Finding signal in the noise, Notification Platform

It was a well documented problem, Jira's email notifications were 'noisy' and overwhelming users. The goal was to solve that first, then use the learnings to build a self-service notification platform that could scale across multiple products.

Problem

Jira email chattiness had been voted one of the top three pain points on the public-facing support tracker. Reducing volume was a clear priority, but the risk was real, any changes could disrupt workflows that customers had built around existing notifications. The solution needed to be general-purpose enough to serve more than one product from the start.

Early results

An initial solution rolled out to a small cohort of early adopters showed promise. Grouping related notifications into single emails resulted in a 60% reduction in email volume. Feedback from that cohort was largely positive, though the research synthesis surfaced nuance worth addressing, users responded well to structured templates, but output from custom notification prompts was identified as poor.

System design

The platform was built around four core ideas. Related content is grouped into a single notification to improve consumption and coherence, while high-priority notifications are always sent immediately and never grouped with others.

Notification templates were optimised for specific use cases and channels rather than relying on a one-size-fits-all format. A diff model, showing before and after event content gives recipients enough context to act without needing to open the source product.

Outcome

Jira email volume was cut by 60%, meaningfully reducing inbox clutter and lowering operational costs. The improved template design lifted the clarity of email content and drove higher user satisfaction scores.

Beyond Jira, the platform's success drove adoption from additional products, validating the decision to build for scale from the beginning.