Author picture Phil Leggetter

How Hookdeck unlocked 100x event scale at Churnkey

Published


Churnkey key facts

Churn can make or break a subscription business, but it rarely happens without warning. Early signals can often identify customers at risk of leaving before they cancel their accounts.

The Churnkey platform actively monitors customer activity, offering early warnings when customers are at risk of leaving. By automating the detection of at-risk customers, resolving issues like failed payments, and providing actionable insights, Churnkey helps businesses understand and address the factors driving customer loss.

To deliver these insights and interventions, Churnkey ingests large volumes of data from sources such as billing systems and customer activity streams. However, as their customer base has grown, so has the complexity of managing these data flows with their homegrown solution. This led the Churnkey team to seek a scalable, reliable solution that could handle the load and allow them to focus on their core mission: helping businesses stay ahead of churn.

The challenge

At the heart of Churnkey's ability to predict and prevent churn is a data science platform that analyzes hundreds of signals from sources such as Stripe, Chargebee, Paddle, and Braintree. This allows Churnkey to collect large volumes of data on payments and customer interactions, generating real-time insights that help them proactively drive retention for their customers.

When Churnkey first launched, co-founder Nick Fogle and his engineering team built a homegrown webhook solution to handle data from their billing systems and customer activity streams. While the system worked well initially, Churnkey's rapid growth—from 80 to over 700 connected businesses—put a strain on its capacity.

In particular, the Churnkey team faced several challenges:

  • Scaling limitations: The existing solution wasn't designed to handle the massive influx of data that came with scaling to enterprise-level customers.

  • Maintenance burden: Engineers were being pulled away from their core responsibilities to manage and maintain the webhook ingestion infrastructure, slowing product development.

  • Reliability concerns: Without robust retry mechanisms and observability, there was always the risk of missing critical events, which could compromise Churnkey's ability to deliver real-time insights to customers.

  • Inefficient data handling: Churnkey's homegrown webhook system couldn't filter incoming events, so every event was passed to their Lambda functions and EC2 instances. This added unnecessary complexity and processing to their infrastructure, making it difficult to scale while maintaining performance.

As Churnkey continued to grow and integrate with more billing providers and data sources, it became clear that their internal system could no longer keep pace. The team needed a solution that could manage increasing complexity, reduce the maintenance burden, and allow them to get new features to their customers quicker.

The solution

Faced with the growing complexity of their webhook management needs, Nick and his team recognized that they needed to bring in a specialist platform to replace their in-house webhook system. This decision reflected one of Nick's core principles: knowing when it makes sense to build versus when it's smarter to buy. For him, the choice was clear—focus the team's energy on solving customer problems, not infrastructure.

We could build something ourselves, but it would probably be an inferior solution, and then we'd be stuck maintaining it long-term. That's not where we want to be. I'd rather keep the team focused on delivering real value to our customers, instead of managing infrastructure problems that have already been solved.

With their infrastructure already on AWS, the team initially looked at Kinesis and EventBridge. However, both solutions proved to be more complex than Nick had hoped. This complexity undermined the goal of using an existing solution since it still required extensive integration and maintenance. Additionally, both solutions lacked key features like retry mechanisms and easy observability, which Churnkey would have to build themselves.

That's when Nick discovered Hookdeck.

We didn't want to end up in a situation where we were spending just as much time maintaining a third-party solution as we would our own. Hookdeck offered exactly what we needed—scalability, reliability, and observability, without the overhead.

Nick began by signing up for Hookdeck and testing it with webhooks from their then-marketing automation platform, HubSpot.

It took me maybe 30 minutes to set up, and I was like, 'Okay, this is cool.' I could immediately see the value. It was a no-brainer.

Following that initial experiment, Nick had the confidence to migrate customer-critical events to Hookdeck. Starting with Stripe, the team soon integrated other billing providers used by their customers, including Chargebee and Paddle, with Hookdeck.

Adopting Hookdeck meant that the Churnkey team both had the confidence to scale to meet the needs of much larger customers and freed them to focus on feature development rather than infrastructure management.

