Insights
Holly Willoughby Together: What All 40 Videos Reveal
The episode the Telegraph called a fifty-minute borefest is the best-performing video on her channel. Five weeks, 40 videos and nine transcripts in, almost nothing about this launch matches the reviews it got.
She launched the channel on 21 July 2026. Five weeks later it has 21,600 subscribers, 588,966 views and 40 videos, and it has been reviewed as though it were television.
The Telegraph: "Holly Willoughby's dire YouTube show makes Meghan look good… Not even Stanley Tucci can redeem this 50-minute borefest." The Times gave it two stars. The i Paper called it "a confused attempt to present herself as relatable".
The Tucci episode they were reviewing has 86,824 views and the highest comment rate of anything on the channel. As television it may well have been a borefest. As YouTube it’s the best thing she has published.
I pulled all 40 videos, all nine long-form transcripts, the comment threads and the UK search data, twice, two days apart. The channel does have a real problem. It isn’t the one in the reviews, and it’s a fixable one.
The 10 findings
- The launch is decaying on schedule. Weekly subscriber adds have gone 6,420, 5,410, 4,000, 1,800, 2,300. Down 57% from week two to week five.
- Guests are worth about 3x the cooking, once age is accounted for. Views per day at the 27 August pull: guest episodes 2,839, cooking 960. On raw lifetime views the gap looks like 6.5x, and most of the difference is the guest episodes having been up for longer.
- Her most-watched videos are her least liked. Her least-watched video is 1.9x better liked than her most-watched, and the comment sentiment runs the same way.
- She is number one for a search barely anyone makes. "Holly Willoughby Together" is too small for Google Ads to report a volume at all. "Toad in the hole recipe" returns 60,500, she has made one, and she isn’t on the page.
- The best story on the channel is in the least-watched video. Two brothers describing their grandmother arriving from Jamaica in 1956, at 7:11 of a video 4,747 people have seen. The title and the thumbnail didn’t mention it, and the title has since been changed.
- The channel has already changed shape twice in five weeks. A trailer, then a guest show, then a guest show with a cooking strand bolted on, and the long-form median moved with each change.
- Shorts are 78% of the library and 40% of the views. The median Short takes 3,746 views against 36,132 for the median episode, and a typical Short is seen by about 17% of the people who already subscribe.
- One video doesn’t behave like any of the others. 27,605 views, 166 likes and a single comment, on a channel whose own Shorts would predict 491 likes and 22 comments.
- The critics and the comments aren’t discussing the same thing. The five most-liked comments on the episode the press reviewed are all jokes about how she left her last job.
- The channel is already testing. The jerk chicken title was rewritten between two data pulls a day apart, from an occasion to a phrase people search for.
The takeout to try yourself: Your first weeks measure the audience you already had. What matters is what the videos do once that list is used up. Holly's subscriber adds fell 57% between week two and week five.
1. The launch is decaying on schedule
Every new channel gets a spike. Hers came from a national press cycle and a Stanley Tucci booking, and it’s fading the way launch spikes fade.
Views have held up better, because a famous guest keeps landing: 37,560 in week one, then 104,579, 171,121, 95,485 and 143,964. Subscribers matter more here, because a subscriber is someone who decided to come back.
The takeout to try yourself: If you’ve just launched, don’t judge the channel until your own contacts have stopped arriving. Until then you’re measuring your network.
2. Most of the views are borrowed
The channel runs two formats. On Fridays a famous person comes round. On Tuesdays Holly cooks, usually alone.
Measured as views per day at the 27 August pull, guest episodes take 2,839 and the cooking takes 960, so the real gap is about 3x.
The cooking is also improving. Toad in the hole is running at 937 views a day, chicken Caesar at 960, and the jerk chicken episode at 2,238.
The guests are worth what their names are worth. These are UK Google Ads figures, pulled on 26 August 2026.
| Name searched | UK searches/mo |
|---|---|
| Stanley Tucci | 135,000 |
| Holly Willoughby | 110,000 |
| Lily Phillips | 90,500 |
| Katherine Ryan | 74,000 |
| Joanne McNally | 49,500 |
| Emma Bunton | 40,500 |
| Holly Willoughby Together | no data |
The takeout to try yourself: Borrowed attention is a real strategy and a fast one. It stops working the day you can’t book anyone, so ask early what the channel is worth when the guest is you.
