Contact us
Sign in

Case Study

Playgendary cuts BigQuery costs and gains better insights into usage with BigQuery Lens

Client
playgendary-logo
Playgendary
Industries
Gaming
Technologies
BigQuery, Compute Engine, Dataflow
Location
Cyprus

Mobile game developer slashes BigQuery spend by over 50% while gaining high-impact insights into its BigQuery usage.

As a mobile game developer with over 3 billion installs and 250 million monthly players, Playgendary knows a thing or two about user acquisition. A critical component of their user acquisition strategy is Google BigQuery, which they use to evaluate the effectiveness of their marketing campaigns.

The brief

Playgendary launches many marketing campaigns, aiming to drive downloads and increased usage of their games. The responsibility of determining whether a campaign succeeded or not falls on Mikhail Artyugin, who leads their BI team. His team does cohort analysis using various factors like device type and registration date, then enriches that with user event data — levels completed, purchases made, etc. — streamed into BigQuery via Dataflow.

While Mikhail found BigQuery fast and performant, he couldn’t easily understand how their costs were broken down at a granular level. Managing a team of data engineers and analysts, he found analyzing job-level data with SQL too time-consuming to do on a regular basis. “Every time I have an idea of what I want to check, I have to write a query on my audit logs. I love SQL but I don’t think it’s a good way of spending my time. Also sometimes I want to analyze trends and not just calculate aggregations for a period, which you can’t easily do with SQL.”

And as their largest service cost, it was important for Mikhail to understand how the BI team was using BigQuery and what could be done to optimize their spend.

 

Playgendary 1

 

Additionally, like many gaming companies, Playgendary faces fluctuation in their compute usage due to unpredictable variables like user activity and game popularity. As new games get introduced, it’s difficult to predict what their spend will be because that is dependent on the game’s popularity. This made it difficult for Playgendary to purchase Committed Use Discounts (CUDs). Weighing the tradeoffs between on-demand flexibility and potential savings from CUDs, Playgendary opted for the former. However, they were hoping to find other methods to optimize their Compute Engine costs.

Overall, Playgendary was looking for a partner to not only help them better understand and optimize their infrastructure costs, but also someone they could brainstorm with around making their infrastructure more performant.

What we did

Leveraging DoiT’s product portfolio and on-demand access to cloud expertise, Playgendary was able to better understand and optimize their biggest cloud cost drivers.

Understanding and optimizing BigQuery costs

Mikhail used the BigQuery Lens to understand how costs were broken down, and which parts of their BigQuery usage should be optimized first. For example, he tapped into BigQuery Lens’ recommendations to identify large tables that went unused for several months, and promptly removed them.

 

Playgendary img 3

 

He also used BigQuery Lens’ Explorer to identify and optimize the most expensive queries per table and user — without having to write any queries himself! After identifying the queries, he copies them into the BigQuery console and examines the query execution flow for the issue — sometimes a bad JOIN or absence of predicate filters.

“BigQuery is a critical component of our cloud infrastructure, but understanding how we could use it more optimally was difficult. Without BigQuery Lens, I wouldn’t have been able to achieve any significant results around cost optimization. The easy-to-use drill down into my team’s BigQuery usage and personalized recommendations made optimizing how we use it really easy.”

 

Playgendary img 3

 

To further optimize their storage costs, Mikhail worked with DoiT Senior Cloud Architect Rajan Bhave to understand the pros and cons of using BigQuery’s new Physical Storage before ultimately deciding that they’d save money by switching.

All in all, in just one month Playgendary saw BigQuery costs fall by over 50% after implementing these optimizations and changes in behavior.

 

Playgendary img 2


Automating Compute Engine savings with Flexsave

To optimize their Compute Engine spend, Playgendary enabled DoiT Flexsave™ and began realizing 1-year commitment discounts on their on-demand compute workloads. As a result, they’ve reduced their overall Compute Engine spend by 25% without sacrificing the flexibility of on-demand pricing, and without any tedious operational or management effort on their end. Most importantly, Playgendary realizes these savings without needing to predict the future success of any new games introduced.

Playgendary case study

The result

The Result

  • Reduced BigQuery costs by over 50%
  • Greater visibility into team’s BigQuery usage and behavior
  • Saved 25% on Compute Engine costs
Mikhail Artyugin, Director of Business Intelligence, Playgendary
“BigQuery is a critical component of our cloud infrastructure, but understanding how we could use it more optimally was difficult. Without BigQuery Lens, I wouldn’t have been able to achieve any significant results around cost optimization. The easy-to-use drill down into my team’s BigQuery usage and personalized recommendations made optimizing how we use it really easy”

Related case studies

see all case studies
Let’s do it

From cost optimization to cloud migration, machine learning and CloudOps, we’re here to make the public cloud easy — without the costs.

Ready to get started?

Get in touch

Company

Offering

Support

Never miss an update.

Subscribe to updates, news and more.