Prism is a fictional streaming service, and so are its rivals Meridian and Halo. This shows the kind of insight a Category Read gives Prism's head of insight. A real one is delivered as a debrief built around your business question and sized to your category. This one draws on 38,400 posts and comments about choosing what to watch.
The business question Why isn't Prism part of more evenings?
That's the headline from 38,400 unprompted posts and comments about choosing what to watch. Prism is the name people give a friend. Meridian is the one they talk about putting on the rest of the week. Three findings explain the gap, then we recommend what to do about it.
The three takeaways
Prism leads the talk about nights people plan around. That's the smaller share.
Event nights are 31% of the talk. Habit nights are 58%.
Next move Test a homepage row of shows that are easy to put on.
If viewers rewatch Prism shows, they aren't talking about it.
14% of Prism posts mention a rewatch. Meridian gets 61%.
Next move Check Prism's own data on what happens after a finale.
Prism shows get recommended a lot. Lack of access is a big blocker to following through.
12% of replies say they've started. Meridian gets 29%.
Next move Make the first episode easy to start from a shared link.
We sorted the conversation by the kind of evening each post described. Two of the five are about starting something new, and Prism is named most in both. The biggest by some distance is simply putting something on, where people describe reaching for a show they already know. Pick an occasion to see who gets named.
Where does Prism come up when people decide what to watch?
of the talk about what to watch
"i know every single line of this and i'm putting it on anyway"
"a new show is a first date and i am in sweatpants"
TikTok comment"putting something on is just scrolling while the tv talks"
Instagram comment"whoever picks has to watch what the other one says fine to. it's been three weeks of fine"
TikTok commentAn imagined set of 38,400 public posts and comments about choosing what to watch, over 30 days, US and English-language. The size of a real read depends on the category and the question. In a real read, researchers code each post to a single occasion and attribute every show named to the service it streams on. Every share here comes from that illustrative set.
Our first reading treated the 41% as Meridian's biggest win. The posts made us less sure. A lot of putting something on is a show playing while someone scrolls (the “tv talks” comment is typical), which is attention of a fairly thin kind.
Meridian seems to win these evenings by being easy to leave on. A better Prism show probably wouldn't change that. A different way of presenting the catalog might.
Prism has a real reputation for the night a new show lands, and people seek it out for that. But those evenings are the smaller part of how people talk about watching. On an ordinary night, the posts describe tiredness, compromise and something familiar playing in the background, and Prism's shows rarely come up.
Whether a tired viewer opening Prism finds anything easy to put on. That's a question about how Prism presents its catalog, and it's where our first recommendation starts.
We took every post that named a specific show and checked one thing: does it mention watching it again? Posts about Meridian shows often do. Posts about Prism's mostly describe getting through a series, sometimes in a single weekend, and then moving on.
What happens to a Prism show in conversation once people have finished it?
Share of coded posts about a named show that mention watching it, or part of it, again.
"signed up for copper lane, watched copper lane, cancelled. sorry prism"
TikTok comment"the quiet hour is on every night in our house and nobody is watching it. it's furniture"
TikTok comment"nightjar was the best thing i watched all year and i never want to see it again. in a good way??"
Instagram commentPosts in the illustrative dataset that name a specific show, coded for whether they mention rewatching, going back to a favorite season or leaving it on again. Rewatch mentions run Meridian 61%, Halo 38%, Prism 14%.
Two things. Mentioning a rewatch isn't the same as rewatching, and saying nothing about it isn't proof someone never went back. And easy to leave on isn't the same as good. The Quiet Hour comment describes a show as furniture.
The Nightjar post is one of the warmest in the dataset. Never wanting to see it again is praise here. So we read this as a difference in the role each catalog plays in people's lives, not a verdict on quality.
What this can't tell usWhether anyone cancels, churns or rewatches. Those are viewing and subscription questions, and the answers sit in Prism's own data.
Prism's originals seem to work as events. People start them on a recommendation, finish them quickly and remember them warmly. Meridian's shows play a different role: they sit in the background of people's evenings for years. Neither is better in itself, but a show people return to keeps giving them a reason to open the app, and the conversation doesn't suggest Prism's shows are doing that.
This is a hypothesis Prism can check and we can't. Does viewing, or subscribing, fall away after a finale? Prism's own data would show that quickly. If it does, the conversation gives some clues about where to start. If it doesn't, the difference here may be about genre more than loyalty.
