Before releasing a game and to improve its reception afterward, it is important to have it tested in several ways. It is thus very important to make sure that the difficulty curve is smooth, that the game mechanics can be learnt properly by the players, that the pacing is neither too harsh nor too soft, and, over all this, that the game is fun and enjoyable.

And then you encounter several constraints, and you need to:

- Be able to stand back enough and judge your game objectively;
- Get as many people as possible to play the game to get the biggest possible picture with a wide array of behaviours;
- Do that soon enough to still have time to modify the game and have it analyzed again before it is released.

Playtesting is the key to these problems.

The idea is pretty simple: get someone who’s not involved in the project to play the game, in as “real” conditions as possible. This means not answering their questions, only giving the most basic explanations (similar to those they would be given in the final game), and making no comment on what’s going on.

Watching them play over their shoulder, we analyze their behaviour, the way they approach the game, their quickness to make decisions and act, the overall understanding of the game they seem to get, along with various objective facts (building order, number of cows lost to the Martians, resources harvested, number of uses of various features, etc.)

â–² A quick explanation of the level to our nice tester.

The levels which are to be tested are chosen beforehand, depending on what type of information we need to gather. In a perfect world, the last playtests are done through the whole main game mode, in real conditions.

At the end of the session, the player is given the possibility to play over the levels they enjoyed most, if they want to. Then we proceed to write down their first impressions, not asking any questions so as not to influence them. To gather a lot of spontaneous information, it is exceedingly important to let the players talk, not pressuring them or filling in the blanks. We actually seldom have a player who has nothing to say about their experience.
After that, we do have a list of questions to ask. These questions are open, in order to avoid any bias. It is important to check your ego at the door, listen to the player, and resist the urge to make them say what you’d like to hear.

â–² The pack is observing, silently.

We have asked brave and eager players to playtest “Beware Planet Earth!”…and we did learn a lot of stuff about our own game. Without saying too much and spoiling the surprise, here are some of the improvements we made after their feedback:

- We regrouped the objects littering the screen without making the game play any better;
- We made visual feedbacks much more obvious for them to stand out;
- We made sure the player never feels overwhelmed by too many decisions to make at any given moment;
- We introduce new Martians and machines more elegantly and obviously through level design;
- We adjusted the difficulty curve;
- We re-balanced the importance and use of some specific feature so as not to make them overlap with the core game play;
- We did no longer hesitate to go nuts while tweaking parameters…and we continued to do so even more from then on;
- We made stuff simpler, simpler, and simpler.

The actual list of improvements is actually much longer, but the conclusion is this: playtesting sessions are the ultimate ordeal for the game before releasing it, and definitely the best way to improve it.

However, it is still important to keep this in mind: never take all the feedback and suggestions from players at face value. In the end, it is us, the developers, who are responsible for the overall consistency of the game, and we must never lose sight of this objective.