Trello has become an important part of how we collaborate at Kevel. We are big fans of the Kanban process* as a way to track the work we need to do and make it more manageable. We've found that the Kanban system is a good way to coordinate working on tasks between teams and individual Kevelers. For the most part, each column on our board is owned by a particular team, and it is the responsibility of that team to move cards in their column over to the right.
For example, when a developer completes work on a task and pushes the change to a staging environment, the card makes its way into the Acceptance Testing column, at which point it becomes the QA team's responsibility to complete QA and keep that card moving to the right.
As a developer, I like the Kanban workflow because it is easy for anyone to create a task card on-the-fly for any idea that may come up while working on a separate task. If we are pair-programming to implement a new feature and we discover an unrelated bug, we can take a minute to create a Trello card and put it in our To Do column. This way, we can be sure that when we find a bug, there is a process for prioritizing and fixing it instead of just saying "Oh, we should probably fix that" and then probably forgetting about the bug shortly afterward.
Cards with development work to be done come in from the left, and I feel motivated to do the work necessary to move them onward to the right.
One nice thing about Trello's UI is that it provides a search bar and filters that make it easy to focus on the cards that are relevant to my workflow.
As we have scaled as a company, we have started to find it more and more difficult to find the right Trello setup for our team. It is not always clear how we can best arrange our Trello boards, columns, and cards.
We use Trello to manage nearly every aspect of our operations, including product design work, support tickets, quality assurance, and even our interview and hiring process. Practically (maybe even literally?) everyone in the company uses Trello on a regular basis.
Our engineering team is split into 3 smaller teams: the UI/API team, the ad serving engines team, and the reporting system team. In addition to the engineers, there are other participants in the Trello process, such as our support staff, QA engineers, and Kevel's VPs of product and customer success.
Our Trello workflow started out minimal, with 5 general columns ranging from "To Crush" to "CRUSHED!" and each column was general enough in purpose that it was relevant to everybody. However, with time, our workflow evolved to the point where we now have 19 columns, and each column is focused on an aspect of our process that potentially only a handful of people care about. In each Keveler's view of the board, the remaining columns are not relevant to his/her workflow, and are typically regarded as Kanban noise.
We experimented with trying to reduce the noise by pulling out handfuls of related columns into separate Trello boards, but that got confusing; it ended up being harder to visualize the "big picture" than it was with the single, monolithic Kanban board that we had. So, we were forced to go back to having a single Kanban board shared by everyone in the company, and for a long while, that was "good enough."
`) keyboard shortcut which toggles between viewing only the columns I care about vs. all of the columns.
My "blind spots" -- the columns that I choose not to see -- are:
Managing these blind spots is trivial. Most of the time, we don't need to look at them at all, because we have a process that allows us to focus on only the columns that we interact with directly.
For example, after we complete work on a task and deploy our changes to a staging environment, my team moves the card from "Needs Staging Deploy" into the (hidden) column "Acceptance Testing" for QA testing. We are then free to forget that that card exists until it either comes back to us for additional development or it passes QA and moves on to the "Needs Production Deploy" column.
If we ever find ourselves wanting to see what's going on in the other columns (e.g. "Do we have any cards in QA right now?"), we can press
` and see all of the columns that are usually hidden. When we're done, we can press
` again and go back to focusing on our part of the board.
For added convenience, I like to use a couple of built-in Trello keyboard shortcuts:
Qtoggles on/off a filter that shows only the cards assigned to me.
I keep this on most of the time, to keep things super minimal. At times, when I'm using this filter in combination with my script, there are only 2 or 3 cards visible to me on the entire board!
> while a card is selected moves the card one column to the left or right. This includes columns that are in one of my "blind spots," which is great because I don't even have to un-hide all of the hidden columns when I'm moving a card forward into a blind spot.
I can just press
> and the card will move on to its next phase in our Kanban process.
I have to be honest: seeing a card completely disappear when I push it forward into another team's territory is pretty satisfying.
Dave is a Software Engineer at Kevel, where he uses Clojure and ClojureScript to build UIs, APIs, and backend services. He is also the creator of the Alda music composition programming language.