The Task Board and Its Views

The seven views

Three of them do most of the work.

Board is the kanban view: columns are statuses, cards are tasks, and you drag between them. It answers 'what state is everything in?' — where work is piling up, what is stuck in review, whether anything is in progress that should not be. It is the right default for a team running a flow.

List answers 'what is there, and what do the fields say?' — it is denser, sortable, and far better for bulk reading. When someone asks a question that involves comparing many tasks on a field, List is the answer and Board is a slow way to get there.

Calendar lays tasks out by date. It answers 'when is everything due?' — which is a different question from 'what state is it in', and the one that surfaces the week where four things land on Thursday.

The other four are situational, and worth knowing exist so you reach for them when they fit. Timeline lays tasks along a date axis for reading a schedule rather than a status. Table is the spreadsheet-style grid. Gallery is a card grid grouped under the same status columns, which suits work that is visual. And Map plots tasks that carry a geographic location onto an OpenStreetMap view — tasks without a location simply are not plotted, so it is empty and unhelpful unless you are actually tracking work in places.

Switch with the view toggle in the header. Your filter and grouping choices carry across, which is what makes switching cheap.

Adding a task

From Board, click the add-task control at the top of the column you want the task to land in. Creating it in the right column saves a drag and, more importantly, records the correct starting state.

From List, use the add-task row.

From anywhere in the app, press c.

All three open the same form. Title is the only required field; assignee, due date, priority and description are all optional and all useful. The rule of thumb is to type the title while you have the thought and fill in the rest when you triage — capture beats completeness, because an incomplete task exists and a perfect one you did not write down does not.

Making a big board readable

Four controls, in the order you will reach for them.

Filter narrows what is shown — by assignee, status, priority, label, due date, custom property. The single most useful filter is 'assignee = me', which turns the team board into your board without leaving the shared context.

Grouping changes what the columns or rows mean. Grouping by assignee turns a status board into a workload board, which is the fastest way to see that one person has eleven things in progress.

Swimlanes add a second axis — horizontal bands cutting across the status columns. Swimlanes by priority or by epic are the usual choices, and they are what make a sixty-card board scannable rather than exhausting.

Column configuration decides which statuses appear and in what order. If your board shows columns nobody uses, hide them; a board that lies about your process gets ignored.

WIP limits

A WIP limit caps how many tasks may sit in a column at once. Kavanah supports them per column, and they are worth understanding properly because they are widely misread as a productivity nag.

They are not. A WIP limit is a diagnostic. When a column hits its limit, the useful information is not 'someone is doing too much' — it is 'work is arriving here faster than it is leaving', which points at the constraint. A review column that keeps hitting its limit is telling you reviewers are the bottleneck, and no amount of pushing on the people upstream will help.

Set limits generously at first — high enough that they only trip when something is genuinely wrong — and treat a trip as a question rather than a violation.

The archived view

Archiving a task hides it from the board without deleting it. The board has an Archived view for browsing and restoring them.

This is the right way to get noise off a board. A board with two hundred done-and-irrelevant cards is a board people stop reading, and the alternative — deleting them — throws away the completion history that velocity, cycle time and estimate calibration are computed from. Archive is the middle path and it is what you want almost every time.