A real YouTube channel audit, start to finish

01 · Where it startedWhat those numbers mean

You have three million views and almost none of it is working for you.

The audience is already here, so nothing below is about getting bigger. What is missing is a second video that can do what the first one did, titles YouTube can actually read, and an ad setup that collects what this audience is worth.

Read it this way

The 7:11 up top is your average view duration. Retention means how long people stay once they start watching. YouTube reports it a few different ways, as a clock time, as a percentage of the video, or at one exact second, but it is all measuring the same thing. Yours is strong however you cut it.

That is the hard one to fix. You cannot teach somebody to be watchable, and you already are.

Everything else holding you back is craft, which means all of it is fixable. It comes down to the packaging that earns the click, and the ad setup that collects once you have it.

02 · The video carrying everything

58 million impressions. 2.2% of them clicked.

Episode 1 - Oven Roasted Crispy Chicken Thighs. YouTube has shown this video to people 58 million times, and about 94% of that came off its own suggestions rather than anything you did. Only 2.2% of those people clicked.

Even at 2.2%, that is 1.3 million views from YouTube alone, before you count search and links. The ones who did click stayed a long time: over eight minutes each, which is about 34% of a 23-minute video.

So YouTube is already doing the expensive part for you. Those 58 million impressions get served whether your thumbnail earns the click or not, which is why one or two points of click-through is worth more here than anything else in this report.

Chef Ange thumbnail: Crispy Chicken Thighs, white condensed caps over a close-up of roasted thighs — Your biggest video, shown at about the size somebody actually sees it on YouTube. **Hold this one in mind.** The title sitting above it is the first thing I want you to change.

Times YouTube showed it, called impressions58,000,000

Clicked through2.2% → 1.3M views

Watched past 8:00 of 23:00~34% retention

Arrived from YouTube search10% of views

And the money

RPM is what you earn per thousand views. Yours on this video is $5.04, which is low for the audience watching it.

You have three mid-roll ads running, meaning ads that play partway through, at 7:30, 19:13 and 23:29. There is nothing before the video and nothing after it. Your average viewer leaves around the eight minute mark, so two of those three ads almost never play. You are getting paid for one of them.

Had you done this prior to the surge of views last May, you probably could have netted half as much again on this video alone.

Report 01 · Monetization

03 · What was actually brokenEight fixes, roughly by impact

None of this needs a better camera.

Each card below tells you what is wrong, why it matters, and roughly what it costs you in time against what it gives back. They run in the order I would do them.

If you only do three things

Drop the episode number and lead with the words people actually type. Fixes 01 and 08.
Move your mid-roll ads inside your average view duration, and switch on pre-roll and post-roll. Fix 04.
Fill the tag field and repeat the keyword in the description. Fix 02.

01Packaging

Your episode numbers turn into hashtags

Your titles open with #014. YouTube treats a leading hash as a hashtag and turns it into a link, and that link goes to a page full of other people's videos. So the first few characters of your title, the ones doing the most work, are pointing people away from you.

Ten minutes · Highest impact

02Search

Tags at 71 of 500 characters

You are leaving 86% of that field empty, and tags are one of the ways YouTube works out what a video is about.

Put the same set of channel tags on every upload, so YouTube learns what your channel is. Then put that video's main keyword in the title and use it three times in the description, so all the words line up.

Twenty minutes an upload · High impact

03Retention

No chapter markers

Your recipe steps are already written into the description. Add a timecode to each one and YouTube turns them into a chapter menu on the player. That matters more on a cooking video than almost anywhere else, because people jump straight to the step they are stuck on instead of leaving to find it elsewhere.

Five minutes an upload · Medium impact

04Revenue

Your ads play after people leave

Two of your three ads sit past the point most people have already gone, so they barely ever play. Move them inside your average view duration, roughly one every three to seven minutes, and switch on the ad before the video and the ad after it.

Five minutes an upload · Highest impact

05Revenue

No affiliate links on a channel built from tools

Every episode puts equipment on camera: the large pot at 6:35, the grinder at 6:46, the hot hands at 14:50. Anyone who wants to cook along with you has to go and hunt all of it down themselves. An Amazon affiliate account takes an afternoon to set up and pays you a cut of anything they buy through your link.

One afternoon to set up · High impact

06Session time

No card when you point people elsewhere

A card is the clickable link that pops up in the corner of a video. You send viewers back to an earlier episode for your base-seasoning ratios, and there is no card on screen when you say it.

Nobody is going to go digging for that video. Put a card up the moment you mention it and you keep them on your channel instead of handing them back to the feed.

Two minutes an upload · Medium impact

07Craft

You skip the details a beginner needs

At 10:12 you never say what the heat is set to. Is that a four, or a max burner? How long between the onions and the chili powder? Put a running clock on screen and say the settings out loud, because the person who most needs this recipe is the one who cannot fill in those gaps on their own.

Changes how you shoot · Medium impact

08Packaging

Your titles describe instead of sell

On a phone, only the first forty characters of your title are visible. You are spending them on an episode number, when they should carry the words somebody searched for and a reason to click.

