Try Taskora free until November 15. No credit card. After that, a free 90-day plan to make sure it fits.

Black Friday at Taskora: sign up free and lock in up to 70% off at launch, plus a free 90-day plan to try it out. Get the deal

Kanban for Support Teams: Build the Board from the Ticket Lifecycle

15 min readproductkanban

Written by Georgii Garanin, Founder, Taskora

Kanban for Support Teams: Build the Board from the Ticket Lifecycle

A support day is a trickle, and it never stops. A chat ping at 9:04, a ticket at 9:15, a walk-up before you've taken your coat off, and by lunch nobody can say what is actually being worked on or what is quietly waiting on a customer reply. Every arrival re-ranks the priority list, and the list lives in four chat threads, two inboxes, and somebody's head.

Kanban for support teams means running the desk on one board whose columns follow the real ticket lifecycle: new intake, triaged, in progress, customer-waiting, verification, done. Every working column gets a cap, waiting on the customer gets its own visible state, and every open request shows in one place. It fits support because support work arrives as a steady drip of ad-hoc requests, and a kanban board is a flow board, so it absorbs that drip the way a sprint board cannot.

If the method itself is new to you, the plain-language guide to what a kanban board is covers the board itself, pulling work only when a slot frees up, and the limits that cap how much is in progress at once. This article stays on one question: what the board looks like when the queue is a service desk.

Why support work breaks sprint boards

A sprint board assumes the work comes in batches. You plan two weeks of committed work, you protect the plan, and the boundary holds until the sprint ends. Support has none of that. Requests arrive one at a time, all day, with no respect for the plan or the boundary. The ticket that was low priority at 9:00 is the whole afternoon at 11:00, and the team re-ranks hourly whether anyone writes the change down or not.

Run a sprint board against that reality and one of three things happens. The plan becomes fiction by day two. Urgent work starts landing off the board, in chat threads, which is the situation you came here to fix. Or the team quietly runs a flow system anyway, one sprint-shaped board with nothing committed, and the ceremonies turn into meetings about a plan nobody kept.

Kanban makes no promise a support desk cannot keep. Work enters when a slot frees up, it ships when it is done, and the board never resets to honor a plan. The full framework comparison, including scrumban, which keeps one planning day on top of the flow, is in kanban vs scrum vs scrumban. The same reasoning makes kanban the default for service teams of every kind: internal IT support, ops, finance requests, people ops. Any team whose work arrives uninvited is already running on flow, whether or not it uses the word.

Build the board from the ticket lifecycle

The textbook column set, Backlog, In progress, Done, describes a project, and a service desk runs a lifecycle. So write down what a ticket actually passes through from arrival to closed, and give each real state a column. Here's one example support board, sized for a three-person desk. The caps are starting points, and the paragraph after the table shows where each number comes from.

Column What enters it Suggested cap A card leaves when
New intake Any request that has not had a human look yet 10 Someone triages it: names the problem, sets a priority, confirms it is real work
Triaged Understood requests waiting for a free owner 6 Someone with a free slot pulls it and starts the fix
In progress One person is actively working the ticket 3 The fix or answer goes out to the customer, and the ball leaves our side
Customer-waiting The customer holds the ball: more detail, approval, or a test 6 The customer replies, or the aging sweep closes the card with an invitation to reopen
Verification The fix is delivered and awaiting confirmation 3 The customer confirms it worked, or a still-broken reply sends it back to In progress
Done Confirmed and closed tickets No cap Nothing. Done is the record, and it grows

Where the caps come from. In progress gets one active ticket per person, so 3. Verification gets the same, and its cap stops delivered-but-unconfirmed fixes from piling up. Customer-waiting gets two per person, 6, because waiting costs nobody's hands and a tighter cap would have the desk nudging customers all day. Triaged gets two per person as well, a menu of ready work rather than a dumping ground. New intake gets 10 for a reason all its own, and the next section is about it.

The column that changes everything is Customer-waiting. Waiting on a customer is a real state of the work, and hiding it inside In progress hides the clock with it. A card that says In progress while it has really been waiting two days for a file makes the board claim hands are on it, and the SLA clock, the promise the desk made about how fast it responds, runs the whole time. Give waiting its own column and the lie has nowhere to sit. Verification earns its place the same way: a fix that has gone out but has not been confirmed is unfinished work, and unfinished work needs a home with a name.

