The argument
A familiar claim runs like this: when people had a few television channels and a fixed schedule, large audiences watched the same programmes at the same time, and the next day they discussed them. Streaming services with huge libraries and on-demand access have scattered viewers across thousands of titles, so there are fewer shows everyone has seen, and less shared conversation as a result.
The claim is plausible and attractive, but it is really a bundle of different ideas: that audiences are more spread out, that fewer titles reach very large audiences, that people have fewer overlapping references, and that this matters for social life. Each can be examined separately.
What “shared culture” could mean
Before testing the argument, it helps to say what is being measured. Possible meanings include:
- Simultaneous viewing: many people watching at the same moment, as with live events.
- Overlap in what people have seen: the proportion of people who have watched a given programme at some point.
- Common conversation: whether people can mention a show and expect others to recognise it.
- Cultural weight: whether a small group of programmes dominates public attention.
These can move in different directions. A hit series might be watched by a very large number of people over months, creating shared knowledge without simultaneous viewing. Conversely, a live broadcast might draw a big simultaneous audience with little lasting conversation.
What the data can show
Ratings and audience measurement. Traditional television ratings estimated the size of audiences for scheduled broadcasts. Peak audiences for individual programmes in many countries were often larger in earlier decades than for typical programmes today, but comparing across eras is difficult. The number of households, the number of channels, methods for counting viewers and the rules for including delayed viewing have all changed.
Streaming service data. Services sometimes announce how many accounts watched a title or how many hours were viewed. These figures are chosen and defined by the company, may use different thresholds for what counts as a view, and usually cover only that service. Viewing across different providers cannot be combined from these announcements, an example of the limits of platform data.
Surveys. Researchers can ask people which programmes they have seen, and measure how much respondents’ lists overlap. This tests the idea of common reference points more directly than ratings do. It relies on memory and honest reporting, and on who is sampled, so features like self-reported behaviour and question wording matter.
Panels and metered households. Linked data on what the same households watch across broadcast and streaming offers the best view of overlap, but it requires large, carefully constructed samples, and not all countries have them.
Social media and search signals. Volume of discussion about programmes can be tracked, as can search interest. These show attention among people who choose to post or search, not everyone, and are strongly affected by what platforms promote.
How the fragmentation idea could be tested
A careful test would try to answer: are audiences more dispersed than before, in a way that cannot be explained by other changes?
- Define the measure. For example, the share of total viewing accounted for by the most-watched titles, or the average overlap between two randomly chosen people’s lists of programmes seen.
- Use consistent methods over time. The measure should be computed the same way for the earlier and later periods, including delayed and streamed viewing.
- Compare like with like. Compare an era with few channels to one with many using the same population, not different countries or age groups.
- Look at other media. If the trend appears in music, books and games too, the explanation may lie in wider changes and not in streaming alone. Related discussion of music is in the article on streaming habits.
- Consider alternatives. Check whether the pattern might be due to rival explanations, such as population growth, the rise of cable before streaming, or changes in leisure time.
Even a clear finding would be partial. Fragmentation of viewing across titles is not identical to fragmentation of culture, since shared references can also come from news, sport, music, memes and social platforms.
Complications and counterpoints
Fragmentation started earlier. The shift from a handful of channels to dozens or hundreds, along with video recorders and later digital recorders, began before on-demand streaming. Attributing change to streaming alone can overlook this longer trend.
Big events still gather audiences. Live sports, major ceremonies and national events often continue to attract large simultaneous audiences. Whether this offsets other fragmentation is an empirical question, and the answer probably varies by country and by genre.
Shared moments can be created differently. Certain series become widely discussed through word of mouth and online sharing even without a common broadcast time. Viral clips and memes can spread a reference quickly, as discussed in the article on viral memes.
Nostalgia can distort memory. People who recall a time when “everyone” watched the same programmes may be remembering the most popular ones and forgetting that audiences then also had many different tastes. Survey recall of past shared experiences is subject to this kind of selective memory.
Algorithms shape overlap. Recommendation systems steer viewers to different titles, which could reduce overlap, but they also push popular titles to many people at once. Which effect dominates is not obvious. The role of recommendation algorithms is a topic in its own right.
Subgroups differ. Averages across a country hide the fact that some groups, for example older viewers or those who watch lots of live television, may have quite different patterns from others. Inferring individual experience from group figures runs into the ecological fallacy.
Reading claims about shared culture
The statement that streaming has ended shared culture is best read as a hypothesis with some supporting signs and some counter-signs, not as an established result. Useful questions are what exactly is being compared, over which period, with what data, and whether the writer has separated a change in how people watch from a change in what they talk about. In most discussions the strongest conclusion the evidence supports is that the ways audiences gather have shifted, and that how much cultural overlap has changed depends on the definition used.