CBS has announced it’s cancelling The Late Show with Stephen Colbert, and The Daily Show is rumored to be next. I remember when a fan petition drive saved Star Trek TOS for a season. That won’t be enough now, IMO. The official story is low ratings. Watching late night shows live on broadcast is not practical for me, particularly 2 or 3 at the same time. I was wondering if DVR views are counted as ratings? If so, could mass distributed cheap DVRs, say built on something like a Raspberry Pi, be used to bump the ratings enough to encourage CBS/Paramount/Skydance/Engulf and Devour to keep the shows on the air?
Opine away, fellow SDMBers! Thanks! And thanks in advance to the mods, who might have to move this thread across fora.
In general, my understanding is that no, watching a show later on DVR is not counted in the ratings but, with some caveats:
Late night is late night no matter the channel watched so the rating should be compared to other shows in the same time slot and same day (i.e. lots of people DVR late shows because they are late…whatever the show).
Some DVRs can report back viewing habits if they have an internet connection and time-shifted viewing can be tracked if they want to (e.g. “Live + 3” (viewing within 3 days) and “Live + 7” (within 7 days)).
Streaming services can (and do) capture precise viewing metrics (things like Hulu Live or YouTube TV).
How all those different metrics are combined to give an “answer” the executives use to assess the success of a show is some secret sauce I know nothing about (and it is important since it is a big part of setting the costs of advertising on the show).
ETA: FWIW:
According to LateNighter.com, citing Nielsen ratings, CBS’ The Late Show with Stephen Colbert topped the 11:35 pm hour in total viewers with an average of 2.417 million across 41 first-run episodes. The Late Show was also the only program to show an increase over the first quarter, with the show up 1%. In the coveted 18–49 demo, Colbert brought in 219,000 viewers.
ABC’s Jimmy Kimmel Live! came in second for the 11:35 pm slot with an average of 1.772 million viewers; however, the show edged out Colbert in the demo with 220,000 viewers. Meanwhile, NBC’s The Tonight Show Starring Jimmy Fallon finished in third with an average of 1.188 million viewers and 157,000 in the demo. - SOURCE
I think they count DVR downloads somehow but you’d need millions of devices from distinct IP and even if you did they are smart enough to realize that something fishy was up.
Colbert will land somewhere and you’ll be able to watch him.
To be precise, the official story is “a financial decision,” though it’s true that the show, as well as pretty much all late night shows, has been hemorrhaging viewers for years.
Only if you’re a member of the Nielsen panel, and time-shifted viewing of a recording is counted separately from the main (live) rating by Nielsen. As @Whack-a-Mole already noted, Nielsen counts “Live+3” and “Live+7” ratings, for members of their panel who watch a recorded show within 3 days, and 7 days, respectively, of the live airing.
A random person watching the show on their cable box, or on a DVR, or on a thousand DVRs, has zero impact on the ratings.
Only in the sense that all television shows have been hemorrhaging viewers for years. There are less people watching broadcast television There are less people watching cable television. There are more people watching everything on streaming services.
Indeed, and YouTube was one of the reasons why CBS was reasonably happy with James Corden’s late-late show, even though his linear ratings were lousy; Corden’s “Carpool Karaoke” and other bits went viral on YouTube.
The methodology for ratings of broadcast television is basically a survey of a random sample by Nielson. It’s not like a streaming service like Netflix where they know exactly who is watching what when. They might use technology in the homes to collect the data (many years ago you would fill out a paper form), but it’s still just a sample.
AIUI, allegedly, Colbert’s show cost $100M/year to produce (w/200-ish employees) and generates about $60M/yr in ad revenue for the network, leaving a loss of $40M/yr.
Also allegedly, a mere single-digit percentage of the viewers are in the key demographic of ages 18-49. The rest skew much older.
This is a problem, and – as said – it’s not just a Colbert problem.
Not just “allegedly.” The article shared by @Whack-a-Mole earlier (linked again below) shows the audience sizes for each of the late-night shows from Q2 2025, and has breakouts for their 18-49 audience. Doing a little math, the percentage of the total viewers who are age 18-49 for each of the shows:
Colbert: 9%
Fallon: 13%
Kimmel: 12%
Gutfeld: 7%
Of the three on the major networks, Colbert’s proportion is the worst, but they’re all bad.
That’s the thing- while I’m sure that was at least an extra bonus, it doesn’t make sense to say that the show will end in 2026 and Colbert will not be replaced if it was all about appeasing Trump. Sure, they have to pay him until his contract runs out in May- but that doesn’t mean they have to give him a platform until May.
O.K., the problem is not the total number of viewers of the show. The problem is that the viewers tend to be older. Advertisers don’t like it when the viewers are largely older. They have discovered that commercials don’t tend to change the buying habits of older viewers nearly as much as the buying habits of young and youngish-middle-aged (18 to 45 years old) ones. So they insist on paying less for advertising on shows with mostly older viewers. Networks want shows that they can charge more to the advertisers on those shows.
A similar thing happened to American television in the early 1970s for shows that appealed mostly to rural audiences. The networks decided to drop such rural-appealing shows and pick up ones that appealed more to urban audiences. This was called the “rural purge”.
I looked up that TLSwSC had 154 episodes in 2024. If the $100M figure is right, that’s $650k/episode, or about $3,200 per person per episode. That doesn’t sound crazy high. I’m rolling equipment costs and overhead in with the per person costs. Plus costs for guests, supplies, etc. Not a crazy number all things considered. I’m a tad skeptical about the ad revenue reports, but will note that I only watch TLSwSC on YouTube, which pays bupkis compared to over the air ads.
In addition to Colbert and the band, there are: writers, producers, directors, camera people, sound people, lighting people, hair, makeup, costume, graphics, video effects, animators, stage hands, electricians/gaffers, safety people, wranglers for guests and audience, network folks like lawyers and PR. I haven’t acted since high school, and never on TV, so I’m sure I’m missing several categories. 5 people in each of 20 categories is 100 right off the top. I suspect for those categories with fewer people, others have more. And probably back-up folks for many categories. I think Colbert has a big writers room. Meyers has at least 8 writers he mentions on his show.
Yeah, 200 sounds big, but maybe that’s why his show is so damned good, in my not at all humble opinion.
Looking at the list below it seems the calculation gets way too bothersome to figure out. The total staff working there has been huge but some consideration of a per-episode basis seems what is needed and that is WAY too much work for me to do for this thread (sorry but not sorry).
And that just tells us someone worked on the show but not how much work was put in.