In Taskora, these six columns are the custom statuses on one stream, which is a project with its own board. New intake is the start status, Done is the final status, and the allowed transitions between them are drawn on purpose. A ticket cannot slip past verification, and the only way from intake straight to done is a drawn move too, the reject, kept for arrivals that turn out not to be desk work. The rework loop matters here too: Verification back to In progress for the still-broken reply. The custom kanban status workflow guide walks the whole setup with an editorial example you can swap for this one.

Cap every working column, intake included

WIP limits, short for work in progress, are what turn the board from a picture into a control. A cap says a column holds only so many cards, and everything past that point has to wait or force a decision. On a support board the caps do three distinct jobs: they stop the desk starting more than it can finish, they make each queue visible, and they trigger the escalation rule.

Intake is the column teams leave uncapped, and it is the one that most needs a cap. An unbounded intake looks harmless because nobody is working the untriaged requests yet. But first response is the SLA that breaks there, in the one column where nobody can see which of the quiet requests is the burning one. The intake cap is a triage trigger. When New intake hits 10, somebody triages before the desk does anything else, and the oldest untriaged request stays under a day old because the trigger fires at least daily.

The escalation rule is the same shape for every column. When a column hits its cap, the desk makes a decision it would otherwise skip: finish something in the column, pull a person in, or make a visible priority call and move something back. Say In progress is full at 3 and a burning interrupt arrives. The lead either queues the interrupt in Triaged with a priority, or displaces the lowest-priority card in progress back to Triaged so its owner can take the interrupt. Either way the board records the cost. A chat-thread desk makes the same trade in the dark and nobody ever sees what the interrupt displaced.

One honesty note, carried over from the definition guide. Taskora gives you the board, the custom statuses, and the transitions, and WIP caps are a habit from kanban practice rather than something we enforce in the product. Put the number in the column name and hold each other to it.

Make waiting visible

Support work spends most of its life waiting, and most of the waiting belongs to the customer. That's normal for a service team. What breaks desks is waiting nobody can see, and the two waiting columns are where the SLA story of the board lives. Customer-waiting holds every request with the ball on the customer's side, a file you asked for, an approval, a test on their end. Verification holds every fix you have delivered and are waiting on them to confirm. A desk that reads those two columns each morning knows which SLAs are at risk without opening a report.

Aging awareness is the habit that keeps them honest. Once a day, sweep both waiting columns, oldest cards first, and nudge everything past your reply window. A card that has waited too long gets a nudge, or it gets closed with an invitation to reopen. The card's history shows when it flipped to waiting, so the sweep is one read from top to bottom.

Shift handoffs ride the board as well. A desk that hands over between shifts, or runs across timezones, can stop writing handoff notes nobody reads. The incoming person opens the board and sees what is in progress with whose name on it, what is waiting on customers, and what the caps say about the day so far. The board is the status update.

Subtasks and dependencies for the multi-step fix

Some tickets are small projects in a ticket's clothes. A laptop rebuild, an access request that touches three systems, a bug that needs to be reproduced, fixed, and verified. Kept as one card, a ticket hides where it really is, and the honest answer "half done" is not a state a board can hold.

Subtasks fix that. The card stays in In progress while its pieces close one at a time, so the board keeps telling the truth about the whole job while the checklist shows the parts. Dependencies handle the ordering between tickets: the fix waits on the vendor's reply, the deploy waits on the change window, the account cleanup waits on a manager's approval. In Taskora, subtasks and dependencies sit on the one plan rather than behind a premium tier, and what to check before you pay for either covers the questions worth asking any vendor, starting with which plan you would have to buy to get them.

One week on a three-person desk

A board is easier to judge in motion than in a spec, so here is one week on the example board above. Treat it as a worked example that shows the moving parts. It's not data from any real desk.

Monday starts with 11 cards in New intake, one over the cap, so the trigger fires and the first half hour goes to triage before anything else. Four of the 11 never make it to Triaged. Two are duplicates of tickets already in progress, so they close on the spot through the reject move. One is a feature request, so it gets recreated as a card in the product stream instead of living in the desk's queue. One is an unsupported ask, so it closes with a pointer to the vendor's docs. The other seven are real desk work: two quick fixes that run through every column within the hour and close, and five that land in Triaged, under the cap. By the end of the half hour, New intake holds only the 3 fresh arrivals that came in during the pass.

