Making games is my hobby, dare I say, my obsession. However, like a lot of other indie gamedevs, I have a full time job. By day I work for a digital marketing agency. Recently, the company has started to adopt a set of principles referred to as ‘The Lean Startup’ – linkie to website here : http://theleanstartup.com/. I’m going to try and avoid using lots of buzzwords in this blog post as articles about Lean Startups seem often to have way more buzzwords than content 🙂
One of the central concepts of a lean startup is the Minimal Viable Product ( MVP ). ‘What is an MVP?’ I hear you cry. Well, simply stated : it’s the least amount of work you can do to get a product ‘out there’. By ‘out there’ I’m referring to getting friends, family, strangers, clients actually using your product and providing you with feedback on it. The general approach is basically a feedback loop – get feedback on the MVP, change the MVP to incorporate the feedback as cheaply ( time or resource wise ) as you can, release a new version of your MVP, get feedback on it, etc; etc;
Think about it very much as step 2 in the underpants gnome great money making scheme.
What I found really interesting about the concept of an MVP is its a methodology I use all the time without ever having thought about it – what do you get when you mix up indie gamedev with an MVP? You get a game prototype. Suddenly all those hours spent cutting out bits of paper and pestering people to play a paper version of game before starting coding made sense. Sending out desperate cries of ‘play my game! tell me what you think’ on social media wasn’t just me being devoid of social interaction – it was a viable business technique!
I read every single bit of feedback I get related to my games like a starving man. I make a point of replying to as many as I can. ( Of course, not everyone gets a reply – ‘Your game gave me cancer it was so bad’, really doesn’t deserver a reply ). And, more importantly, I implement a LOT of the feedback I get, and along comes another version of the game.
I’m now going to talk a little about how I used the idea of an MVP to build Wastelands.
Wastelands, my most recent #1GaM entry, was built as an MVP. Initially when I was approached by a co-worker with the idea of building Wastelands, the phrasing of the goal was ‘a dopewars clone, set in a zombie apocalypse scenario, where you sell drugs to your real life friends, who can die in the game’. My first thought was yuck, zombies. As much as I love the zombie apocalypse scenario I think it is become more and more clichéd. My second thought was, this is a really awesome idea minus the zombies so I wrote a little story about aliens. We refined this somewhat between us and got down to deciding how to turn this idea into an MVP. So, the major hook is having your friends in the game. But whats the best way to go about that?
There are so many ways of getting your friends into a game these days, its almost indecent its not used more to bring interesting NPC’s into your game. Twitter, facebook et al all provides ways of pulling this data down. But working with with those techs is time consuming. We are thinking MVP. What is the laziest way of having this tech in the game without resorting to the full twitter/facebook implementation? ( ‘laziest way’ will probably be frowned upon by practitioners of the lean startup, but it’s exactly how I think of it 🙂 ).
It’s simple – the first MVP released was shown to co-workers inside the company. So with this limited audience, we found a global set of friends for the entire company – the board of directors. No need to worry about a facebook or twitter integration, no need to worry about any sort of online/offline integration. Simply an array containing the name of each director, along with an image.
The feedback was good – the concept was interesting, though immediately we had to deal with the idea of a public release – the board of directors where all very good internally – but would mean nothing to anyone else. So we evolved the MVP and the next version of the game contained some text boxes at the start of the game that allowed you to put a friends name in, and select a sex – this allowed us to populate the world with your friends, along with a not very specific image of them. You can guess the next bit of feedback we got from the tests.
‘We want to see our friends instead of random NPC faces !’ they cried. Still with another MVP in mind, we looked for another lazy way to achieve this. The idea came pretty quickly. By adding an email address box next to each friend box, we could query gravtar and pull an image of a friend to place it in the Wastelands world.
This was just one example of feedback and implementation inside the Wastelands MVP. There were many other things that came to light – resulting in a game which wasn’t exactly like our first vision, but one which had been developed hand in hand with players, with us able to drop and add features dependent on actual feedback. We had something viable that people enjoyed playing.
lamentconfig
This worked fine, but building static levels like this struck me as a bit boring. If in doubt I try and use procedural generation as much as possible so I decided to do that instead 🙂 I tweaked the game to have to distinct map layers – the first one contains the desert environment, the second will contain the resources that you are collecting. I used the diamond-square algorithm to generate some realistic looking desert and the resource layer.
