Why Your Average CSAT Score Can Hide the Problem
Average CSAT hides response rate, sample size, distribution, segment differences, volatility, and unrated conversations. Learn what to show beside it.
— thoughts and learnings —
Posts by Augist McCarthy
Customer Support Researcher. Augist studies how support teams run queues, escalations, and quality reviews at Supportman, and turns what she finds into practical playbooks for Field Notes.
Seven Metrics Every Customer Support Dashboard Should Show
Build a balanced support dashboard with volume, responsiveness, resolution, surveyed CSAT, response rate, satisfaction coverage, QA, and DSAT.
How AI Can Help Managers Coach Every Support Agent Without Generic Advice
Use AI to find repeated conversation patterns and prepare individualized coaching—while managers retain context, judgment, and responsibility.
Why Some Agent Feedback Backfires
Feedback can reduce performance when it targets identity, lacks evidence, or offers no next step. Learn how to coach support agents from task-level evidence.
Team Report or Agent Report? Why Support Leaders Need Both
Team reports diagnose systems. Agent reports support individual development. Learn how to use both without turning support metrics into an unfair ranking.
DSAT Is an Operations Signal, Not Just an Agent Score
Learn how to classify DSAT by agent, product, policy, process, tooling, and customer factors—then turn negative ratings into measurable fixes.
Why Positive CSAT Belongs in Your Support Training Program
Customer praise can do more than lift morale. Learn how to turn positive CSAT comments into behavioural examples for onboarding, QA, and coaching.
The Service Recovery Paradox Is Not a Strategy
Can a recovered service failure create more loyalty than a flawless experience? Sometimes—but the effect is conditional and never a reason to tolerate preventable failure.
AI Quality Assurance for Support: How to Calibrate the Judge
Automated support QA needs a custom rubric, human calibration set, attribute-level agreement checks, confidence thresholds, and an appeal process.
How to Validate Predicted CSAT Before You Use It
A practical guide to calibrating AI-predicted CSAT using customer ratings, error analysis, subgroup checks, confidence thresholds, and human review.
The Five Dimensions of Customer Support Quality
CSAT tells you how respondents felt. A support QA rubric explains why. Learn the five dimensions that make conversation quality measurable.
What Is 100% CSAT Coverage?
Learn how surveyed ratings and AI-predicted CSAT work together to create a satisfaction signal for every eligible support conversation.
Your CSAT Score Is Missing Most of Your Customers
Most customers never complete a CSAT survey. Learn how nonresponse and positivity bias affect support reporting—and how predicted CSAT fills the gap.