3. Her most-watched videos are her least likedHer least-watched video is better liked than her most-watched
| Video | Days live | Views | Like rate |
|---|---|---|---|
| Stanley Tucci | 30 | 86,824 | 1.95% |
| Lily Phillips | 6 | 72,804 | 1.12% |
| Joanne McNally | 20 | 62,498 | 1.76% |
| Katherine Ryan | 27 | 51,445 | 1.87% |
| Emma Bunton | 13 | 36,684 | 2.06% |
| Ellen Roome | 23 | 25,732 | 2.01% |
| Toad in the Hole | 16 | 15,110 | 2.93% |
| Chicken Caesar | 9 | 8,755 | 2.80% |
| Jerk Chicken | 2 | 4,747 | 3.62% |
The most-watched video on the channel is liked by 1.95% of the people who watch it. The least-watched is liked by 3.62%. Median like rate on the cooking is 2.93% against 1.91% on the guest episodes, and the pattern holds across every pair in the table regardless of age.
The comments say the same thing. The jerk chicken episode drew twelve comments and every one of them is warm, including the two guests thanking her. The Lily Phillips episode, at fifteen times the audience, drew "shouldn't be giving these people a platform disgusting", "vile" and "revolting".
More people watched. Fewer of them liked it.
The takeout to try yourself: Put like rate next to views on every video. If views rise while like rate falls, the extra people are the wrong people.
4. Number one for a search nobody makes60,500 searches a month, and she isn’t on the page
"Holly Willoughby Together" returns no volume figure at all from Google Ads in the UK, which is what the keyword planner does with a phrase too small to report, and UK Google Trends finds it in five weeks out of the last 113. She owns that results page outright: six of the top ten YouTube results are hers.
"Toad in the hole recipe" takes 60,500 UK searches a month. She has made a toad in the hole video. She doesn’t appear on that results page anywhere.
The UK results page is held by Kitchen Sanctuary at 624,287 views, Jamie Oliver at 934,388, and at number five, This Morning, with 27,410 views. Her old show ranks for the dish she just cooked.
Her episode is titled "I Feel Like I'm On OnlyFans | Holly Makes Toad In The Hole From Scratch". The dish is in there. The word people actually type, "recipe", isn’t, and neither is anything that tells YouTube this video answers that question.
The same gap sits under every cook day: jerk chicken recipe takes 14,800 UK searches a month, chicken Caesar salad recipe 6,600. Those are real audiences with real intent, and the channel is making exactly the right videos for them without competing for them.
The takeout to try yourself: Check what your audience types before you name the video. A title that wins the click and misses the query only works while someone is pushing it.
5. What the transcripts showNine transcripts, and the best story on the channel
I read nine long-form transcripts. The guest episodes open on a cold-open montage of the best lines: Joanne McNally's "Have I just hooked up with Ali G?", Lily Phillips' "You once made £800,000 in a month", Emma Bunton's "I've got knickers bigger than this."
The cook episodes open on housekeeping. Toad in the hole begins:
"Sorry. I said I could do it and then I'm like I'm a moral. Okay. Thanks. So, um what have we got here?"
Chicken Caesar opens mid-fluster, "I'm kind of making it up as I go along… Where's my jelly cam?" Jerk chicken opens on the weather and the schedule: "It’s full-on summer out there. Carnival season is just around the corner."
The best story on the channel is at 7:11 of that jerk chicken video. Craig and Sean describe their grandmother arriving from Jamaica in 1956, on her own, to "no Irish, no blacks, no dogs", working as a cleaner in the morning and a nurse in the afternoon, cooking for everyone at the weekend because that was how she showed love.
4,747 people have seen it. Nothing in the title, the thumbnail or the first minute tells you it’s there.
The takeout to try yourself: Find the best ninety seconds in your last video and ask whether the title and thumbnail promise that moment. On most channels the answer is no.