38% of the shows named in a recommendation are Prism's, more than either rival. So we read the replies directly underneath. Under Prism recommendations, 12% say they've started. Under Meridian's it's 29%, and under Halo's 24%. Most of Prism's replies are people working out whether they can watch it at all.
What happens between someone recommending a Prism show and someone starting it?
How the replies under a Prism recommendation break down, as shares of coded replies. More than half ask where to watch or mention not being subscribed. That points to access rather than a lack of interest.
"where is it on?? i'm not signing up to another thing for one show"
TikTok comment"is it worth a month or am i being dramatic"
Instagram comment"adding to the list. the list is where shows go to die"
TikTok comment"ok started ep 1 at midnight because you said it would ruin my week. it did"
TikTok commentReplies sitting directly under a recommendation of a named show, coded by what each reply does: asks where to watch, says they're not subscribed, says they'll wait or add it to a list, says they've started, or something else.
Some “where's it on” replies are just chat, and plenty of people adding a show to their list probably mean it.
What we're comfortable saying is narrower. Prism recommendations produce fewer “started it” replies than its rivals', and more than half of Prism's replies are about access.
The recommendation is doing its job, and people are clearly interested. The hesitation comes afterward, when watching means finding the show, checking which plan it's on or paying for another month. “Worth a month?” came up often enough that we think it's the question sitting under a lot of these replies.
Whether making the first episode easier to reach changes what happens next. It's the cheapest of our three recommendations to test.
On an ordinary weeknight, people describe reaching for something familiar. Test a homepage row built for that moment: shows that are easy to drop into, episodes that work on their own, nothing that needs a free weekend. Then see whether it gets used on the evenings Prism currently misses.
Owner · Product and merchandisingHabit evenings make up 58% of the conversation about choosing what to watch. On the biggest of them, putting something on, Prism has 6% of the shows named. Meridian has 52%.
Prism's catalog is built around big originals. If there's little in it that works as background viewing, a new row will only show the gap, and the real answer is a commissioning one.
Compare weeknight use of the new row, and what people go on to watch, against the current homepage.
The conversation suggests Prism's shows get finished and put away. Prism's own data can show whether interest falls away after a finale. If it does, look at what makes Meridian's shows easy to return to, and what Prism could borrow without turning its originals into background TV.
Owner · Insights, commissioning and merchandisingOnly 14% of posts about Prism shows mention going back to them, against 61% for Meridian. People describe Meridian shows as on while they cook, or on a ninth time round, and Prism's as finished in a weekend.
The gap may simply reflect the kind of shows each service makes. Prism's originals are what make it distinctive, and copying Meridian's comfort catalog would trade that for a fight Prism starts years behind.
Look at what people who finish a Prism original watch in the following weeks, and read what they say once they've finished. Any reason to come back would start there.
Take the ten Prism titles that come up most in recommendations and make the first episode easy to start from a shared link. Run it with a holdout, so you know what removing the friction is actually worth.
Owner · Growth and distribution38% of shows named in recommendations are Prism's, but only 12% of the replies underneath say they've started, against 29% under Meridian's. More than half of Prism's replies are about where to watch or not being subscribed.
A free episode may mostly attract people who were never going to pay. Keep it to ten titles, not the catalog.
Compare starts and sign-ups from linked recommendations against a holdout set of titles.
Researchers framed the business questions, built the vocabulary and chose the sources. They also ruled on what counted as relevant, which takes judgment, not a filter setting.
Occasions and reply types are defined and tested by hand before anything is counted.
AI sorts the posts and comments against that frame and surfaces patterns and outliers for the team to check. It's what makes reading at this volume practical.
It doesn't decide what a pattern means.
Researchers go back to the posts behind each pattern, argue with their first reading and decide what the evidence can support. That's where the recommendations come from.
In a real read, every figure traces back to a coded post.
This example imagines a read built entirely on public conversation. Its figures are shares of an invented coded dataset, created to show how a Category Read works.
A read like this doesn't need any of Prism's own data. Recommendation 02 is a question for that data, and we'd treat it as a separate next step.
Selected public conversation, scoped to the brief, where those sources are supported and configured. Organic Intelligence runs on its own or alongside qualitative, quantitative and brand tracking work.
A real Category Read starts with what people say when nobody has asked them, and ends with things worth testing, checking or doing.
Prism, Meridian and Halo are fictional streaming services created by Basis for this example. Every figure, quote, show title and performer name here is invented and illustrative. Figures are shares within a fictional coded conversation dataset, not estimates of consumers, viewing or business performance.