#no estimates, agile or waterfall?
In mid May this year, I started with a summer challenge, to do a full lenght feature film during the summer. You can read more about the film ”Att vinna ett kex / Lyckohjulet” (in Swedish).
Foto: Casimir Artmann - Jonstops marknad
The question for todays blog pos is which project method is best suited for this project?
Lets start with #noestimates as a first try.
A feature film need to be at least 80 minutes long, so we can’t deliver less, and we can’t add more scenes to a published film later. We have a number of other constraints as a very low budget, limitations when to shoot and where, as well as the availability of cast & crew.
To start filming without estimates for a plan, with these constraints is more or less sheer stupidity. Therefore #noestimates as a concept are thrown out of the window.
A feature film is by default a very waterfall project, as you have main deliverable [the film] and five very well defined phases; development, pre-production, production, editing and distribution.
The question is how agile can you be in your process, under these circumstances?
Given a very tight time plan, we had to start pre-planning before the manuscript was developed. We know that we only could set aside ten full days of shooting, and only a limited number of days per location. So the script was adjusted to these limitations.
The decision to use only available technical equipment was partly a self imposed constraint, partly due to budget reasons, but also from a wish to make it as simple as possible on set, where we had a very limited crew.
Our script was only little more than 35 pages compared to 60-80 pages for a normal full-length feature, but we still had about 60 scenes for the film (quite normal). You can compared this to a situation where you don’t have very detailed requirements before you start your coding. This meant that we had to rely much on set design in existing locations, and be ready to improvise acting, (but not dialogue) on set.
If you are unsure, you make a proof of concept, and we made two of them to verify our approach before starting filming with actors.
Using text to speech software, we made a transcription of the manuscript and added sketched storyboard images to the voice over. This gave a a rough idea of how long the film were and if the story hold together story wise.
The other PoC was to try the camera equipment at a nearby festival, to validate that what we had [for equipment] worked for those scenes.
One of the main tasks in the pre-planning phase was then to plan which day on nine days each scene was to be shot. I put aside one full spare day for scenes we couldn’t film as scheduled, but could be delayed.
For this planning I had to make a rough estimates for how long time each scene would take, based on how long the scene was and the complexity of the scene. Knowing this would not be the exact truth on set, wi still needed a plan. Then schedule each scene depending on availability on locations, cast, time of day and working time limitations.
From an agile perspective, we had a small self contained team (5 persons behind the camera) where everybody had several different roles. We had ten sprints, where we delivered a number of features each day for a daily release of video & audio clips.
For each day, we had a call sheet describing which scenes to shot, who should be on set, props and set design. E.g. very much a Kanban approach to each working day, even if we didn’t have a physical board on set.
Our production phase is now almost finished, and we have shot most of scenes we planned in with eight days, within time, budget and quality. (With an exception for three scenes that would be filmed on the tenth spare day and three scenes to be filmed later this week as planned).
Even if the whole film is not filmed have we started an initial assembly edit, so a it’s not a true waterfall project either.
Some of the lessons learned, so far :
More time is needed for preparation with actors and preparing shots. Lack of preparation took extra time on set, where other people had to wait.
You need more time for internal communication before starting filming and during production than we set aside.
The team was to small and we would have been at least six persons behind the camera instead of only four. You miss things and people getting stressed.
Don’t shot eight days in a row, even if actors and part of the crew got days off. Not planned from start, but things happens during production that we didn’t ancipitated.
Not to forget, we have succeeded with the challenge so far.