Blog

What is support triage

Support triage sets the order of the queue and decides whether a message needs a person at all. The four decisions it makes, and the signals to sort on.

Six requests land overnight: a customer locked out of their account, a billing question, a feature request, a newsletter, a delivery failure from a mail server, and a bug report from your largest account. Sorting them, deciding which one gets an answer first and which one needs a person at all, is support triage.

Four questions make up the sort: is this a real request, how urgent is it, who owns it, and does it need an answer at all. Triage runs before the answering starts and sets the order of the queue. It is also the step most teams skip, because an inbox has a default order whether anyone chose it or not.

How triage differs from tiering

Triage is order. Tiering is ownership. Confusing the two builds the wrong thing: a routing rule where a sort order was needed, or a set of levels where somebody simply had to read the queue.

Triage Tiering
The question Who gets an answer first? Which level of the team handles it?
What it changes The order of one queue The path a request travels
Who needs it A team where more arrives than one person can answer A team with separate front-line and specialist roles
Fails when Nobody reads the queue The front line is not busy enough to route to

Most teams of five need the left column and have no front line to move requests between. What is tiered support covers the levels and when they pay for themselves.

What triage actually decides

Four decisions, and the first two settle whether the message belongs in the queue at all.

Decision The question What it changes
Is this real A person asking for help, or a receipt, a newsletter or a bounce? Whether the queue gains an item
How urgent Is someone blocked now, or can this wait until Thursday? Its place in the order
Who owns it Which part of the product does this touch? Who reads it next
Does it need a reply Can the docs, the status page or a fix answer it? Whether it costs a human a reply

The order is deliberate, because the cheap decisions come first. Discarding a bounce notice takes a second and saves a reply nobody needed to write.

The word comes from medicine, where it means sorting patients by how urgently each one needs care, from the French trier, to sort (Merriam-Webster). The support version keeps the sorting and drops the stakes.

Ticket triage is the same work under the name a help desk uses. What a support ticket is covers the object itself.

The signals a small team can triage on

A queue of a dozen threads does not need a priority field. It needs a few questions that everyone answers the same way.

Signal What it suggests What it leaves out
Someone cannot work at all Ahead of a question or a preference How many other people are blocked
No workaround exists Fix it before the reports that have one Whether the fix is yours to make
The same report arrives twice One issue, not two Who writes in next
Several accounts report it The impact is growing Which account matters most

A service desk computes this from two inputs: impact, meaning how many people the problem affects, and urgency, meaning how long until it matters (UC Berkeley ServiceNow knowledge base). On a team of five both inputs come from reading the message, and the matrix is the part you can skip.

Google’s SRE book uses three questions before it treats a problem as an incident: does it need a second team, is it visible to customers, and is it still unsolved after an hour of concentrated work (Managing Incidents). Those three work on an inbox too, and they give a team a written threshold for the day one request jumps the queue.

Nothing in that list needs a severity scale, a custom field or a routing rule. It needs one person to read the message and choose.

What happens when nobody triages

An inbox defaults to newest first, so the loudest customer wins by accident. The thread that just arrived sits on top, the message from Tuesday keeps sinking, and the customer who wrote once and waited ends up at the bottom. Writing again, copying a colleague, or asking in the team chat is what moves a request up, because each of those adds a newer message.

The quiet customer does not complain. They wait, then they stop expecting an answer, and you find out at renewal. The only trace is the oldest unanswered thread, which is the thread first response time eventually records.

The second failure runs the other way. When the order is unclear, two people open the same thread and start the same reply, and one of them throws the work away. A queue nobody trusts is a queue everyone checks again.

When triage is worth writing down

Formal triage is a full-time job at volume, and the SRE book says so in blunt terms: tickets should be a full-time role for a stretch of time a person can hold, and when the volume is too high for one person the fix is another person (Dealing with Interrupts).

Below that volume the queue is small enough to read. When five people can see every thread, anyone can sort by reading, and a written process adds a status change where a sentence in chat was enough.

What the queue looks like What to do
One person reads every thread before anyone answers Nothing. Reading is the triage
Two people answer the same thread in a week Give each thread an owner, not the team a process
Requests arrive faster than one person can sort them Give the queue to one person for a day or a week
You promised a reply time and you are missing it Write the order down and say who is on it

The last row is the one that forces the issue. A promise to answer within a working day is a commitment about order, and what a support SLA is covers what that promise contains.

The overhead case deserves the same plainness. A team that reads every thread together, fixes the bug the same afternoon and answers in the same conversation has no triage problem. Adding a rotation, a label and a morning queue review to that team produces a status field and no change in the answer.

If the wish list has grown to priorities, reports and routing rules, that is the drift toward a help desk, and shared inbox vs help desk covers which one your team needs.

Where Chirp fits

Chirp is a shared inbox for small software teams that track their work in Linear, so the queue is one list and the order is visible without a meeting. j and k move through it, e archives, and s, a, t and b snooze, assign, tag and block. A saved view on g 1 to g 9 puts the same queue in front of everyone.

Press i in a thread to link it to a Linear issue, create one, or search for the one that already exists. The thread snoozes until the issue is done, then returns with a reply drafted for you that names it. Rules close, tag, assign and snooze threads and write drafts, and no rule can send mail.

The Issues view (g i) sorts linked issues by how many customers are waiting. That is the triage question in one line: an issue moves up when it is holding up more people, rather than when the newest message about it arrived. Chirp costs $9 per user per month.

FAQ

What does support triage mean? It is the step that decides the order of the support queue and whether a message needs a person at all. It comes before answering, and it is what stops the newest message from always being the most important one.

What does triage mean in simple terms? Sorting by who needs attention first. A hospital sorts patients by how urgently each one needs care, and a support team sorts requests the same way with the stakes replaced by blocked work.

Is triage the same as tiering? No. Triage sets the order of the queue, and tiering decides which level of the team handles a request. A team can triage without any levels, which is what most small teams do.

How do you triage without a priority field? Read the message and ask the four questions: is this real, how urgent is it, who owns it, and does it need a reply at all. The signals in the table above are enough for a queue you can see in one screen.

What are the three types of triage? The phrase usually points at medicine, where triage systems sort casualties into a few groups by how urgently each one needs care. Support borrows the word rather than those categories, so the useful split here is the four decisions above.

Does triage mean first in, first out? It means the order is decided rather than defaulted. Oldest first is a defensible order and it answers the customer who waited longest before anyone else. Leave it to the inbox and the newest message wins, which is a default rather than a decision.

Next step

Triage only works when there is one queue for the order to apply to. What a shared inbox is covers the arrangement that gives one address a shared history and a visible owner, which is what makes the sort worth deciding.