The data we almost weren't allowed to share

The data we almost weren't allowed to share

The scene: a dating-app client sitting on the richest behavioral dataset of any brief I'd worked, and a legal team that said no to everything. The setup was a goldmine locked behind privacy fears. The angle was to find the one aggregate, fully anonymized cut that legal could bless and journalists would devour: messaging patterns that predicted a second date.

We spent two weeks getting one approved insight rather than fighting for ten. The pitch was a single irresistible finding: 'Messages sent on Sunday at 9pm are 40% more likely to lead to a date than any other time.'

The result: 52 placements, the most of any campaign that year, 41 with links, including a Today show web feature and a Cosmopolitan piece. The trends page drew 21,000 sessions. A lifestyle editor wrote, 'This is the kind of stat people text to their friends, which means I have to cover it.'

Working with legal early, not around them, was the whole unlock.

Lesson: Your richest linkable asset may be one privacy-safe slice of data you're sitting on. Get one insight approved, not ten rejected.