How Gianmarco Soresi Got 1 Million Views On His YouTube Special
An analytical deep dive into the stand up comedian's debut hour. Plus, Netflix data.
Hello and welcome to the 21 new subscribers since last week’s column. It appears many of you are here because my name came up in Dropout CEO Sam Reich’s recent tweet/Thread/ LinkedIn post asking for someone who analyzes the virality of online comedy, so a special thanks to those of you who mentioned my name in the comments.
Today we’re looking at the debut stand up comedy special from Gianmarco Soresi, who coincidentally enough was recently on an episode of the Dropout show Game Changer.
Self-released on YouTube last week, the special has already garnered over a million views.
I’ve put together some interesting data that paints a compelling picture, illustrating the expert way in which this special was released, edited, promoted, and distributed. Gianmarco and his team were kind enough to share with me some of the back-end analytics, so buckle up for some fascinating insights ahead that are typically only ever viewed by the individual themselves.
We’ll uncover a few key metrics propelling its high viewership, and compare its performance to a few recently released stand up comedy specials on Netflix, with a look at their in-house stats.
First of all, if you haven’t gotten a chance to watch Gianmarco’s special, I’ll link it here:
Quick Stats (at the time of pulling this data on 9/26/25)
Date Posted: 9/19/25
Runtime: 1:06:20 (76.3 minutes)
Total Views: 995,307
It should be noted that on YouTube, views are only counted if the user watches for more than 30 seconds.
Total Watchtime: 383,600 hours
Average Watch Time: 23 min, 16 sec
Average Percentage Viewed: 35.1% (percentage watched of the total runtime)1
Impressions Click-Through Rate: 3.6%
More on this, later.
Here are some key things I observed after watching the special and looking at the data.
First off, do you notice how the special just kinda…starts?
It’s become common – dare I say, “hack” – to start a stand up special with a certain kind of preamble. This typically includes but is not limited to: a frazzled venue arrival, a random celebrity cameo, and 1-2 minutes of what my friend Steffen calls, “Woo! Comedy”, which consists of the comedian yelling questions like, “AUSTIN, HOW THE **** ARE WE?” and “HOW’S EVERYBODY DOING TONIGHT?” in order to collect a bunch of “Woo”s from the crowd, prior to settling into their actual material.
In fact, many of these tropes were parodied by Will Angus in an Almost Friday video:
Not here.
No preamble, we’re straight into The Constitution, and laughing within 5 seconds.
Even if you have no clue who Gianmarco is, you can pretty quickly piece together what you’re watching – high quality footage of a guy telling funny jokes on stage.
This almost feels obvious to point out, especially when posting a video on YouTube, where you’re competing with a hundred other things for the undivided attention of a stranger for an hour, but you’d be surprised how rarely this type of intro is utilized.
Now you might say to yourself, but Gianmarco has a massive head start compared to the average comedian, I mean his YouTube channel has 1.2 million subscribers:
But if we look at the special’s back-end data from YouTube Studio, you’ll notice that 78.6% of the special’s viewers are not subscribed to Gianmarco’s personal channel:
Within a day, non-subscribers were watching the special three times as frequently as his subscribers. This makes sense for a few reasons.
Firstly, 1 million subscribers does not mean 1 million daily, active Gianmarcophiles. For all we know, half of them subscribed in 2022, and don’t go on YouTube anymore.
Think about your own social media habits.
Do you check each of YouTube, Instagram, TikTok, Snapchat, Threads, LinkedIn, Facebook, etc. every day, for hours on end? Even if you do, do you scour through the posts of every individual you follow on each platform? Most likely not. But you’re still “following” dozens of other users on all of them, even if you never check them.
This is why I encourage creators, casting directors, advertisers, and brands to reframe a vanity metric like “Followers” or “Subscribers” as “Perceived Total Reach”. The real value of building an online audience is not the number at the top of your profile, but how many of them you consistently reach and mobilize (Mobilized Audience Rate).
