View project Drag

Bright Box

This is our approach to rapid, automated and cost-effective testing and iteration of content.

A combination of production software and machine learning, along with human insight and oversight, Bright Box has been developed and optimised over several years. Going way beyond simple A/B tests, Bright Box efficiently and rapidly scales the deployment of potentially dozens of ad and audience variants.

Why? Because the reality is that there are simply too many variables to accurately model and predict how audiences will respond to our communications. As a result, clients sometimes feel pressured to pick the lowest risk option, which might well be the least noticeable and memorable.

Market research can help, but participants often struggle to identify or articulate their implicit feelings, while The Hawthorne Effect describes how differently we behave when we know we are being observed.

Bright Box instead tests in live, real-world environments, harnessing the instant and often under-utilised feedback loop afforded by digital channels - what we call ‘signals’ - to test, iterate and improve on risky fire-and-forget marketing.

Even small campaigns often benefit markedly from limited, controlled and inexpensive trials. While bigger campaigns offer scope for continual in-flight iteration, day-by-day and even hour-by-hour.

With Bright Box, we’ve invested heavily in software and data, so our clients don’t have to. Get in touch to find out more.