Celtra, the data-driven ad platform for two-thirds of the Fortune 500, processes over 2 billion analytics events daily. Join Celtra Director of Engineering and Analytics Grega Kešpre as he shares hard-won insights about harnessing the cloud to deliver a world-class data processing pipeline.

Keeping up with your data is do-or-die. For Celtra, a creative and analytics platform for rich media mobile display ads, that means creating, trafficking, tracking, and optimizing $400M in mobile and cross-screen media buys for over 2,500 top brands.

As the company continues growing, Celtra Director of Engineering and Analytics, Grega Kešpret, has not only stayed a step ahead with their data analytics needs, but has built a future-proof solution.

Two or three years ago, Kešpret notes, they hadn’t reached the volume of data they work with currently. “But we started slicing and dicing massive amounts of data,” he says, “and we started doing more with it, developing better reports for our clients as they asked for them.”

The problem, he says, was on the one side, messy, unstructured event data. On the other, precomputed, preaggregated client reports. “But we didn’t have the middle point, the data warehouse in between, and it became a huge burden for us,” Kešpret explains.

Choosing a cloud-based solution was a several-year project, he says, and they went through several tries before landing on their final choice (Snowflake).

Too many companies try to cobble something together from various open-source solutions to meet their needs, or torture the in-place solution into shape so that it works somehow, he says, before realizing they need to start over with something actually intended for their needs.

The biggest challenge is the abundance of solutions out there. From a very narrow marketplace of just a couple of vendors, the market has exploded.

“It’s hard to navigate through this space to develop the correct, most cost-effective solution,” he says. “Do you need fast answers, predictive analytics, scans of huge amounts of data?” You need to know exactly what you need your data to do, and where you need it to go, and to whom.

“At the end of the process we understood our requirements really well,” says Kešpret. “We could filter out a lot of data warehouse solutions that wouldn’t fit our needs.”

Cloud solutions have lowered the barriers for new companies, he notes. “You can have five or ten guys develop Instagram or Snapchat,” Kešpret says. “And then you have a service that has millions or billions of users that needs to work with a lot of data — and ten years ago this was not possible.” Back then, you’d have to invest in the bare metal, the hardware, and the resources, including dedicated, on-site personnel.

Today, he says, you can use the cloud to give you all that. “If you have the knowledge, you can automate. And you don’t need as much investment initially to make it possible,” says Kešpret. As a result, data warehousing and analytics has been hugely democratized.

Join Kešpret and Jon Bock of Snowflake to learn how to build a data warehouse solution that supports growth and scales as you go.

Don’t miss out!

Register here for free.

In this webinar you’ll learn:

  • How to build a simpler, faster solution to support your data analytics
  • Diverse reporting and ad hoc analytics in one system
  • How to take advantage of the cloud for flexibility, scaling, and simplicity


  • Grega Kešpret, Director of Engineering, Analytics, Celtra
  • Jon Bock, VP of Marketing and Products, Snowflake
  • Evan Schuman, Moderator, VentureBeat

Register today and learn how the top SaaS strategies can streamline your business.