Many Somalis got specifically 5.2 % Ethiopian admixtures after the new update, which I think is an ancient component they've carried for a long time. Maybe Ethiopians carry a higher percentage of this component, so the system automatically makes that cluster to be of Ethiopian population-like source.
Either way, their algorithm doesn't capture the complexities of the differentiated intricate, yet similar structures of the ancient admixtures in Somali individuals, and between other closely related Northeast African populations.
This is due to negligible research on Horner DNA and the un-sophisticated algorithm of the biotechnology firm. They should spend more money on improving their systems by funding real research rather than what seems to be an expensive yet very effective marketing strategy.
But this suugo-theory is improbable if we apply Occam's Razor.