6. The channel has already changed shape twiceWeek by week, three different channels
Sorted by week:
| Week | Long-form | Shorts | Long-form median |
|---|---|---|---|
| 1 (21 to 27 Jul) | 0 | 1 | trailer only, 81,424 views |
| 2 (28 Jul to 3 Aug) | 2 | 9 | 68,902 |
| 3 (4 to 10 Aug) | 2 | 4 | 43,858 |
| 4 (11 to 17 Aug) | 2 | 8 | 25,551 |
| 5 (18 to 24 Aug) | 2 | 9 | 38,822 |
| 6 (25 Aug) | 1 | 0 | 3,325 |
Week four is where the second format arrives. Holly starts cooking alone, the toad in the hole episode goes out on 11 August, and the long-form median drops to 25,551. Week five recovers to 38,822, and it recovers because Lily Phillips is in it rather than because the cooking improved.
Every long-form episode until now has gone out surrounded by Shorts: seven, nine, ten, thirteen, fourteen, sixteen of them. The jerk chicken episode on 25 August went out with one.
The takeout to try yourself: Count your uploads by format and by week before you judge any trend. If you added a format halfway through, an average across the whole period describes a channel that no longer exists.
7. Shorts are 78% of the library and 40% of the viewsWhat 31 Shorts are actually reaching
Thirty-one of the forty videos are Shorts. Between them they hold 39.8% of the channel's views, 37% of its likes and 18.1% of its comments.
The median Short takes 3,746 views. The median long-form episode takes 36,132. One episode is worth about ten Shorts.
Twenty-two of the thirty-one Shorts are under 5,000 views. Three have ever passed 10,000, and two of those are the launch trailer and the video in the next section.
Compare that with her subscriber count. A typical Short on this channel is seen by about 17% of the people who already subscribed. It isn’t reaching anyone new. And it hasn’t moved: the Shorts median was 5,481 in week two and 3,362 in week five.
They’re clips of the episodes, published in bursts of seven to sixteen around each one. That’s a trailer strategy aimed at an audience that has already found the channel.
The takeout to try yourself: Compare your Shorts median against your subscriber count. If a typical Short is seen by fewer people than already subscribe, posting more of them won’t change that.
8. One video doesn’t behave like any of the othersThe one video whose numbers don’t add up
Bought views show up as a view count that rises while the likes and comments don’t. Every video on the channel was checked against the channel's own rates for its own format.
| Emma Bunton shoes Short | Actual | Expected at channel rates |
|---|---|---|
| Views | 27,605 | — |
| Likes | 166 | 491 |
| Comments | 1 | 22 |
Its like rate is 0.60% against a Shorts median of 1.78%. One comment on 27,605 views is the part that can’t be explained by the clip being weak, because a weak clip that 27,000 people watch still collects twenty comments.
A second video shows a weaker version of the same pattern: the "Episode 1 With Stanley Tucci Is Now Live" Short took 17,865 views, 163 likes and 13 comments, a like rate of 0.91%. Both are announcement clips rather than content clips, which is the sort of video a channel puts money behind at launch.
The launch trailer is the control. It took 81,562 views, the most of any Short, and it was liked by 1.91% of them and drew 160 comments, which is exactly what the channel does organically. So this isn’t a channel buying views across the board.
Putting money behind a launch clip is an ordinary thing to do and there’s nothing wrong with it. Public data can’t prove that money was spent here, and it never can. What it can show is that two videos were watched by people who behaved unlike the channel's own audience, which matters because those two view counts aren’t evidence that the videos worked.
The takeout to try yourself: Divide likes by views and comments by views before you celebrate a video. A view count that jumps while both rates fall is traffic that arrived from somewhere else.
9. The critics and the comments aren’t discussing the same thingWhat the top-liked comments are really about
The reviews were about the programme, and all three were reviewing the Stanley Tucci episode, which is the best-performing video on the channel and holds its highest comment rate.
The obvious conclusion is that the audience disagreed with the critics. The comments don’t support that. They aren’t about the show at all.
Sorted by likes, the top of the Tucci comment section reads:
- "Who came here immediately paused the video and went to the comments?", 99 likes
- "Did Stanley have to queue to get into the studio?", 90 likes
- "Are you ok?", 80 likes
- "First of all, are you okay?", 58 likes
- "Just here for the comments", 49 likes
Every one of those is a reference to how she left her last job. The comments that are actually about the episode sit underneath them: "Really enjoyed this keep them coming" on 9 likes, "This was so fun. Looking forward to more!" on 3, "Love this, it's like a warm hug" on 0.
