A central location to come home to
After launching You Suck At Parking, Happy Volcano wanted a central home for their analytics data. David Prinsmel, Happy Volcano’s game director, and their data engineer, Rahul Jani, found that Player Warehouse let them be more flexible around storing and processing their data. They could pull it all into one central location and then create their own visualizations off the back of that information.
Here’s what they found along their journey.
Focus on your onboarding process first
If you want players to stick around, you need to make sure they feel comfortable in your game. So your onboarding process is vital to your success.
Through their data, Happy Volcano realized that people were far more likely to drop off if they didn’t play through some of the single-player races before hopping over to multiplayer.
“It makes sense when you think about it,” David said. “As a player, it’s frustrating to join a multiplayer race and get absolutely thrashed immediately. You just end up leaving. So we’re now working on how we can improve that journey. The opposite is also true – If player finishes the single-player campaign but never plays an online race, they’re going to leave after the campaign. Using the data, it’s now clear which day we need to convert them to multiplayer.”
Of course, every game is different. But for Happy Volcano, they want to encourage players to move over to multiplayer eventually, which is why those are the metrics they put the most weight on tracking.
Stick to your gut
“Fun is key, first and foremost,” David said. “You want a really simple, fun concept at the core of your game. Make sure you can describe it in a single sentence and make a prototype. Then build on that one core mechanic that’s unique – because that’s your hook.”
But remember, prototypes are never perfect. Sometimes, the data can show that a prototype isn’t working as well as you hoped. But that might indicate that you simply need to refine the idea, rather than scrap it completely.
“It’s really important to have lots of different perspectives early on,” David said. “But it’s also important to know which of those perspectives to listen to. A prototype should show you whether an idea has legs and is fun to play, but it definitely won’t get the retention of a properly polished title.”
So if you test your prototype and you find that the feedback is good, but that there are problems – that doesn’t mean you should kill the idea completely. Maybe the controls are too confusing, maybe the learning curve needs sanding down, or maybe there’s a missing element, like a progression system. Data can help point you in the right direction, but won’t give you the answers upfront.
Give your game time to grow
“I’ve seen that some people don’t give a game a chance to breathe,” Rahul said. “They try to close the project too early. Or they sand off the edge, which actually makes the game different.
“Emotion comes before data,” he added. “So include data early, but don’t listen to it too soon. It might just be the magic sauce that makes your game fun.”
With You Suck At Driving, Happy Volcano realized that, while they had a battle pass and players could unlock cosmetics, there needed to be another sense of achievement – a big horizon for players to earn. New items, a new currency, and less focus on monetization.
“We realized that we had all the winning cards in our hands. We just didn’t quite play them outright,” David said. “For example, our last update included far more for the players to strive for – it was essentially a completely free DLC.”
Capture everything you can for later
While it’s important not to overload yourself as a designer, a new data engineer will want to capture far more of the data points. Not only will this help them create a benchmark, but it’ll help them learn the tool.
“From my point of view, I learnt a lot about how GameAnalytics is set up by making sure I was tracking everything I could,” said Rahul. “If I hadn’t tried to capture all that data, I wouldn’t understand the tool as much as I now do.”
Now that Rahul understands how the tool works, it is much easier to rely on the data.
“One of my major frustrations is that the way that data that comes in keeps changing between different hardware or consoles. So it might be different to the same data collected from another platform. You end up working with a lot of black boxes,” said Rahul. “But GameAnalytics is my rock. I know how it works, and I can trust the parameters that I’ve set up with it. It’s nice to have something to come home to.”
You don’t need a huge team to do advanced analytics
Even if you’re collecting vast amounts of data, you don’t necessarily need a huge team. Happy Volcano uses our Player Warehouse – which brings all their data into one central location – to make sure they capture everything they can.
“If I didn’t have access to Player Warehouse and that data, I wouldn’t be able to output anything useful,” Rahul said. “But the fact that I have access to that means that I don’t have to set up a database on our side.”
This simplicity has meant that Happy Volcano has been able to keep their team numbers down, while still getting the information and insights they need.
“You’d usually have to have another data engineer to sort out the analytics pipeline,” Rahul added. “Using a data warehouse means that you can just hire one person to manage this one tool and use that raw data.”
Get your engineers and developers talking early
Setting up your analytics can take time. For example, Rahul described how he learnt that it’s important to know in which format your game engine is outputting your data – is it consistent with other platforms? How are the variables coming in?
“You’re probably going to need to clean up the data in some way, because if the strings aren’t in the right format, it’s not going to work,” Rahul added. “An analyst will, of course, need to sit down with the developers and see how these strings are coming in.”
Give your analysts time
Starting early will make sure that your engineer has time to get everything set up properly and to completely understand the tools. But it’s also key to making sure that they can discuss with the developers what’s happening inside your game.
“Put aside time for analysts to learn what your tool does and think through all the consequences of the different data that you’re capturing,” Rahul said. “How it ends up in the final warehouse is really important, because you’ll eventually need to output in some way.”
For example, you might output that data and create a historgram graph needing timestamps. Or you might want to process funnels or design event interaction using a more complex business intelligence tool.
Pick out a few key metrics for your designers
When you’re first capturing data, it’s easy to get overwhelmed. As a designer, it’s important to pick out just a few to start with and start expanding from there.
“The toughest part of using analytics was figuring out where the focus should be,” David Prinsmel explained. “We ended up with about 80% of stuff that wasn’t particularly useful at this stage of the game. Instead, it was much more important to look at metrics like where players were dropping off in the journey.”
It was key for Happy Volcano to gather all that data, but just as important to filter out what was relevant to them. By focusing on a few key metrics, such as retention rates and whether the player joined a multiplayer race, Happy Volcano could spot the hurdles their players were facing.
Start your analytics journey
If you’d like to get started making sure you’re ready to find those insights that will improve your game, check out our SDKs and start using our free tool. And if you’re ready for more advanced analytics, check out our DataSuite.