The context:

Confused.com is a UK-based insurance comparison platform serving millions of customers across motor, home, and van insurance products. Despite having sophisticated audience segmentation and decisioning engines, the company faced a critical challenge: scaling personalised content to match the granularity of its data capabilities.

The challenge:

Confused.com was hampered by a fundamental scalability problem in its email marketing operations. While the platform possessed incredibly sophisticated decisioning engines capable of segmenting audiences down to the individual level, it was still sending the same handful of generic messages to millions of customers.

The core issue was operational and technical. With a small team of copywriters responsible for millions of emails across multiple insurance products, and with each conversion representing hundreds or thousands of pounds in lifetime value, traditional manual A/B testing simply couldn’t deliver meaningful personalisation at scale. Testing five subject line variations over a fortnight left enormous performance gains on the table.

The stakes are uniquely high in insurance marketing. When a customer receives a quote but doesn’t convert, it’s not just one lost sale but potentially years of lifetime value lost. In motor insurance alone – where Confused.com processes a high volume of quotes daily – even small improvements in conversion rates compound rapidly.

Confused.com needed a system that could test thousands of variations, identify successful linguistic patterns that consistently drive engagement, and apply those insights systematically across every campaign, trigger, and customer touchpoint.