The same comment tops the Katherine Ryan episode with 248 likes. It’s the most-liked thing anyone has written anywhere on the channel, and it’s a joke about people coming to read the comments.
That’s what her comment rate is measuring. The guest episodes draw the most comments and the least discussion of the video, and a new viewer who lands on the Tucci episode is shown five jokes about her before a single opinion about the show.
The cooking videos are different. The comments there are about the video. The jerk chicken episode drew twelve comments and all twelve are about the food or the guests, including Craig and Sean writing "Thanks for having us holly! You're a Jamaican chef now!"
The chicken Caesar comments are production notes: "The camera angles and constant shots of the crew are so distracting and make this hard to watch" on 12 likes, "Holly seemed so anxious in this one", "please use a tripod the camera is so shaky", and someone pointing out her name is misspelled at 0:59.
The three sources disagree. The press says the show is boring. The like rates say the cooking is the best thing on the channel. The comments say most people arrived for something that has nothing to do with either.
The takeout to try yourself: Sort your own comments by top-liked rather than newest, because that’s the order a new viewer sees them in. Check whether the first five are about the video at all. If they aren’t, your comment count is measuring something else.
10. The channel is already testingA title changed between our two pulls
Between the two data pulls for this article, the title of the jerk chicken episode changed.
On 26 August it read "Holly Willoughby Makes Her First Ever Jerk Chicken W/ Original Flava". On 27 August it read "Holly Willoughby Learns How to Make Jerk Chicken With Original Flava".
It’s a small edit with a specific effect. "How to make jerk chicken" is a phrase people type into YouTube. The first version described an occasion. The second one answers a question, which is exactly the gap section four is about.
In that same day the video went from 3,325 views to 4,747. One video over one day can’t tell you the title caused it, and I’m not claiming it did. What it does tell you is that someone on that channel is watching the numbers and changing things in response to them.
A television review can’t see this. An episode is finished the moment it airs, and a YouTube video is a test that keeps running.
The parts of the channel that have been changed most are the parts improving. The cooking format is three videos old and the third is running at more than twice the daily rate of the first two. The Shorts are the least changed thing on the channel and the flattest.
The takeout to try yourself: Treat a published video as a test that’s still running. Change one thing on an underperformer, note the date, and look again a week later. Most channels never go back, and learn nothing from the videos that didn’t work.
The inconvenient truth
Five weeks in, she has built something the reviews didn’t credit: 21,600 subscribers, guests most channels couldn’t book, and the highest like rates on the channel on her cooking videos. The Tucci episode the Telegraph called a borefest is the best-performing video on the channel and drew the most comments of anything she has published.
The structural problem is the one the reviews missed. Most of the views on this channel are borrowed from someone else's name. The show's own name is still too small for Google to report a search volume. And the videos her audience loves most are the ones almost nobody is shown.
Every one of those is fixable, and most of what looks wrong here is normal for a channel this young. Finding which of them is costing the most, in a month where the numbers move, is the job our YouTube channel management service does. The critics reviewed it like a finished programme. It’s five weeks old and still being built.
Method and limitsSources, dates and what this can’t see
Public data only, pulled twice, on 26 and 27 August 2026: a full channel export, the YouTube API for exact like and comment rates, nine long-form transcripts, twenty comments each from five videos and UK Google Ads for search volumes.
There’s no Studio access here, so nothing estimates impressions, click-through or retention. Results pages were checked twice on the day and positions move.
At five weeks old every trend is short, and the newest videos have days of data rather than weeks, so per-video comparisons here are made on views per day and labelled with age. YouTube changed how it counts a view on 24 August, so the jerk chicken episode is counted more generously than the eight that came before it.
Analysis and write-up: Stewart Read.
Meet Stewart
Stewart Read
YouTube Growth Strategist
Stewart has run YouTube channel management for 8-figure businesses - turning channels into reliable sources of leads, clients, and measurable revenue, not just views.
Using a class-leading suite of channel analysis tools he built himself, he pinpoints exactly what makes your audience click, watch, and convert - then turns those findings into videos that bring buyers to your business.
Channel Intelligence
The monthly service behind this analysis. A written briefing on what your channel returned to the business, a walkthrough of the findings, and a plan for the videos to make next.
See how Channel Intelligence works