Take a closer look at the previous chart:
If you pay attention to the blue and yellow2 lines that diverge halfway between the ‘Sep 18, 2025” and “Sep 20, 2025” labels on the X-axis, you’ll notice the video starts off only being watched by Gianmarco’s subscribers, who watched the video for 25 minutes and 3 seconds on average, or 37.8% of the total special:
That figure might sound depressingly low if you’re a comedian accustomed to live shows, where 95%+ of the audience stays for the full show, other than bathroom breaks and walkouts.
But remember, these are not ticket buyers. These are strangers on their iPhone or TV.
You’re competing with dinner, and homework, and laundry, and family time, and video games, and porn, and Sunday Night Football, and the WNBA playoffs, and the gym, and date night, and the Yankee game, and so many other factors that it’s a miracle any of them stayed for 23 minutes, or the length of a full episode of a TV show.
Hell, posting a YouTube video of your standup is way closer to playing a saxophone outside of Grand Central than it is to crushing a spot at the Comedy Cellar.
Gianmarco’s high watch time metrics signaled to YouTube, “Hey, the people who watch this video are spending a lot of time watching this video. We should show it to a bunch of other users, including the ones that don’t follow this account, so that they also spend a lot of time watching this video, so that we can make them all watch a shit ton of advertisements provide entertaining content to our valued customers.
And so his special started getting shown to others via places like the YouTube home page. In fact, let’s take a look at how viewers found the special:
The vast majority of people came from “Browse features”, which basically means the default YouTube home page or Subscription tab on the left-hand side of the site.
Here’s what that Traffic Source data looks like if we drill one layer deeper:
Another very important metric worth calling out: ZERO views were obtained from YouTube advertising, as you can tell from the very last row of data.
This means every single view of this special was earned organically.
It is far more common than you may think to pay for eyeballs on YouTube, especially when there is outside pressure from investors, producers, etc. to reach a vanity metric like X number of views. Paid spend can absolutely be effective if you target the right audience, who in turn generate more organic viewers based on their high watch times and shares. But oftentimes this method does more harm than good, feeding a video to a million of exactly the wrong people for the sake of a high perceived total reach.
The fact that the only money spent here was on the filming and editing of the special itself is a testament to the quality of Gianmarco’s material and performance.
Another thing worth highlighting is the Impressions Clickthrough rate (CTR).
At the risk of sharing something incredibly vulnerable, below is my current YouTube home feed, which contains ZERO video from channels I currently subscribe to:
When I opened YouTube just now, all six of these videos just got logged as an “impression”. If I clicked into one of them, YouTube Studio would count that as a “click”. So the metric labeled Impressions Clickthrough Rate (CTR) tracks how many of the people shown this thumbnail (impressions) actually clicked on it (CTR).
In general, a good Impressions CTR to target with a YouTube video is 2.0%.
Gianmarco more than doubled that, as viewers coming from YouTube’s browse features earned an Impressions CTR of 5.0%.
In other words, 1 in 20 strangers who were shown this title and thumbnail felt so compelled by the title, image, and posture that they pressed play:
On a separate but related note, I spent a lot of my weekend looking at Netflix’s semi-annual, “What We Watched” report posted on their website in mid July.
They included the raw data in a spreadsheet at the bottom of their page, and made it freely available to download and tinker with. Unfortunately I’m a Virgo, and that’s my idea of a good time, so I spent most of my Friday night isolating all the comedy specials released this year3, watching a little bit of each of them, and seeing if I could identify what separated the top viewed specials from the least viewed specials.
What I found was very interesting.
First of all, they round to the nearest 100,000 and no special gets recorded as anything lower than 100,000 views. That’s their zero.
Second of all, they count views very differently than YouTube.
Netflix calculates views by taking total hours watched and dividing it by the runtime.
So if a comedy special racks up 269,140 hours watched and the runtime is one hour, they’ll report that as 300,000 “views”.
Let’s take a look at some of their “most viewed” specials released in 2025. The first five columns are Netflix’s, and the right-hand column is mine, tracked manually:
I’ve written before about optimizing for the time it takes to get to a first laugh (TTFL) when posting social media clips of comedy, but never did I expect to prove the same theory with Netflix data, with a slightly different metric I’ll call First Laugh on Stage.