Tuesday's aging sweep finds one ticket that has sat in Customer-waiting since Friday with no reply. The nudge goes out before new intake gets touched. Old waiting is where SLAs die. New intake is where they only threaten to.

Wednesday at 11:00, a whole team loses access to a shared drive. The request enters New intake like everything else and gets triaged inside the minute, because the cap rule says intake gets a human look first. In progress is full at 3, so the lead makes the call in the open: the lowest-priority card in progress, a license renewal nobody is waiting on, goes back to Triaged, and its owner takes the outage. Triaged lands exactly on its cap of 6, which is the board saying the next accepted request has to bump something out.

Thursday, two fixes go out and their cards move to Verification. One customer confirms within the hour and the card closes to Done. The other replies that it still fails for one person, so the card goes back to In progress instead of lingering half-fixed. That's the rework loop earning its keep.

Friday afternoon the board reads 4 in New intake, 5 in Triaged, 2 in In progress, 5 in Customer-waiting, 1 in Verification. Every cap is respected, nothing is invisible, and the desk can say what Monday will need before Monday does.

What to ask before you put this board on a tool

A board this specific only runs on a tool that lets you build it. Four questions decide that, and the fuller comparison, including the tier gates to watch for, is in kanban software: what to compare.

Can the board run our statuses? New intake, Customer-waiting, and Verification are real states of support work, and a tool that offers a fixed Backlog, In progress, Done will make you hide them. In Taskora, every stream carries its own statuses, with a start status, a final status, and the allowed transitions drawn between them.

How do WIP limits work? Some tools enforce them and some leave the discipline to the team. Ask the question, then set a cap in a test board and see whether the tool or the team has to hold it.

Are subtasks and dependencies on the plan we would buy? The multi-step fix needs both. The answer decides whether the structure arrives mid-quarter as an invoice, and it is worth having before the move rather than after it.

Does the board stay fast at thousands of cards? Support desks earn history faster than most teams, because every closed ticket stays as a card and a year of closed tickets turns into a board of thousands. How Taskora's kanban holds its performance at thousands of cards covers why that is a design question and what breaks in boards built on renumbering.

Take the example table and these four questions as your checklist, and try Taskora to build the example board before you move anything.

Start with one queue

One honest limit first. Kanban organizes work that has already reached you. If what the desk actually needs is a help desk system's intake machinery, like email-to-ticket capture, canned replies, or customer satisfaction surveys, a board will not fake that, and a tool built for it is the right answer.

If the day this article opened with is the day your team lives, the queue is already a board. It's just written in chat threads. Rebuild one week of real tickets as a stream, name the statuses your desk already says out loud, put the caps in the column names, and let the board carry the interrupts it has always been carrying in the dark. Try Taskora on your real queue instead of a sample board.

FAQ

Is kanban good for a help desk?

Yes, because help desk work arrives unplanned, one request at a time, which is the situation sprint boards fail and flow boards absorb. Build the columns from the ticket lifecycle and give waiting its own state, and the board answers the question every desk gets asked in the weekly review: what is actually open, and where is it stuck.

What columns should a support kanban board have?

Yours, written down from what a ticket actually passes through. The example in this article runs New intake, Triaged, In progress, Customer-waiting, Verification, Done. If a ticket waits on the customer at any point, that waiting needs its own column, and the same goes for delivered fixes awaiting confirmation.

How do I set WIP limits for a support team?

Start with one active ticket per person in In progress, looser caps on the two waiting columns, and a cap on New intake sized to force daily triage. Then tune from a week of real counts, setting each limit at or just below what the column typically holds. In Taskora that stays a manual practice, as the caps section above covers.

Can I run this board for IT support and internal service teams?

Yes, the shape carries over. IT support tickets follow the same lifecycle, dependencies express the change-order sequencing, and internal service teams get the same benefit: visible queues, visible waiting, and a cap that turns overload into a decision instead of a mood. Kanban for service teams is the same method with the intake renamed: a portal, a shared channel, or a manager's forwarded email.

Does Taskora triage tickets with AI?

Not today, and we won't claim it does. AI assist is in development and rolls out to early members first. Until then, one capped intake column and a focused half hour a day is the triage system in this article, and it works without any AI at all.

All posts