QA Email and Chat Scorecard

For QA Specialists, QA Reviewers, Campaign Leads and anyone who writes to customers. After reading this page you will be able to score a written reply (email, ticket or text) and a chat on the Narqix scorecards, and check your own message before you send it.

Track: All tracks · Role: QA Specialist, QA Reviewer, Support Agent, Back Office · Time to complete: 20 min

Where this comes from

The criteria and pass marks are set on Quality Standard and QA Scoring in Narqix Standards. The Customer Support track uses the same weights with support wording on the Support QA Scorecard page. Both names are shown below, so the sheets match whichever page your team trains on.

Item	Rule
Written replies (email, tickets, texts)	Pass 88%
Complaints and sensitive replies (any channel)	Pass 90%
Chat	Pass 85%
New team member	100% of the first 20 customer-facing messages reviewed, often before sending
Disputes	Within 3 working days, to your Campaign Lead

Written reply scorecard (pass 88%)

Criterion (Standards name)	Support QA Scorecard name	Weight	Full marks	Half marks	Zero
Every question answered; issue resolved or next step owned	Complete resolution	25	Every question answered; solved or clearly owned next step	One minor question missed, or next step vague	Main issue not addressed
Accuracy	Accuracy	20	Policy, prices, dates and order or file data all correct and checked	One minor detail wrong, no customer impact	Wrong information that affects the customer
Tone and personal touch	Empathy and personal touch	15	Personal first line proves the message was read; empathy sized to the problem	Generic but polite	Cold, robotic or bare saved reply
Structure and clarity	Clarity and structure	10	Answer first; short paragraphs; numbered steps	Answer buried or long paragraphs	Hard to understand
Grammar, spelling, correct names, no unfilled placeholders	Writing mechanics	10	No errors; name right; no blank fields	1 to 2 small errors	Wrong name, blank field or many errors
Next step: who does what, by when	Ownership of next steps	10	Who, what and when, with time zone	Next step without a time	No next step when one is needed
Process: tags, status, notes, verification	Process and documentation	10	Correct status, tags, priority and internal note	One process item missing	Wrong status or no note on a handover

Chat scorecard (pass 85%)

Criterion	Weight	Full marks
Greeting within 30 seconds, using their name and issue	10	Fast, specific greeting
Pace: no silent gaps, check-in times stated and kept	15	No gap longer than promised; holding messages used
Resolution	25	Solved, or an owned next step
Accuracy	15	All information correct
Short-form tone	15	Warm, clear, one idea per message
Saved replies personalised	10	Saved replies adapted to the conversation
Closing with outcome and next step	10	Outcome summarised, next step stated, "anything else" offered

Auto-fails (the item scores zero)

Data sent to the wrong person, or pasted into an unapproved tool
Required verification skipped before changing account details
False or misleading information (for example, saying a refund was sent when it was not)
Rude, sarcastic, dismissive or discriminatory language
Refund, credit or promise above authority without approval
A promise the writer knew would not be kept
Payment card data typed, stored or repeated
Compliance step missed (for example, a do-not-contact request ignored)

The scoring sheet

WRITTEN QA · (CAMPAIGN) · reviewed by (REVIEWER NAME) · (DATE)
Writer: (WRITER NAME)    Channel: (EMAIL, TICKET, TEXT OR CHAT)
Ticket or conversation ID: (ID)    Sent: (DATE, TIME AND ZONE)
Complaint or sensitive: (YES OR NO)    Pass mark used: (85, 88 OR 90)

AUTO-FAIL CHECK: (NONE, or WHICH ONE, WITH QUOTE)

SCORES (full, half or zero)
(CRITERION 1) ((WEIGHT)): (SCORE)    Evidence: "(QUOTE OR FACT)"
(CRITERION 2) ((WEIGHT)): (SCORE)    Evidence: "(QUOTE OR FACT)"
(CRITERION 3) ((WEIGHT)): (SCORE)    Evidence: "(QUOTE OR FACT)"
(CRITERION 4) ((WEIGHT)): (SCORE)    Evidence: "(QUOTE OR FACT)"
(CRITERION 5) ((WEIGHT)): (SCORE)    Evidence: "(QUOTE OR FACT)"
(CRITERION 6) ((WEIGHT)): (SCORE)    Evidence: "(QUOTE OR FACT)"
(CRITERION 7) ((WEIGHT)): (SCORE)    Evidence: "(QUOTE OR FACT)"

TOTAL: (SCORE) out of 100    Result: (PASS OR BELOW PASS)

KEEP DOING: (ONE SPECIFIC STRENGTH, QUOTED)
FIX FIRST: (ONE SPECIFIC CHANGE)
Better version: "(THE SENTENCE OR PARAGRAPH AS IT SHOULD HAVE BEEN)"
Coaching session: (NOT NEEDED, or BOOKED FOR (DATE))

Worked example: support email

The customer asked two things: where the order is, and whether they can change the color. The agent replied: "Your order is in transit and will arrive soon. Thanks for your patience!"

WRITTEN QA · (CAMPAIGN) · reviewed by (REVIEWER NAME) · Oct 21, 2026
Writer: (WRITER NAME)    Channel: EMAIL
Ticket or conversation ID: (ID)    Sent: Oct 20, 2026, 3:05 PM ET
Complaint or sensitive: NO    Pass mark used: 88

AUTO-FAIL CHECK: NONE

SCORES (full, half or zero)
Complete resolution (25): 0    Evidence: color change question not answered; "soon" is not a resolution
Accuracy (20): 10    Evidence: tracking shows a delivery exception; "in transit" is incomplete
Empathy and personal touch (15): 7.5    Evidence: polite but generic
Clarity and structure (10): 5    Evidence: short, but no real answer
Ownership of next steps (10): 0    Evidence: no owner or time
Writing mechanics (10): 10    Evidence: no errors
Process and documentation (10): 5    Evidence: tagged, but no note about the delivery exception

TOTAL: 37.5 out of 100    Result: BELOW PASS

KEEP DOING: clean, error-free writing
FIX FIRST: answer every question, using the full tracking detail
Better version: "Your order hit a delivery exception on Oct 19 (address could not be verified). I have asked the courier to re-attempt and will email you the new date by 5:00 PM ET today. On the color: yes, we can switch it before it ships again. Reply 'switch' and I will update it."
Coaching session: BOOKED FOR Oct 22, 2026

Back office task scorecard (pass 90%)

Criterion	Weight
Accuracy against the source	35
Completeness	20
Timeliness	15
Process (SOP followed, naming, logging)	15
Communication (update sent, questions batched)	15

Below 80% means coaching.

Score yourself before you send

Run the five-point self-check from the Quality Standard on every message:

Request: does it answer exactly what was asked, every question?
Source: did I check the facts against the order, file or policy?
Place: right recipient, right ticket, right status?
Record: note, tags and status updated?
Customer test: would I be comfortable if the client read this?

Coach's note: The most common written deduction is a missed second question. Before you send, number the customer's questions on a scrap note and tick each one in your reply. It takes ten seconds and protects 25 points.

Never do this: Never send a message with an unfilled placeholder or the wrong customer name. Never paste customer data into an unapproved tool to "improve" a reply. Never promise a date the tracking or the policy does not support.

Check yourself

What are the pass marks for written replies, complaints and chat?
Which criterion carries the most weight on both scorecards?
A reply is perfect but tells the customer a refund was sent when it was not. What is the score?
List the five-point self-check.

Answers

Written replies 88%, complaints and sensitive replies 90%, chat 85%.
Resolution (every question answered, or an owned next step): 25 points.
Zero. False information is an auto-fail.
Request, source, place, record, customer test.