The result

Hookdeck cut unnecessary inbound events by 50%

Now that they had a reliable platform for ingesting events, the Churnkey engineering team could turn their attention to improving efficiency. That's when they realized that a significant portion of the webhook data they were processing wasn't needed.

We found that about 50% of the events hitting our infrastructure weren't useful. Largely, that's because many partners don't give us the granularity that we need. So, before Hookdeck we were ingesting a lot of data that we then had to dispose of.

The excess load not only strained Churnkey's infrastructure but also led to higher cloud costs.

Hookdeck's event filtering helped solve this issue. By discarding unnecessary events before they hit Churnkey's AWS infrastructure, the team reduced cloud costs and simplified their filtering logic.

Hookdeck helped Churnkey scale to meet a new level of demand

With Hookdeck managing event ingestion, Churnkey was ready to take on the larger volumes of data that came with bigger clients.

This rapid growth introduced new challenges in managing unpredictable data volumes.

The amount of data we had to handle grew a hundred times as we moved upmarket. We had no idea how big the volume would get or how fast it would grow, but thanks to Hookdeck, we were able to increase our throughput and serve those clients without any missteps.

Enhanced reliability

Churnkey processes time-sensitive data, leaving little room for missed events. With so many moving parts, occasional webhook handling failures—whether from traffic spikes or system issues—are inevitable.

Hookdeck's automatic retries with exponential backoff ensure that Nick and his team can trust every event will be delivered in time, without overwhelming systems that are offline or facing issues.

Hookdeck gives us the confidence we need to uphold our customer promise of real-time insights.

The impact of Hookdeck's real-time observability

And if things do go wrong, the Churnkey team must be able to detect and respond to problems before they impact customers. That's where Hookdeck's end-to-end observability of the event lifecycle comes in.

We can see potential problems in real-time and get ahead of them before they affect customers. The visibility Hookdeck gives us has been crucial for keeping everything running smoothly.

But it's not just about passive monitoring—Hookdeck's Issues and Notifications feature alerts the Churnkey team on Slack the moment an issue is detected. “The Slack integration is incredibly helpful,” Nick noted. “We get immediate alerts, which means we can respond at the moment something occurs.”

Simplifying development and testing with webhooks

Outside of production, Hookdeck plays a role throughout Churnkey's software development life cycle:

Development: Hookdeck's CLI helps the Churnkey team receive webhooks locally, saving time during development.

We used ngrok tunnels before, but Hookdeck's CLI really stepped things up—it's faster, more reliable, and makes it easier to onboard new developers.

Testing and staging: Isolated environments let the Churnkey team develop and test new functionality without risking their production application. (See Hookdeck Projects to learn more).

It gives us peace of mind knowing we can test new integrations without touching production—it definitely reduces the risk of things going wrong when we go live.

With Hookdeck, the Churnkey engineers have solved their immediate scaling and reliability challenges. Now, as they look to the future, the team is focused on expanding the platform's capabilities even further.

The future

As Churnkey continues to onboard more customers to its automated retention platform, the challenge goes beyond just handling a higher volume of events. With a growing and diverse customer base, there's a greater chance that new clients will rely on data sources and payment systems that Churnkey hasn't integrated with before.

With Hookdeck, it's a lot simpler to handle new data sources when customers ask us. We don't have to build custom integrations. Instead, Hookdeck gives us a standard way to integrate any data source, so we can scale and onboard faster without the usual headaches.

Scaling 100x while reducing costs with Hookdeck

Whether scaling for future growth or serving their hundreds of current customers, Hookdeck gives Churnkey a stable platform, freeing them to focus on delivering value instead of managing infrastructure. Here's how Hookdeck supports Churnkey's operations:

  • Seamlessly handled 100x more data as Churnkey onboarded enterprise-scale customers.

  • Cut unnecessary inbound events by 50%, reducing both cloud costs and server strain.

  • Ensured reliability through automatic retries and real-time observability, preventing event failures from impacting service.

  • Simplified and derisked development and testing with Hookdeck's CLI, and organizations and projects.

  • Enabled rapid onboarding of new integrations without the need for custom infrastructure.






Try Hookdeck for free or share your event processing requirements and book a demo

The Hookdeck Event Gateway