflection via KB and bots. Enter values for instant results with step-by-step formulas.
Formula
Deflection Rate = (KB_Views - Tickets_From_KB) / KB_Views
We assume that a user viewing a help article intends to solve a problem. If they do not file a ticket within that session, we count it as a 'Deflection'. This calculator estimates the volume of these successful self-serve interactions and multiplies them by your Cost Per Ticket to find total savings.
Worked Examples
Example 1: High Deflection
Problem:15,000 KB Views, 2,000 Total Tickets. Cost $8/ticket.
Solution:Est. Deflected Users = 14,600. Rate = 97.3%. Savings = $116,800.
Result:Massive ROI
Example 2: Low Deflection
Problem:1,000 KB Views, 800 Tickets. Cost $15/ticket.
Solution:Est. Deflected = 840. Rate = 84%. Savings = $12,600.
Result:Needs Better Content
Frequently Asked Questions
What is Ticket Deflection?
Ticket deflection is the process of resolving a customer's issue through self-service resources (like a knowledge base, community forum, or chatbot) before they need to contact a live agent.
What is a good Deflection Rate?
Industry benchmarks vary, but 20-40% is average for new KBs. Mature organizations with AI search can achieve 70%+ deflection rates.
Does Deflection lower NPS?
Not necessarily. If the self-serve content is good, NPS goes up. If it's a 'doom loop' of bad articles preventing contact, NPS goes down.
What is Cost Per Ticket?
Total Support Budget / Total Tickets. Includes agent salaries, software costs, and overhead. Typically ranges from $5 (email) to $25+ (phone).
What is 'Implicit Deflection'?
When a user finds the answer via Google and never even enters your support flow. Hard to measure but very valuable.
Background & Theory
The Deflection Equation
Deflection is the bridge between **Scale** and **Quality**. You cannot hire linearly with user growth.
- Total Demand: All user issues.
- Self-Serve Capacity: Issues solvable by docs/bots.
- Human Capacity: Issues requiring empathy/complex debugging.
Deflection Rate measures how well you filter the "Total Demand" into the "Self-Serve" bucket.
Key Variables
- Findability: Can users find the article? (SEO, Search bar quality).
- Clarity: Is the article understandable? (Reading level, screenshots).
- Completeness: Does it solve the whole problem?
Practical Tips
- Just-in-Time Help: Put tooltips on confusing UI elements. Prevent the question from forming.
- Ticket Tagging: Analyze incoming tickets. If "How to export PDF" is a top tag, write a better article for it.
- Feedback Loops: Let agents flag confusing docs. They are the frontline sensors of documentation failure.
History
The Call Center Era
In the 1980s and 90s, "Deflection" wasn't a concept. The goal was "Call Avoidance"—often achieved by burying phone numbers (a dark pattern). Support was viewed purely as a cost center, and the only metric that mattered was Average Handle Time (AHT).
The Rise of the Knowledge Base
With the internet (Web 1.0), companies put FAQs online. This was the birth of true self-serve. Metrics shifted to "Page Views" of help docs. However, views didn't correlate with solved problems. A user could view 10 pages, get frustrated, and still call.
AI and The Modern Era
Today, Deflection is a sophisticated science. We use "Intelligent Search" and LLMs (Large Language Models) to suggest answers inside the ticket submission form ("Contextual Deflection"). We track user journeys to see exactly where they drop off. Deflection is now a key component of Product-Led Growth (PLG), where the product explains itself.
Common Misconceptions
- Myth: High deflection means we can fire agents. Reality: It means agents can focus on complex, high-value issues instead of password resets.
- Myth: All KB views are deflections. Reality: Many views are from frustrated users who eventually call anyway. You must account for the "leakage."