Negative-Rated Conversations, Intercom

A table of the actual 1-2★ Intercom conversations from the last 30 days, with customer, agent, tag, and the rating remark, so you can read the words behind a bad score. How to read it, why it matters, and how to act on it.

Metrics type: Supporting Metrics • Category: Helpdesk / Customer support

The individual conversations rated negatively (1-2★) in the last 30 days, listed with customer, handling agent, topic tag, and the rating remark — the qualitative detail behind the CSAT number.

At a glance

Negative-Rated Conversations is a customer-satisfaction metric tracked from your Intercom workspace. Where CSAT and CSAT Trend give you the number, this card gives you the evidence: a table of the real conversations a customer scored 1 or 2 stars, with the remark they left. For the Blitz support lead this is the most actionable card in the satisfaction set — every row is a named customer, a recoverable relationship, and a clue about what is going wrong. Worked weekly, it turns CSAT from a scoreboard into a queue of follow-ups.

What it counts	Each conversation with a negative conversation_rating (1 or 2 stars) in the last 30 days, one row per conversation. Columns typically include the customer/contact, the handling admin, the topic tag(s), the rating, the free-text remark, and a deep link back into Intercom. The card also surfaces concentration — how many negatives cluster on a single tag or topic.
Sample type	API-derived from the Intercom conversations endpoint. Vortex IQ filters on conversation_rating.rating ≤ 2 and enriches each row with the contact, assigned admin, and conversation tags.
Why it matters	A percentage cannot be coached or recovered — a specific conversation can. Reading the remarks tells you whether bad scores stem from shipping, sizing, refunds, tone, or slow replies, and which customers are worth a personal save. Five negatives on one tag is a process problem, not five unlucky days.
Reading the value	Scan for clustering first: if the alert has fired, several negatives share a tag or topic — start there. Then triage by customer value and recency. Each row is a to-do: reopen, apologise, fix, and where relevant route to the team that owns the underlying issue.
Currency	mixed (table: counts, ratings, and text remarks)
Time window	30D
Alert trigger	>5 on one tag/topic
Sentiment key	—
Roles	owner, operations

Calculation

Vortex IQ selects conversations whose conversation_rating.rating is 1 or 2 within the last 30 days and builds one row per conversation, joining the contact, the handling/assigned admin, and the conversation’s tags. It then groups negatives by tag/topic; when more than five negatives share a single tag, the alert fires, flagging a systemic issue rather than scattered one-offs. There is no positive-share percentage here — this card is the raw, named list, deliberately qualitative so you can read intent and recover customers.

Worked example

A representative reading of Negative-Rated Conversations for Blitz. Over 30 days the table holds 14 negative-rated conversations. Triaged by tag, seven of them carry sizing and remarks like “boots ran a full size small, no help offered” and “had to ask three times for an exchange label.” That clears the >5-on-one-tag threshold and trips the alert. The pattern is not an agent problem — replies were prompt — it is a product-information and returns-policy problem. You do three things: reopen the seven sizing threads and offer free exchanges, brief the team on a standard sizing-and-exchange response, and flag the size-guide gap to the merchandising side of the business. The remaining seven negatives are scattered across tags and read as genuine one-offs, handled individually. To trace whether sizing complaints correlate with specific SKUs, open Vortex Mind; to ask “list this month’s 1-2★ conversations grouped by tag” in plain English, use Ask Viq.

Sibling cards merchants should reference together

Card	Why merchants reach for it
ic_csat_score	The headline number these rows explain — read the score, then read the words behind it.
ic_csat_trend	When the trend bends down, this table is where you find out why.
ic_csat_by_agent	Cross-check whether negatives cluster on one agent or are spread across the team.
ic_top_tags	See whether a topic driving negatives is also a high-volume topic overall.
ic_reopen_rate	Negatives plus reopens point to conversations closed before the customer was satisfied.

Reconciling against the vendor’s own dashboard

Where to look in Intercom’s own dashboard:

In Intercom, open the Inbox and apply a filter on conversation rating set to the negative scores (1-2★), date range last 30 days. That gives you the same list of conversations. The Reports → Customer satisfaction view shows the negative share and lets you click through to the underlying threads and read each remark.

Why the Vortex IQ value may legitimately differ:

Reason	Direction	What to do
Negative threshold. Vortex IQ treats 1-2★ as negative; if your workspace uses a 3-point scale (good/ok/bad) or counts 3★ as negative, row counts will differ.	Variable	Confirm which ratings your workspace classifies as negative.
Tag attribution. A conversation may carry several tags or none; how negatives group depends on tagging discipline, which differs from Intercom’s own grouping.	Variable	Audit tagging consistency before reading the clustering.
Rating-date vs conversation-date. Vortex IQ lists by rating submission date; an Inbox filter may key off conversation created/updated date, shifting a row in or out at the window edge.	Marginal	Align the date basis.
Time zone. Workspace time zone vs merchant reporting time zone can move a borderline row across the 30-day boundary.	Marginal	Confirm time-zone match.

Cross-connector reconciliation: when negatives cluster on delivery, refund, or stock, pivot to the cross-platform cards — Complaints on Out-of-Stock SKUs and Support Spike on Failed Payments — to see whether a commerce or payment problem is generating the bad scores. For divergence investigations, use Vortex Mind.

Known limitations / merchant FAQs

Q: How often does Negative-Rated Conversations update? The card refreshes on the standard data refresh (typically every 30-60 minutes). New negative ratings appear as a fresh row once the customer submits them — which may be hours after the conversation closed.

Q: Why does my Intercom rating filter show more or fewer conversations? Usually the negative threshold (1-2★ vs a 3-point scale), tag-based grouping differences, or the rating-date vs conversation-date basis at the window edge. Match those before assuming a mismatch.

Q: The clustering alert fired but the tags look unrelated — what now? Check tagging discipline. If agents tag inconsistently, the grouping is noisy. Tighten your tag taxonomy (see Top Topics) so concentration is meaningful, then re-read the card.

Q: Can I customise the alert threshold? Yes. The >5-on-one-tag trigger is configurable per profile in the Alert Rules tab. Lower it if Blitz wants to catch emerging issues earlier, or raise it during high-volume sale periods to reduce noise.

---

Tracked live in Vortex IQ Nerve Centre

Negative-Rated Conversations is one of hundreds of KPI pulses Vortex IQ tracks across Intercom and 70+ other ecommerce connectors. Nerve Centre runs the detection layer; Vortex Mind investigates the cause when something moves; Ask Viq lets you interrogate any number in plain English.

Start for free or book a demo to see this metric running on your own data.