Age (hours) of the oldest still-open conversation.
At a glance
Oldest Open Conversation answers one blunt question: what is the longest a customer has been left hanging in your inbox right now? It reports the age, in hours, of the single oldest conversation that is still open. This is the card that catches the thread that fell through the cracks — the one nobody owns, the one that slipped past a snooze, the one the customer has quietly given up on. For Blitz, a forgotten conversation about a wrong-size football boot can become a public complaint; this card surfaces it before that happens. A healthy support desk keeps this number low and stable.
| What it counts | The elapsed hours since the most recent customer message (or created_at if never replied to) on the single oldest conversation where state = open. The value is the maximum age across the open queue, not an average. |
| Sample type | Real-time read of the conversations endpoint, sorted by waiting time; the card returns the top (oldest) record’s age. |
| Why it matters | The oldest open conversation is your worst customer experience right now. It is almost always a sign of a routing gap, an unassigned thread, or a snooze that expired unnoticed. Long tail age, not median, is what produces angry escalations and bad reviews. |
| Reading the value | Lower is better. A number that creeps past your SLA window — and especially past the 3-day alert line — means something has been abandoned. Open it, find out who should own it, and close the loop. |
| Currency | number |
| Time window | RT |
| Alert trigger | >3 days |
| Sentiment key | — |
| Roles | owner, operations |
Calculation
Vortex IQ queries the Intercomconversations search endpoint for all records with state = open and computes, for each, the waiting age — the time elapsed since the customer’s last inbound message, taken from conversation.statistics.last_contact_reply_at, or created_at if the customer has never received a reply. The card reports the single largest age, converted to hours. The underlying record carries the conversation ID, assignee (or unassigned), customer, and the tags, so you can jump straight to the offending thread.
Worked example
A representative reading of Oldest Open Conversation for Blitz. Most of the week the card sits around 6–10 hours, which is normal for an inbox that closes overnight. On Tuesday it jumps to79 hours and trips the 3-day alert. Opening the record, the support lead finds a conversation from Saturday about a backordered cricket bat that was snoozed “until restock” but never re-surfaced because the restock note was added in a different tool. The customer has since emailed twice. The lead replies with a firm delivery date and a discount code, closes the loop, and the card drops back to 8 hours — the next-oldest genuinely active thread. To understand why the snooze failed to re-open the conversation, she opens Vortex Mind; to ask “how many conversations are older than 48 hours right now?” she uses Ask Viq.
Sibling cards merchants should reference together
| Card | Why merchants reach for it |
|---|---|
ic_alert_sla_breach | The oldest conversation is almost always also an SLA breach — these two cards point at the same neglected threads. |
ic_alert_unassigned | Unowned conversations are the most common source of a runaway oldest-open age. |
ic_snoozed_conversations | Expired or mis-set snoozes are a frequent cause — check the snoozed queue when the oldest age spikes. |
ic_open_conversations | The total open backlog gives context: a high oldest-age inside a small backlog is a routing problem, not a capacity one. |
ic_open_backlog_trend | The trend shows whether old conversations are accumulating over time or this is a one-off. |
Reconciling against the vendor’s own dashboard
Where to look in Intercom’s own dashboard: In Intercom, open Inbox, select your open view, and sort by “Waiting longest” (or “Oldest”). The conversation at the top of that sorted list is the same record this card reports, and its waiting time should match the card’s hours value. There is no single “oldest open” number in Reports, so the live Inbox sort is the right place to reconcile. Why the Vortex IQ value may legitimately differ:| Reason | Direction | What to do |
|---|---|---|
| Age basis. Vortex IQ measures hours since the customer’s last message; Intercom’s “waiting longest” sort uses time since last admin action. For a conversation mid-reply these can diverge. | Variable | Compare on a conversation where the customer sent the last message for a clean match. |
Snoozed exclusion. This card counts only state = open; a snoozed conversation that is older is not counted until it re-opens. | Vortex IQ may read lower | Filter the Intercom Inbox to “Open” only, excluding snoozed, before comparing. |
| Time zone. Both systems must use the same reporting time zone for the hours figure to align exactly. | Marginal | Confirm time-zone alignment in the Vortex IQ profile. |
| Refresh lag. A conversation closed seconds ago may still appear in one view before the other refreshes. | Marginal | Force a refresh on both views. |
Known limitations / merchant FAQs
Q: How often does Oldest Open Conversation update? It is real-time — re-evaluated against the live clock on each dashboard refresh. The age ticks up continuously until the conversation is closed or snoozed. Q: Why does my Intercom Inbox show a different oldest conversation? Usually because Intercom sorts by last admin action while Vortex IQ measures from the customer’s last message, or because the Inbox view includes snoozed conversations that this card excludes. Sort the open-only Inbox by waiting time to reconcile. Q: The age looks huge but the customer isn’t actually waiting — why? A conversation can be technically open while waiting on the customer (you replied, they never came back). If that distorts the metric, close or snooze those “awaiting customer” threads so the card reflects genuinely neglected conversations. Q: Can I change the 3-day alert line? Yes. The>3 days threshold is configurable per profile in the Sensitivity tab — tighten it to match a same-day support promise or relax it for low-volume B2B queues.