Let’s contrast these with some lower viewed Netflix comedy specials released this year.
For comparison, I’ll include Gianmarco’s special, and place where it would sit using Netflix’s version of tabulating view count, as if it were released on the platform:
Gianmarco’s special is outperforming the viewership of several other comedians at his “tier” of name recognition in the public eye, even with their perceived Netflix “boost”.
However, it should be noted this is by no means an apples to apples comparison.
Some of these specials were released months before the Netflix “What We Watched” report was released in July, whereas Steph Tolev's was released just three weeks prior.
Netflix themselves would probably tell you that the real key metric is the preexisting starpower of the individual comedian, but I’m not so convinced.
Bert Kreischer and Gabriel Iglesias are unquestionably mega-famous comedians that border on rockstars, especially when you look at the rooms their specials were filmed in vs. the rest. But are Andrew Schulz and Ari Shaffir similarly considered bonafide A-listers? Regardless, the first two take stage immediately with striking images, and the latter two have their crowd laughing in a matter of seconds. Just like Gianmarco.
Furthermore, six days after its release, Gianmarco’s YouTube special is outperforming the average episode of John Mulaney’s recent talk show, even if you use Netflix’s KPIs.
Mulaney is a veteran comedian with significantly bigger starpower, and multiple specials of his own that I’m sure performed great in the past. I probably watched New In Town for 15 total watch hours just by myself, back in the day. But every episode of Mulaney’s show has a pre-show provocation, an intro theme, and a minute of preamble before the first written joke. I know this is the traditional way late night shows operate, and that these were recorded live shows, but perhaps an underrated reason those shows have a fraction of the reach they used to is because it takes them three minutes before anything funny happens. Meanwhile, they are competing with millions of other content offerings that take just a few seconds to get interesting.
The real issue here is measurement.
By calculating views this way, Netflix is pretending everyone who clicked “Play” watched the entire special. Obviously, this isn’t the case.
YouTube zags the other way. A “view” is credited after thirty seconds. That means someone who tapped out at the thirty-one second mark is treated the same as someone who stayed for the full hour and six minutes of Gianmarco’s special.
Both platforms overestimate the total perceived reach, but in different ways.
Netflix inflates by pretending completions are universal. YouTube inflates by counting every start as equal. Neither tells you what the hell any viewers actually did.
Rather than complain about this without proposing a solution, I personally think a better way to measure “true” viewership would be four numbers, side by side:
Unique Starts: How many individual users pressed play.
This still allows Impressions CTR to be a worthwhile metric to monitor, and encourages A/B testing of different thumbnails, titles, and video descriptions.
Median Watch Time: When did the middle-est viewer tune out? Median is a better measure of central tendency than the average, because it is more resistant to outliers (cut to my STAT 100 professor, kvelling)
Completion Rate: How many unique viewers made it to the end, or at least the end minus the credits.
Are your viewers Jared Goff? Or are they more of a Jameis Winston crowd?
Rewatchability Rate: Of the users with a 90%+ completion rate, what percentage of them completed it again? Or were there certain segments they return to?
I’ve personally watched several Netflix comedy specials multiple times4. This should absolutely be a factor in measuring success, as anything that lends itself to rewatchability is an extremely valuable asset (see: Office, The)
You can keep Total Watchtime in the mix, but it shouldn’t be the headline metric.
If Netflix reported Median Watch Time and Completion Rate, we would have a far more honest picture of how these specials perform. And if YouTube highlighted completions alongside views, it would put the viral “million views” claim in context.
Right now, both platforms are hiding the same, sad truth: most people don’t finish.
Meanwhile, when you buy a ticket to a live show, the completion rate is ~98 percent.
The same goes for the movies. When studios cite box office numbers, it’s understood that almost everyone who bought a ticket sat down and watched the entire film. You don’t need to caveat it. One million tickets sold is a million “completions”, so to speak.
Streaming numbers are a totally different ballgame, and completely non-standardized.
For Netflix, 300,000 “views” could either mean 300,000 people watched the full hour, or it could mean millions of people sampled it for a couple of seconds and dipped.
