Household decisions
Why Recommendations Built for One Person Fail a Household
Updated Sep 18, 2026
Most streaming recommendations are built around a single account's history, which is a reasonable approach for one viewer and a poor one for a household of two or more people whose tastes only partly overlap.
A shared account blends two histories into a confused signal
If one household member mostly watches tense dramas like Succession and another mostly watches lighter comedy, a single account's smart recommendations end up averaging the two into suggestions that don't strongly fit either person — the system can't tell whose turn it is to be recommended to.
Individual profiles solve viewing, not picking together
Separate profiles fix the blended-history problem for solo watching, but they don't help when the two of you are choosing something to watch together — neither profile's personalized suggestions are built for "what will both of us actually enjoy tonight," because that's a different question than "what does this one account tend to like."
The overlap between two people's taste is a real, useful signal
Knives Out sits at a genuine overlap for a lot of households — mystery for one person, comedy for the other. A recommendation system built for households would need to actually look for that kind of intersection deliberately, rather than defaulting to whichever profile watched more recently.
What a household actually wants is deeper personalization, applied to the pair, not the individual
The fix isn't fewer smart suggestions — it's advanced personalization that treats "this household" as the unit being recommended to, not "whoever's logged in right now." That's a genuinely different design problem than single-viewer discovery, and most apps haven't built for it.