HUMAN Blog

How White Ops Uses Snowflake

Written by Ram Narayanan | March 2, 2018

To access the data we need to target and stop bots quickly, White Ops uses Snowflake, a data warehouse built for the cloud.

At White Ops, our mission is to protect the internet by verifying the humanity of every online transaction. But we can’t fight the good fight alone — we rely on our technology partners for a variety of services that make it possible for us to stop bots faster, preventing them from preying on you and your business.

One of our earliest and most valuable partnerships is that with Snowflake, an SQL data warehouse built for the cloud that compiles all the data that we need to store and analyze in one easy-to-reach place. Snowflake’s Elastic Data Warehouse helps us more easily and quickly pull insights from our data and gives our analysts unfettered access to the data they need. Thanks to Snowflake, we’re faster, smarter, and better positioned to help our customers avoid the potentially devastating effects of bots and other fraudulent activity than ever before.

The Need for Speed

Bots and internet fraud are constantly moving targets — for every threat that’s eliminated, a new one pops up in its place. That’s why we’re always monitoring for new and changing threats so we can respond to new patterns the second we notice them. To do this, we need an accessible data storage solution that allows us to process massive amounts of data and generate new detection algorithms in real time.

Snowflake’s Elastic Data Warehouse is exactly that solution. It keeps all of our result data in one place, scales elastically, and queries diverse data with standard SQL. Because Snowflake uses SQL as its core language and is delivered as a data warehousing cloud service, everyone on the team from data scientists, to analysts, to researchers can access our data directly. By providing data access to a wide range of team members, we can move more quickly than its competitors, avoiding roadblocks and generating solutions faster.

Avoiding Bottlenecks

Data solutions we’d used in the past didn’t allow open access. That meant that when a White Ops security engineer faced a big data question, he or she had to send a request to the vendor and wait until a developer could build a custom map-reduce job. This created a productivity bottleneck, keeping our engineers from responding to threats as quickly as they needed to.

We needed a way to simplify and standardize this process to increase the productivity of our research and development teams, and allow us to respond to customers more quickly. Now that our data is easily accessible in Snowflake, our researchers and analysts can turn their insights into action as soon as they surface them, instead of having to queue up queries.

And because all of our data is safely stored in Snowflake, our detection team can easily see and compare ad fraud trends in one place, and even perform statistical studies directly with that data. Ultimately, having better access to data in Snowflake means we can avoid bottlenecks, work faster, and are better able to analyze our own data so that we can combat fraud threats with greater accuracy.

Better Data, Better Fraud Detection

After implementing Snowflake’s storage solutions, we have seen three major areas of improvement: we can write more algorithms that detect ad fraud in less time, we have improved system health and quality of results via our QA process, and we have deeper, more accurate data that is accessible across team members and user groups.

With the help of Snowflake, White Ops has been able to improve our competitive advantage and better focus on our core competency: the fast delivery of a broad spectrum of ad-fraud detection algorithms for our customers.