Liked-Learned-Lacked-Longed for
Method
A retrospective format that collects what each person liked, learned, lacked, and longed for, then picks what to change next cycle.
Why it exists
Open a retrospective without columns and the talk runs in two directions only: it was good, or it wasn't. What the team figured out during the sprint, the cycle that just ended, and what it wishes had existed get written down nowhere.
The room tends to go like this. The facilitator asks how the cycle went and for twenty seconds nobody speaks. Eventually one person raises a single complaint, and the remaining 25 minutes go to that one thing. Nothing gets decided about what to change next, and what's left is a mood.
4L replaces that open question with four columns. Each column asks for a different kind of answer, so someone who finds a complaint hard to voice can enter through what they learned instead.
How to run it
Write the four columns up with their meanings, have everyone write silently, then read, cluster, and choose one column at a time. Thirty minutes to an hour covers it.
- Split the board into four columns. Liked is what went well this cycle, Learned is what the team found out, Lacked is what would have helped but wasn't there, and Longed for is what the team wants going forward.
- Everyone writes on sticky notes in silence for five to ten minutes. One or two notes per column is about right.
- Read them out column by column and stack similar notes on top of each other. Twelve notes commonly collapse into three clusters.
- Each person spends two or three votes across the clusters. Only the two or three that draw votes get discussed further.
- For each one chosen, write a single action for the next cycle with an owner and a date. With no action item, the same note comes back at the next retrospective.
The format fits a broad sweep of one cycle, and it fits a team running its first retrospective. It falls short when the job is driving an incident down to its root cause — four columns gather symptoms, and nothing in them pushes the question a second time. Past ten people, reading the notes alone eats 40 minutes.
How it sounds at work
"Let's run this one as 4L. In the Lacked column, write the situation rather than a person's name."
Asking for the rough edges to be framed as structure instead of blame.
"Everything in Longed for is deployment automation. Let's raise that as an item for next cycle."
The wishes clustered in one place, so they should be promoted into an action.
"An empty Learned column means nothing new got tried that cycle."
Reading the blank space in a column as a signal in itself.
Pagedeck spends 30 minutes on this format after the review on the last day of the cycle. Theo posts under Learned that pulling the payment logs into one place cut root-cause hunting to 20 minutes, and Noa posts under Lacked that a mockup built on fake data had to be redrawn once the real data arrived. Mia counts the votes, takes the top two clusters, and the team agrees to pull one real record before the next mockup.
Easily confused with
| Term | Columns | How it differs |
|---|---|---|
| Keep-Problem-Try (KPT) | Keep, problem, try | One of its three columns is already the action, so it reaches a conclusion faster. 4L takes learnings and wishes in separate columns to gather more, and keeps the move to action as a later step |
Where to go next
- Retrospective
see the meeting this format runs inside and what that meeting is for
- Keep-Problem-Try (KPT)
compare it with the other format used most often in the same slot
Last updated 2026-09-16