Making the impossible, possible: A GameAnalytics case study

57 Million EVENTS PER DAY

3500 QUERIES PER DAY

20% REDUCTION IN ENGINEERING HOURS

We’ve had the pleasure of speaking with Ioana Hreninciuc, CEO of GameAnalytics, to learn just how they use Imply to make their next-generation data stack possible.

For those of you who are new to GameAnalytics, they are currently the number one analytics tool for anyone building a mobile game, from indie developers, games studios to established publishers.

And being one of the leading analytics companies, they have a lot of data to play with. As it stands, GameAnalytics receives, stores, and processes game events from 2+ billion monthly players in nearly 100,000 games. These events, created by their SDK, all pass through a system they call the Data Collection API. This forwards the events to Spark ETL processes, and then through AWS Kinesis directly into Imply, eventually ending up with real-time statistics and graphs on a dashboard. This displays user activity, game revenue, and more.

Simplified GameAnalytics data pipeline, expanded further down on this publication.

But getting to this stage was no walk in the park. I interviewed their CEO, Ioana Hreninciuc, to find out what exactly they did and why, and how they’re now using Imply to help grow their company.

Here’s what I learned.

So Ioana, why did GameAnalytics decide to move technology?

In the last five years, our user base has grown eight times over, and it is now more than 2 billion game players worldwide.

But we’re a fairly small team. We have around 35 people working for GameAnalytics in total at the moment, which isn’t a lot when managing such a large data set. One of the biggest purposes for a tool like Imply is turning all of this data into reports for game developers. We were using Druid open source before. And although this was fine to start with, it became more and more difficult to manage as we scaled.

So we needed to find something that not only would scale with us, but also something that our team could easily work with.

What was your ideal solution?

Three things came to mind when looking for something new. Manageable, scalable, and affordable. Let me elaborate.

For the manageable side of things, we needed something that could help our entire business. From an internal use case, we wanted specific reports which our management, product, finance, and of course, our backend team could use. And for an external use, we needed something for our end-users. We now have 100k developers who need to access data from 2 billion monthly players. And this is still growing.

Which leads on to scalable. I’ve already mentioned a couple of times, but we are growing. And this was the biggest pain point for our last solution. The technology wasn’t able to manage the amount of data we were collecting, as it was hitting its limits in regard to scaling. This also impeded us from developing new, more interactive features.

Lastly, it needed to be affordable. We’re a growing company for sure, but we provide our core software free of charge, and we needed something that wouldn’t break the bank – or at least help us be as cost effective as possible.

Why did you choose Imply?

We actually stumbled across Imply when speaking directly with your CEO, Fangjin Yang. He went over the tool with us and it just seemed to tick all of the boxes. But it wasn’t just one thing that sold it to us. It was the fact that Imply covered so much, including:

• The software features.
• The monitoring features.
• The deployment process.
• The DevOps setup.
• And the fact that Imply Cloud comes with all of that built in.

All of this saved us a bunch of engineering time. It’s easy to buy software, but it’s very hard to buy good engineers. Because Imply Cloud has all of these features built-in, it saves us on the headcount. Which is something incredibly important to us. We have an ambitious roadmap. And everyone needs to be making products, not dedicating most of their time to Ansible, Terraform code, and other operational, non-product related work. Of course, we need all of these tools as well, so it’s very valuable when we get them “built-in” rather than having to develop everything ourselves.

That being said, it did take us a moment to use Imply properly at the beginning. Mainly because you really need to know what you want in order to get the most value from it. Once we took a step back and thought about how we really want to use it, well it was easy sailing from there on out.

How do you get the most of our Imply?

As mentioned before, our entire company gets value out of Imply. I use it every week, easily. Mainly to find out how our top customers are doing. We push snapshots from our CRM into Pivot. And sometimes we’ll use it to answer questions like “Are our customers using this feature, or this one?”.

The business development team tends to get some use out of it, too. We have some pay-per-use products. So for finance purposes, we get our reports from Pivot on some of the metrics that our customers have – and then we’ll bill them based on those. For example, if one of our customers has a MAU (monthly active users) of over say a million that they’re collecting data for, well we’ll charge them an additional fee.

And then for more generally, our management team uses it for various information on our customers. Our product team will use it for feature usage. So Pivot, to us, is a valuable tool adopted by almost everyone in the company. It was an easy switch to make, and has saved us a lot of time, money, and resources.