We monitored our team’s mood for 2 months, look what we found
We measured the type of tasks together with the mood of each member of the development team for 2 months.
Introduction
Since I joined FXStreet one of my main objectives was to make sure that everyone in the team was motivated and happy with the projects in the company, always balancing different type of tasks. Among other tools to measure motivation, the use of what is call “Continuous Retrospective” proved to be an interesting approach.
In each stand-up, team members draw a scribble in our white board to illustrate their mood explaining why they feel that way. We obtain two different insights from this data; Firstly, we can monitor every team member. Secondly, after a sprint come to an end the overall team vibe can be analyzed just simply looking at the whiteboard.
“Continuous Retrospective” where in each stand-up team members draw a scribble in our white board to illustrate their mood and a short explanation why they felt that way.
After some time doing this I started to think about some interesting questions; are the type of task directly correlated with the mood?, are innovation tasks the ones that always make people happy?, the day of the week affects on a person’s mood?.
The Experiment
I wanted to go one step further and try to correlate the mood with the type of task we were doing, so this is what we tried. We have been tracking all the explained above for more than 2 months.
The experiment is based in 2 variables, mood:
- Green .- Those days where things were really good, you did not have any issues, you solve a difficult task, or simply you were happy.
- Blue .- A day where you did not have any excitement and things went normal.
- Red .- A bad day, things went wrong in unexpected ways, a bug that cannot be reproduced or simply you did not feel well.
and task type:
- Bug (B).- Those tasks related with bugs or critical situations.
- Normal (N).- User Stories in sprints with technology already implemented by the team.
- Innovation (I).- Those User Stories where we need to investigate new technologies, architectures, etc..
- Management (M).- Meetings or other types of tasks that do not involve programming.
My goal was to discover any anomalies further than every time a person has a “Bug task” the outcome will be red and on the other hand people involved in innovation will be always green.
I monitored 40 days in total and the results are in this spread sheet.
Before jumping into the final conclusions, just a couple of things to mention about the team. We are 7 full stack developers and 1 front-end developer. We rotate tasks every sprint, so if a person is on duty in a sprint, taking care of emergencies or bugs, next sprint will take care of an innovation user story and so on. This is something crucial to take into account when analyzing the data set.
Conclusions
In order to get some conclusions I started working on the spread sheet basically counting things (I created a couple of google sheet’s scripts to do it for me) like: nº of green days / doing I tasks, nº of blue days / B tasks, total number of green, blue or red days, days of I tasks and so on. This can be extended to nº of green Mondays vs nº of green Fridays or even tracking a particular developer and how he or she feel about B tasks.
First thing that impressed me was, that the amount of green days with an N tasks outnumbered the ones with I, also B had the same number as I. One possibility for this is the “immediate reward” concept, usually Bugs are small issues that once solved leave a good sensation, if a developer solves 3–4 bugs in a day that will be a good day. A interesting improvement could be to split B tasks between easy and hard ones (e.g. track a degradation in performance).
What I did not realize was that apart from the relation between mood and activity this exercise gives other types of feedback. For example, it can tell you how many time your team devotes to Innovation or if in the last two months people have been happy or if the team as a whole is having a hard time.
With only two months of data it is not enough to set any realistic conclusions but in my personal opinion this is a good way to monitor the health of the team with objectivity. You could keep a historical record to measure everything and make decisions based on statistics.
The project
Since I really believe this method can help other companies I would like to start a open source project to collect data from other companies in order to create a big data set that will keep those records and help to extract that information easily.
One of the first things I would improve is the data input, the process now is a little pain. I carry a pen and paper to the stand-up and jot down stand still every input from every person (Mood/Task) then, when I go back to my desk, I import the data into a spreadsheet.
Another thing that we could improve is the data analysis, if enough people send their information we could apply a Machine Learning algorithm in order to forecast your team’s mood if you know what type of task they are working on.
If you are interested please drop a comment and we can discuss it privately and create a git repository and start working on it.
