Week 9: Social Watching

Two identical shows, side by side, will still get picked unevenly the moment one of them says other people are watching it. This week is about what a social cue is actually doing when the content itself can't explain the difference.

19 April 2027

Open the slides

This week’s question#

Week 2 already showed that a “popular” badge can shift a pick among three different shows. That leaves an obvious gap: shows differ from each other, so maybe the badge was just nudging you toward whichever option seemed slightly better already. This week closes that gap directly by asking a narrower question: does a social cue still move a choice when the two options are, as far as you can tell, exactly the same show?

A cue with nothing else to point at#

When two options are genuinely interchangeable, any preference for one over the other can’t be explained by the content — there’s nothing about the content to prefer. That makes an interchangeable pair the cleanest possible test of what a social cue is doing on its own. If a crowd-size number or a friend’s name still shifts which option gets picked under those conditions, the cue isn’t summarizing anything about quality — it’s operating as a signal in its own right, one that people treat as informative by default even when there’s nothing behind it to be informed about.

Two different kinds of social cue are worth distinguishing here. A crowd cue (“12,400 people watching now”) reports on a mass of strangers — it works, when it works, by suggesting that popularity itself is evidence of worth. A named-friend cue (“Alex is watching this”) reports on one specific, known person — it works through a completely different channel, closer to trust in someone you know than to trust in numbers.

A familiar scene#

A streaming app shows you two shows you’ve never heard of, side by side, with near-identical descriptions. One has a small label: “Trending now.” You click that one, and if asked why, the honest answer is something like “it seemed more popular” — even though “popular” told you literally nothing about whether you’d enjoy it. The label did the deciding; the show just came along for the ride.

The interactive experiment: Two Identical Rows#

Below are four rounds. Every round shows the exact same two shows, described identically, in the same two slots. Sometimes neither is flagged. Sometimes one carries a cue — a crowd-size number, or a friend’s name — and which slot carries it changes between rounds, so this isn’t measuring which side you tend to click.

What just happened?#

• Round 1 had no cue on either option — a baseline reading of which slot you reached for with nothing else to go on.
• Rounds 2 and 4 carried a crowd cue, on opposite slots. If you picked the cued option in either round, that’s the crowd-cue effect on its own, since the two shows underneath it never actually differed.
• Round 3 carried a named-friend cue. If your pick changed here compared to the crowd rounds, that’s a difference between “many anonymous people” and “one specific person you’d recognise” as social evidence — two distinct cues, not one general “social proof” effect.

That’s this week’s question, answered as cleanly as a two-tile demo can manage: the shows never changed, the slots alternated, and whatever moved your pick had to be the label sitting next to an otherwise identical option.

What this demonstration can’t tell you: four picks from one visitor aren’t a controlled study, and picking (or not picking) a cued option here says nothing about how “susceptible to social proof” you are in general. What repeated, controlled studies of social proof (not this four-round demo) actually show is that both crowd cues and known-person cues shift choice rates on average across many people — not that any single round proved something stable about your own judgment.

Reflection#

Think of a show, restaurant, or product you picked mainly because you saw a popularity signal — a rating, a “trending” badge, a friend’s mention — rather than because you’d evaluated it yourself. If that same option had been presented with no signal at all, sitting next to something equally unfamiliar, do you think you’d have picked it anyway?

Next episode#

This week was about a social cue that costs you nothing to follow. Next episode is about a cost that’s much harder to wave away once it starts compounding: what watching trades away when it runs up against a full night of sleep. That’s next episode’s question.