By source.type (email, chat, messenger, etc.). Where your customers actually reach you.
At a glance
Volume by Channel is a Conversation Intelligence metric tracked from Intercom data. It splits the last 30 days of conversations by the channel they arrived on — email, in-app Messenger, live chat, social, and so on — and renders the mix as a donut. For the support lead and founder at Blitz it answers “where is demand coming from?”, which drives staffing, automation, and where to invest in self-serve. A shifting mix often precedes a shift in cost-to-serve: chat is fast and cheap to deflect, email is slower and heavier.
| What it counts | Conversations created in the trailing 30 days, grouped by conversation.source.type (for example email, conversation / Messenger, chat, push, facebook, twitter). Each conversation counts once, in the channel it originated on. |
| Sample type | Backend API data from Intercom (conversations grouped by source.type), refreshed on the standard data refresh. |
| Why it matters | Channel mix determines your cost and speed of service. A growing email share usually means slower resolutions and higher handling cost; a growing Messenger/chat share means faster, more deflectable contacts. Knowing the split lets you put bots and macros where the volume is and staff the channels customers actually use. |
| Reading the value | Read proportions and movement. The donut shows today’s mix; compare the dominant slice to last period. A channel suddenly swelling — say email jumping from 30% to 55% — is a signal: an incident is pushing customers to a heavier channel, or a chat widget broke. Cross-reference Conversation Volume Trend to see whether the channel grew or just took a larger share of a flat total. |
| Currency | number |
| Time window | 30D |
| Alert trigger | — |
| Sentiment key | null |
| Roles | owner, operations |
Calculation
Calculated automatically from your Intercom data. Vortex IQ pulls every conversation withcreated_at in the trailing 30 days and groups the count by conversation.source.type, the field Intercom sets to record where the conversation originated. Each conversation is attributed to a single channel — its origin — even if later replies come through a different surface. The donut renders each channel’s share of the 30-day total. See the worked example below for a typical reading.
Worked example
A representative reading of Volume by Channel for Blitz on Intercom. In a normal month the donut shows roughly 55% Messenger (shoppers messaging from the storefront), 30% email, 10% live chat, and 5% social. The support lead is comfortable: Messenger is the heaviest slice, so that is where the saved replies and the order-status bot live. Mid-month the mix shifts — email climbs to 50% and Messenger drops. Reading it alongside Conversation Volume Trend (which is flat) tells the founder the total did not grow; customers simply moved to email. Investigating, they find the storefront Messenger widget failed to load on mobile after a theme update. They fix the widget, and over the next week the donut returns to its Messenger-led shape. For deeper investigation, use Vortex Mind to correlate a channel shift with a site change; for natural-language exploration, ask Ask Viq “why did email volume jump this week?”.Sibling cards merchants should reference together
| Card | Why merchants reach for it |
|---|---|
ic_volume_trend | Confirms whether a channel grew in absolute terms or just took a bigger share of a flat total. |
ic_top_tags | Pairs the “where” with the “what” — which topics dominate each channel. |
ic_median_first_response | Heavier channels (email) usually carry slower first responses — read them together. |
ic_volume_by_hour | Adds the “when” so you can staff each channel for its peak. |
ic_avg_parts | Shows which channels generate longer back-and-forths and higher handling cost. |
Reconciling against the vendor’s own dashboard
Where to look in Intercom’s own dashboard: In Reports → Conversations, apply a breakdown or filter by “Channel” (Intercom groups on the samesource.type family) over a matching 30-day range. Intercom’s channel labels (Email, Messenger, Phone, Social) map onto the raw source.type values Vortex IQ groups by, so the proportions should align. The Inbox channel filters show live state, not 30-day created volume, and will not match.
Why the Vortex IQ value may legitimately differ:
| Reason | Direction | What to do |
|---|---|---|
Channel label grouping. Intercom rolls several raw source.type values into friendly buckets; Vortex IQ may keep them split (e.g. push separate from Messenger). | Variable | Map the raw types to Intercom’s buckets before comparing slice sizes. |
| Period boundary. Rolling 30 days vs a fixed Reports range changes which conversations are in scope. | Variable | Set Intercom to the identical custom range. |
| Bot-originated conversations. Conversations opened by a bot or proactive Series may be attributed differently or filtered out per your profile. | Variable | Match Intercom’s bot-inclusion setting to your profile. |
Known limitations / merchant FAQs
Q: How often does Volume by Channel update? The card refreshes on the standard data refresh (typically every 30-60 minutes for live integrations). Q: Why does my Intercom dashboard show a different number? The usual causes are channel-label grouping (rawsource.type vs Intercom’s friendly buckets), the rolling-vs-fixed period boundary, and whether bot-originated conversations are counted. Reconcile those in Reports → Conversations.
Q: Which channel does a conversation belong to if the customer switches mid-thread?
The channel of origin — source.type at creation. Later replies through other surfaces do not re-attribute it.
Q: Can I customise the alert threshold?
This card is a read-only mix with no threshold. For anomaly detection on total volume use Volume Spike; thresholds there are configurable per profile in the Sensitivity tab.