The news: Google blamed a faulty, untested policy update—not overloaded infrastructure—for triggering a Google Cloud outage that took down Gmail, Cloudflare, Shopify, and many others. It admitted to skipping standard risk safeguards, per CNBC.

“We apologize for the impact this has had not only on our customers’ businesses and their users but also on the trust of our systems,” Google posted in its incident report.

The bigger picture: This isn't an isolated event. Last year, a CrowdStrike update bricked millions of Windows PCs globally resulting in $5 billion in cumulative losses. While Google’s was a cloud-level failure, CrowdStrike’s was a local software update that rippled across cloud-reliant systems because of Microsoft’s dominance.

• Google Cloud: Disrupted APIs, cloud storage, and productivity tools—but core OS functions were unaffected.
• CrowdStrike: Took down endpoints and caused entire systems to fail, grounding planes and halting operations like payments and payroll.

These incidents show the dangers of over-centralization. When one key provider fails, damage spreads quickly.

"Cloudflare relies on Google Cloud, and many other services depend on Cloudflare, creating a network of interconnected systems," said Mike Hicks, Principal Solutions Architect of Cisco ThousandEyes. "This can result in situations where users experience outages in services that are not directly connected to the root issue, as demonstrated in this case—users encountered outages in services with no direct relationship to Google."

Why it matters: Both incidents were preventable but worsened by scale, human error, and missing fail-safes.

• Google’s misstep highlights fragility in relying on cloud services when deployment protocols break.
• CrowdStrike’s collapse reveals the hidden risks of third-party agents in critical global infrastructure.

Everything’s connected: Centralized systems, no matter how robust, create single points of failure with cascading consequences. While cloud computing and third-party integrations offer efficiency and scalability, they also introduce systemic risks when governance, testing, or redundancy measures fail.

"Businesses have their brand reputation, revenue and employee productivity on the line every single day. The best possible thing any business can do is to proactively monitor any services they deliver to customers, and all the dependencies that their service might have (think 3rd party APIs, CDNs, etc.) so *when* (not if) something breaks, they’re not caught flat-footed," Hicks noted.

Our take: The next outage is a matter of when, not if—and the time to prepare is now. CMOs and CTO's should pressure vendors for transparency, diversify martech stacks to reduce dependency, and ensure business continuity plans cover cloud failures and system disruptions. 

This content is part of EMARKETER’s subscription Briefings, where we pair daily updates with data and analysis from forecasts and research reports. Our Briefings prepare you to start your day informed, to provide critical insights in an important meeting, and to understand the context of what’s happening in your industry. Non-clients can click here to get a demo of our full platform and coverage.