As concepts go, it’s hard to fault, really; they give a sense of achievement that might otherwise be lacking until the ending of a game, they provide a game-skill measurement for social discussion, and – this counts the most for us! – they provide an alternate avenue for game exploration.
That last point is pretty important, because we need our game content to count for more. One playthrough of each level isn’t going to cut it on the hour estimation or the player’s sense of consumption. They want a buffet table of content, and we need them to chew what we give them for as long as possible, so to speak, or else the kitchen will run dry.
So, with achievements in mind, the problem is…how do we design them in a way that makes sense? What are the commandments of achievement design?
1. Thou Shalt Not Make Elitist Achievements
That means, no “#1 ranking online” achievements, no “10,000 kills online” or “stayed undefeated by myself for 3 days.” People want to max out their gamescore. If you make it too hard, they will just try a different game. Go for things the player can do, given some time playing the game. Ranked achievements are basically “popularity” achievements, frequently broken by the design of the ladder system involved, or the closing of your multiplayer server a year later. If the community goes kaput (as almost everyone eventually does), your achievement design is broken.
2. Thou Shalt Follow Interesting Situations
Don’t bother with “beat mission 1, beat mission 2”. The player is doing that anyway, so you would just be encouraging them to follow the path and leave everything else alone. Instead, look for interesting things – things the player might not normally try. Reward the player who finishes a level with only funny hat attacks, or the one who destroys all of the doors. Portal rewards camera vandals, for instance, essentially adding a new “find the cameras” minigame to the overall game progression. Beyond: Two Souls rewards finding different outcomes.
Think of Achievements not as rewards for doing the obvious, but as carrots you can use to lead players to undiscovered content.
3. Thou Shalt Not Overload on Minor Things
Though the TRC usually helps prevent this by limiting the total number of achievements, some games seem to go overboard with quantity over quality, so to speak. A hundred achievements for collecting specific pieces of laundry is little more than micromanagement, rather than an actual sense of “achievement” as the name suggests.
Try to seperate out your achievements to cover different topics, and set the points accordingly. If 19 achievements in your game are worth 5 points or less…why exactly are they there? And why would your audience stick around to achieve them?
4. Thou Shalt Not Allow Everything to be Achieved in the First Playthrough
The whole goal here is to keep the player coming back to try things they wouldn’t normally. Okay, maybe it takes careful GameFAQs study to do it, but if some players are achieving everything the first time through, there is probably something wrong with your achievement design.
Why? Because, once again, nothing out of the ordinary has happened. Finding all the bananas may seem interesting and out of the way, but if the player finds them all on the path, then you have to ask yourself if your content is really capable of more than one run through.
5. Thou Shalt Not Insist on Repetition
“Grinding” – the act of repeating simple tasks in a game – is generally not enjoyable without superb feedback. There is some room there, for the hardcore. Maybe you have to shoot 10 tanks with each gun – that would be an effective way to lure players into trying different weapons. But at some point it becomes a merit badge that the game didn’t crash. If shooting 10 tanks becomes 10,000? Congratulations, you’re trying to reinvent Desert Bus.
As designers, we should be aiming for variety. Lead the players to our most interesting content, mechanics, or challenges, instead of having them repeat the same old one a thousand times more. This is a game, not a mortgage, and extreme repetition only serves to age existing mechanics.