Research we publish.

Modelled studies run with simulated customers. The brand studies look at public brands everyone knows; the cases are anonymised and built to show the method. None of them are client results.

Brand research

Simulated customers react to brands they already know. Each has a full report.

Consumer brandsLiquid Death · OLIPOP · Spindrift

Identity outsold function two to one, until the signal inverted.

Purchase motivation split 2:1 identity over function, and mass distribution then inverted the signal for the highest-value buyers.

Consumer brandsFIGS · Vuori · Patagonia

Remove one barrier to belonging, and the next one appears.

Identity barriers to purchase don't dissolve, they migrate. Open one gate and the next emerges.

Consumer brandsAG1 · HUM

Retention wasn't the product. It was the ritual.

The thing being bought isn't the supplement; it's permission to stop asking whether it works.

Modelled cases

Anonymised scenarios built to show the method. Not client results.

Apps & subscriptionsA retail bank

The customer most able to pay was the first to leave.

The highest-income saver churned first; the lowest-income one held on. A study of savings identity and autopay retention.

Consumer brandsA health brand's creators

The 8,000-follower doctor beat the 30-million celebrity.

Follower count didn't buy purchase authority. An unpaid clinician scored 83/100; the celebrity scored 31.

Food & retailAn online food listing

One keyword scrolled past 83% of the market.

A single diet keyword in a product title signalled a tribe, and 83% of shoppers who didn't belong scrolled straight past.