VP Product at BlaBlaCar
The software industry has a unique privilege: shipping a flawed product is not that big of a deal. The cost of fixing a software product and redistributing it again is so low that your success is not determined by the quality of your product, but by how fast you are able to fix flaws.
Software practitioners call it the speed of iteration. If you ship a product today, called iteration 1, how soon will you be able to get customer feedback in order to spot flaws, fix them, and ship iteration 2?
Debates remain regarding what your first iteration should look like, and how you should progress towards your end-vision. Many metaphors have been used to settle those debates.
Let’s take the example of a car. How would you proceed to build it, in an iterative manner?
One approach is to build the separate parts, one by one. First the wheel, then the chassis, then the body-side moldings, etc.
The problem with this metaphor is that it doesn’t really bring the promised value of an iterative process. Shipping the wheels isn’t enough to get feedback on it, so you won’t be able to fix anything at that stage. In fact, you won’t get any feedback until you’ve built the entire car.
A better approach could be: first build a scooter, then a bike, then a motorbike, and finally a car.
This second metaphor addresses the feedback loop issue described earlier. At each stage, you have a product that can be used. But other questions persist: How do you know where to start? And how do you jump from one iteration to another?
Surprisingly, answers to these questions can be found in the entertainment industry. In France, one person in particular has demonstrated what it’s like to launch a product and iterate on it, adding one by one the features that were missing at first.
His name is Alexandre Astier. His product is a TV series called Kaamelott.
I watched it multiple times, from beginning to end.
At first, for entertainment only. But some episodes about leadership and fatherhood took me by surprise, touching me deeply. So I watched the TV series a second time, to unpack that feeling.
The following viewings revealed that Kaamelott could be an endless source of inspiration, if you were willing to pay attention. And it became clear to me that Astier was actually giving a masterclass in product development.
The last iteration of Kaamelott, a movie, was supposed to come out today. Since it got postponed to an undefined future because of Covid-19, this article is an attempt at honoring the fascinating evolution of Kaamelott over the years.
Kaamelott is a new interpretation of a famous legend: the Knights of the Round Table, and their quest for the Holy Grail.
Alexandre Astier started by a short 15-minute movie, called Dies Irae, in 2003. For budget reasons, this short movie was highly restrained in terms of production — half a dozen actors, entirely shot in one room.
Yet, this first iteration already contained a distinctive feature: the comical contrast between the solemnity of the setting and the mundanity of the dialogues. In opposition to Hollywood movies, the characters weren't expressing themselves through brilliant declarations, they were chatting like regular people and misunderstanding each other most of the time.
This short movie got the attention of French producers. They offered Astier to turn the concept into a TV series, with a prime time slot on a popular TV channel. But it had to comply with many constraints:
Short duration (each episode should last three minutes);
Stand-alone episodes (people should be able to understand an episode without having watched the previous ones);
Funny tone (each episode should follow a comic structure, with a punchline at the end);
Low budget (the production should be optimized at all costs).
Alexandre Astier saw the opportunity to develop his product further. He accepted all the constraints, and decided to make the most out of them.
He shot one hundred 3-minute comic episodes inside a castle. While doing so, he focused on the one thing that wasn’t impacted by the constraints: introducing and developing endearing characters.
The ratings for this first season were positive. A second season was rapidly requested and broadcast, attracting millions of viewers every night.
With the success of the first two seasons, the tone had been set, all the characters had been introduced, and the distribution channel had proven to be highly effective. It was time for Kaamelott to move to the next iteration.
In software development, you pay close attention to the people who use your product very often. They are called power users, and they usually demand dedicated features. For Kaamelott, a similar audience appeared: hardcore fans — people who didn’t miss a single episode.
The emergence of this loyal crowd was a great argument for Astier to break a first constraint: stand-alone episodes. He introduced the notion of chronology in a soft manner. In season 3, we started to see twin episodes: for example, the 23rd and 24th episodes were in fact one story, broken into Poetics - Part 1, and Poetics - Part 2.
In season 4, Astier offered his power users a full chronology during the entire season. For the first time, Kaamelott had a plot that they could follow from one episode to the next — the ultimate reward for hardcore fans.
The chronology allowed Astier to loosen the comical constraint too. He could now plant seeds of tragedy in one episode, let them grow during a couple of episodes, and reap the dramatic results later on. It’s no coincidence if, in season 4, we saw King Arthur hit by two consecutive blows: the betrayal of his best friend, and the departure of his wife.
Thanks to the unstoppable success of seasons 3 and 4, Astier was finally able to renegotiate all the initial constraints he had started with.
In seasons 5 and 6, the most visible change was switching to 50-minute episodes.
The second change was expanding from the castle and setting the story into more spectacular outdoor sceneries, in the mountains, on the sea shore, and even in Rome.
The third change was more controversial. At the risk of surprising and disappointing its hardcore fanbase, Kaamelott evolved towards a more melancholic tone. Fewer punchlines, darker themes — depression, death of newborns, loneliness of power, etc.
The comic tone hadn’t disappeared entirely, but in seasons 5 and 6, when the viewer would be crying, it wouldn’t always be out of laughter.
The effects on viewers paid off. Some power users felt betrayed, but this iteration attracted a new audience, seduced by this unique mix of fun and serious.
Coming back to software development, the evolution of Kaamelott offers four lessons:
Ship a full product at each iteration. The first 15-minute short movie was not a 15-minute excerpt from the final movie. It was the core concept of Kaamelott, delivered in a constrained yet meaningful way. When wondering what iteration to ship next, the question to ask yourself is: Considering the constraints I have to work with, how can I produce something of value?
Keep your final vision in sight. Feedback isn’t the only driver of iteration. You should always keep your ultimate vision in mind. From the beginning, the Kaamelot crew cared a lot about the realism of the costumes. Even if very few people noticed, this obsession led them to constantly refine the appearance of each character, as their knowledge of the 5th century clothing increased.
Allow the constraints to choose your path. It’s unlikely that you already know every single detail of the end product you have in mind. This uncertainty can be paralyzing. That’s why you should look at constraints in a positive light. They help you focus on fewer things, by telling you: You cannot work on this, so start with that.
Use the success of an iteration to build the next one. Each iteration should make your product more solid in one area. You should then use this area as a stepping stone to build the next iteration. For example, the proximity with the characters developed in the first seasons of Kaamelott made it all the more dramatic when relationships turned sour.
Today, I should have been in a movie theater, to watch the first Kaamelott movie.
The last episode of the TV series was released in 2009. For this movie, Astier absolutely wanted to have the right budget and the final cut. He waited a decade for those conditions to be met. This determination makes me curious about this upcoming iteration.
What constraints were lifted? How did that help him make another step toward his final vision?
I guess we will have to wait a few more months to get the answers.
Create your free account to unlock your custom reading experience.