Home › Track record

Our track record

A prediction is only worth something if you can check it. Every month we save the brands our screen picked, then come back to see how their US Google searches moved compared with the brands we track but didn't pick.

Live picks

We started saving our picks as they're made in October 2026. Each month's list is saved with its date, and we don't change it afterward.

Reconstructed months

For the months before October 2026, we applied the same rule using today's data, today's brand list and today's US size and store checks. That favors hindsight, so treat these months as a sanity check, not proof. The real test is the live picks above.

In 4 reconstructed months (14 different brands), picks grew faster in US Google searches than the typical brand we track but didn't pick 23 of 39 times. That's not statistically different from chance yet: a coin flip would get about half, many brands repeat from month to month, and the early months did better than the later ones.

Every month's picks

Brands are compared on Google's 0 to 100 scale, which a spring 2026 data break compresses, so small differences are within rounding noise. US searches since = average US Google search interest over the latest 8 weeks compared with the 8 weeks before the pick, each brand on its own scale. Brands with very little search interest at the time of the pick are left out of the score. Data source: Google Trends. Picks are brands doing better than their peers in Korea and still small in the US; see how we pick.