What to "Keep, Drop, Add" to your next retrospective

2020-02-10 Maxence Dalmaisintroduction

It is no secret that we like as much facilitating retrospectives as thinking and discussing retrospective activities. So when somebody suggested us to do an article designed as the retrospective of an activity, we could only be excited. Because we like to go meta, we will here discuss what to “Keep, Drop, Add” from the “Keep,Drop, Add” activity. We choosed it because lots of team are using this single activity as the whole retrospective. Lots of ressources out there already explain that this is not the best idea, but we want go further and apply the iteration process and start from what you might be doing at the moment to something better.

Let's apply the keep drop and add principle to your retrospective
Let's apply the keep drop and add principle to your retrospective

The “Keep, Drop, Add” activity is presented as a variation of the “Short Subjects” activity presented in the got to read book Agile Retrospectives: Making Good Teams Great by Esther Derby and Diana Larsen. If you have participated in any iteration retrospectives, chances are you are already familiar with this format. If you don’t let me explain it brefiely. The idea with the “Sort Subjects” activity is to ask participant to brainstorm privately and write items related to 2 to 5 topics. Depending on topics, the individually written items can be proposed actions, as this is the case with the “keep, Drop, Add” variation, or more personal perceptions if using the “Mads, Sads, Glads” ones. Most of the time, after everybody disclossed their items, and similar ones are grouped, we will ask participant to dot votes the most important items for them in order to know which one to keep for the next phase of the retrospective.

By now, you are probably aware that this format as several issues. After all, you otherwise won’t be reading this article. Let’s list the main ones. When using the “Keep, Drop, Add” variation, you are asking people to come up with action items. If you then directly dot votes to the items, pick the 3 to 5 more popular ones and pick them as your actions for your next iteration, and call it done, you will have a very fast retrospective but it won’t turn to be really effective. By doing that, you are missing two points that are fundamental to a retrospective. The first thing is the Collective Intelligence. By asking people to think privately and directly pick items that individuals came up with, you are not using a collaborative process that could refine the actions. If you don’t spend time having an open discussion about items, what they are trying to solve, and how you can improve them, probably your actions won’t be as good as they could. Secondly, if using the “Keep, Drop, Add” activity as the whole retrospective, you are not creating a shared vision. If you ask directly individuals to come up with actions, they will usually think about their own context, and not think for the whole team. The 5 steps framework suggests you should start by “gathering data” and “generate insight” before “decide what to do”. Another way to think about it is to “Start with Why”, whereas if you ask to come up with actions you are starting with what.

ok, but concretely, what to “Keep, Drop, Add” to my retrospective?


  • Using “Short Subjects”, sometimes.

You might think that we are against using any “Short Subjects” based activity during a retrospective. If that’s the case, let me assure you that it is wrong. Every activity can be effective depending on the context and how does it fit with the whole retrospective. Using Short Subjects is an effective way to come up with a list of items that can be the basis of a discussion

  • Give everybody the same voice

Something good with the private brainstorming technic described before is that it allows everybody to express themself. The same thing goes with dots voting. We believe this is essential for your team’s health to respect this principle.


  • The single activity retrospective format

You should never run a retrospective with a single activity. Let me repeat it. NEVER. Even when using a complex activity as the Sailboat, you should never use a single activity. Take some time to explain the context to participants, to make them comfortable and be sure that they understand each other’s priorities. Finally, be sure to gather some feedback regarding the retrospective.

  • Rushing the retrospective

This point is similar to the previous one, but instead of being an easy rule to follow, we are here speaking about a general mindset. If you run a retrospective because the scrum guide says to do so but are not convinced about it, please ask for someone else to facilitate the retrospective. The retrospective is about aligning people and reach a global agreement regarding changes that they will conduct in the next iteration. By definition, this is a slow process. So please don’t rush your next retrospectives.


  • More Variety

That could have been the initial point of that article, but I guess it will make more sense to you. When using always the same activities, people tend to get bored and the same topics will come up again. The retrospective activities should fit the context. By definition the context evolves, so should the activities. Even when sticking for convenience with “Short subjects” based activities, changing the topics so that the discussions start with people’s perceptions of things, will dramatically change your retrospective outcome.


We gave you 5 actions to apply to your next retrospective. Why not more? Because 5 is already a big deal. Once you will have applied successfully theses pieces of advice, it will be time to try something more. And so on. It is only step by step that you will become a retrospective hero!