Scaling paid search automation for effective search campaigns
Previously, PPC Samurai hosted its website in a data center located in France to meet General Data Protection Regulation (GDPR) requirements. Unfortunately, this approach meant latency problems for users outside of Europe who were routed through France, regardless of their location. The company has since migrated all of its data to Google Cloud.
“By migrating to Google Cloud and applying Cloud Load Balancing to divert traffic between EU and US regions, we reduce our worst case latency from 400 milliseconds to 200 milliseconds, which means faster PPC workflow automation for digital agencies,” says Povey. “The increasing influence of ecommerce is driving demand for our PPC management platform around the world. Google Cloud provides a scalable infrastructure for us to deliver consistent app performance to digital agencies everywhere.”
From a development standpoint, Google Cloud helps PPC Samurai to scale up services more quickly. Povey adds that he can provision compute resources as needed on Google Cloud to build new features such as an inline action menu in the dashboard instead of waiting two weeks for a server from the legacy data center.
“Additionally, rolling deployment on Google Kubernetes Engine (GKE) helps us move from approximately five-minute downtime per upgrade to zero downtime, which means the application is available to customers more than 99.95% of the time.”
PPC Samurai hosts its web application backend on GKE and stores 5 TB of ad performance data such as clicks and impressions on the Cloud SQL database.
Managing app performance for customers
In its early stage, PPC Samurai participated in the Google Cloud for Startups program and used Google Cloud credits to deploy a test cluster for its US region. Results from the two-week trial convinced Povey that Google Cloud is the right cloud platform for PPC Samurai.
“From its user interface to API support, Google Cloud offers a great consistent experience compared to other cloud providers,” says Povey. “For example, we can manage all aspects of our Google Cloud workloads on the Google Cloud Console, instead of a different login for each service.”
Povey adds that error reporting and log monitoring in the operations suite was a welcomed improvement to prevent downtime for users. PPC Samurai uses Google Cloud Marketplace to proactively detect bugs and create alerting policies so that the system is always running smoothly. In addition, the team uses Cloud Logging as a health check for low memory states and restarts clusters that are consuming too much resources without affecting operations.
Povey adds, “Just as PPC Samurai helps PPC experts save time with automated workflows, Google Cloud simplifies our cloud management workflows with Cloud Logging and Cloud Monitoring, so we can spend time resolving business-critical issues.”
Keeping data up to date for ad performance insights
Ad performance insights helps digital marketers track how each search campaign is performing and what needs to be improved to meet client goals. With digital agencies managing tens and thousands of accounts, it has never been more important to have an intuitive platform to work with. Setting up a workflow on PPC Samurai is as simple as a few clicks. However, on the backend, each workflow may require multiple query activities.
A workflow to reduce wasted ad spend, for example, requires calculating the ratio of cost of converting data versus non-converting data based on conversion metrics such as ad group cost, keyword cost, or search query cost. This is a laborious task if done manually.
To help PPC managers use their time wisely, PPC Samurai uses Cloud SQL to update and process data from platforms such as Google Ads by six AM so clients can start their business day with insights from fresh data. In addition, the company stores the historic data for up to three years, so agencies can analyze past ad performance and build relevant workflows.
“We lacked scalable storage in our legacy environment. As a result, we frequently ran out of disk space, which meant risking critical downtime. We also did not have great visibility into database-related performance issues for our app users,” says Povey. “On Cloud SQL, we use Insights to identify performance issues and operate as efficiently as possible. This has allowed us to halve the number of CPUs required in our Cloud SQL instances.”