On YouTube, one million “views” could mean one million walk-outs after 30 seconds.
That’s the uncomfortable truth in 2025.
The number of people who actually finish a special is often closer to a couple of nights at the Beacon Theatre than the inflated digital figures suggest.
The live, ticketed event is still the only environment where you can point to the audience and say with confidence, “They saw the whole thing.”. This all goes back to the earlier point about Perceived Total Reach vs. Mobilized Audience Rate.
Streaming sells you on increased reach. But live shows prove there’s even an audience.
I wrote about this last fall, but the smart bets are on comics and formats that already work in front of a highly mobilized crowd.
With this in mind, Netflix’s role as curator and distributor should be to identify live shows with highly mobilized fanbases, and leverage its marketing budget, SOV, and international distribution to catapult their reach in a way that they wouldn’t possibly be able to on their own.
What they’re not doing is paying attention to why the best ones actually worked.
Many people love complaining about how the “algorithms” are “constantly changing”.
I have a hard time believing Netflix’s back-end analytics suite for measuring retention is meaningfully different from YouTube Studio.
There’s only one KPI to optimize for once you already have someone’s monthly subscription revenue, and you’re not selling them ads: Time Spent on Platform.
Look again at the top viewed specials from 2025. All of them take the stage immediately. Even the Kill Tony special begins with the two hosts at their mics.
Netflix’s viewership data makes it seem like they think people still use their service as an at-home cinema. We may have back in the DVD rental days, but we don’t anymore.
In 2025, people watch Netflix in largely the same way they watch YouTube: on their phones, on their tablets, and on their TVs, with one hand on the remote and the other hand on their phone or their dick. The idea that a casual user is giving anything on their platform their undivided attention for the entire runtime is straight fantasy.
What’s frustrating about it is, they’re making lots of good bets. Kill Tony was unquestionably a hit. Liza and Steph and Rosebud are all fantastic comedians. But part of giving somebody a megaphone is showing them how it works. Something tells me each of their specials would have reached significantly more people if they had simply “cut to the chase”, and gotten to their scripted material a little bit quicker.
And I know there are probably some creatives and directors and artistic types scoffing at this, or thinking I’m trying to “growth hack” an art form.
But was this not what we were taught in like, basic middle school essay writing? To come up with a creative way to grab the reader’s attention? To my mind, the art of the “hook” is agnostic across all mediums. Of course the start of your creative works should be compelling to a stranger, whether it’s a movie, a book, or an Instagram Reel.
That’s why the time until that first laugh on stage matters so much.
It’s social proof. Crowds attract crowds. Other people are laughing, I’m in good hands here.
Gianmarco understood that. He treated YouTube like what it is: a place where you only get five seconds to hook people. And by every metric that matters, it worked.
In general, a good target Average Percentage Viewed (APV) is 40-50%.
Or is it green? I don’t know, I’m color blind.
For the three nerds that care: I filtered the “Shows” tab to shows with a runtime between 40 minutes and 2 hours, and then used the Date column to limit it to those released in 2025.
Off the top of my head: James Acaster’s four-part Repertoire, all of Tim Minchin and Bo Burnham’s specials, Catherine Cohen: The Twist? She’s Gorgeous, Natalie Palamides: Nate, Adam Sandler: 100% Fresh, Hannah Gadsby: Nanette, John Mulaney: New In Town, Oh Hello on Broadway, Middleditch & Schwartz, Dave Chappelle: Deep in the Heart of Texas,
It's worth noting that by virtue of being on Netflix, those Netflix "views" are likely much larger. People don't really sit around in a group to watch YouTube but often something on Netflix has many people viewing it.
It's also worth analyzing the "family friendliness" of comedians -- Netflix probably gives an inorganic boost to family friendly comedians because if you're in the mood to watch stand up in a group, you're more likely to pick a comedian appropriate for the most prudish or youngest person. Also worth comparing "adult" comedian success on YouTube vs Netflix for that reason since more adult content is probably more likely to be consumed solo with a controlled private screen
thanks for another AMAZING analysis as always !!!