Every upload · Highest impact

04 · Who was actually watching

Your audience is not who you think it is.

Across the channel, 63% of your viewers are women, 81% are 55 or over, and 66% are in the United States. On your breakout video it goes further: 65% women, 70% over 65, and 98% of them were not subscribed.

That last number is the one to sit with. Almost everybody who watched your biggest video was a stranger passing through, and there was nothing else on the channel packaged well enough to catch them.

You have never once made a video with these people in mind. That is not a criticism, you had no reason to know. It is the biggest open opportunity on the channel.

Female

63%

Channel-wide

United States

66%

Not subscribed

98%

On your biggest video. Nearly every viewer was new.

Ad rate

Premium

Food advertisers bid hardest for this exact viewer

Why it matters

Older, American, food-focused viewers are the audience advertisers bid hardest for. You already have them, and you are still earning a middle-of-the-road rate. Fix 04 gets far more of the ads you already sold to actually play, and that is most of the gap.

On top of that, you have not touched a younger audience at all yet, so there is a second audience sitting there you have never gone after.

05 · The planSmall enough to actually run

One video a week. Saturday, 10:30am PT.

That slot is not a guess. It came straight out of your own audience tab, and it is when most of your subscribers are online and when they stay longest.

Bank a backlog

Four to six finished videos before the streak starts. One a week at quality is harder than it sounds.

Hours on packaging

Thumbnails are close to as important as the video. A thumbnail swap has taken a client video from 19,000 views to 750,000 in a week.

Rebuild the description

Playlists, top videos, affiliate links, business email, socials, hashtags and a keyword paragraph on every upload.

Study competitors on purpose

Pick a handful of channels that beat you on one thing each, whether that is titles, thumbnails or intros, then copy that one thing on purpose.

Read every retention graph

Your engagement tab shows you the exact seconds where people lean in and where they drop off. It is the best free feedback on the platform, and you are not reading it.

06 · What changed

Revenue per thousand views more than doubled.

Your audience did not change and neither did how you shoot. What changed is your ad placement, your descriptions, your tags and your thumbnails.

An honest caveat before you get too excited: ad rates across the whole platform climb toward the end of the year, so some of this jump is just the calendar. The search and ad placement gains below are ours.

Report 01

$5.71

RPM, trailing 90 days

2.2×

Report 02 · Oct 2021

$12.62

RPM, trailing 90 days. Individual videos hit $14.01 and $15, with days as high as $19.

What that is worth

At a $12.62 RPM, every million views is now worth more than double what it was in Report 01. Same channel, same recipes, same viewers who were already showing up. The only thing that moved was the setup around them.

Subscribers

26,294

Up 2,379

Lifetime views

3,023,258

Up 223,795

Search traffic

13.5%

Up 4.5 points from 9%. The tag and description work landing.

Avg view duration

6:57

Down a little, and on purpose. More short videos in the mix.

Revenue per 1,000

$12.62

Was $5.71

Chef Ange daily views July to October 2021: flat through summer then a step up from mid-September — Your daily views, July to October 2021. Flat all summer, then **a step up from September 18** that held, as the packaging and description work started landing.

YouTube retention graph for the steak video showing 60% average percentage viewed — Retention on the steak video. **60% average percentage viewed**, with 73% still there at 0:30 and no sharp drop anywhere in the first five minutes. That is the shape you want.

07 · What became of fix 01Checked against the Internet Archive

The title did change. Here is the receipt.

Fixes 01 and 08 said to drop the episode numbering, lead with the words people actually search for, and make the title a reason to click. Your biggest video carries that title today.

The Internet Archive photographs public web pages and keeps the copies. Both snapshots below are your own video page as it really looked on those two dates, so none of this rests on my memory or yours.

Archived December 30, 2020

~~Episode 1 - Oven Roasted Crispy Chicken Thighs~~

2,669,533

Views at that snapshot

Archived December 27, 2022

The Best Chicken Thigh Recipe EVER!

3,378,556

Views at that snapshot

What this proves, and what it does not

It proves the change happened and that it stuck. The episode number is gone, the search words lead, and the title now says exactly what the video gives you.

What it cannot prove is how many of the 709,023 views between those two dates came from the new title. The Archive only photographed the page twice in that stretch, and the video was pulling in views the whole time either way. Nobody can split that number cleanly, and anybody who says they can is guessing.

That video sits at 3,692,960 views today and is still the biggest thing on your channel.

Read off the live video page, September 2026

08 · What was still openThe honest part

One number refused to move: click-through rate.

An audit that only reports wins is a sales pitch. Your click-through across the channel came in at 3.2%, about half a point below the first report, and that is despite your thumbnails clearly getting better.

In Report 01 I asked you for 6%. I am resetting that to 5%, because hitting a number you can actually reach matters more right now than chasing the one I picked first.

Steak video3.3%

Mashed potatoes2.0%

Channel, trailing 90 days3.2%

Report 02 floor, reset from 6%5%+

Chef Ange thumbnail: Garlic Parmesan Wings, sharp foreground with a blurred background and hands reaching in — Garlic Parmesan Wings, posted 25 September 2021. Sharp in front, blurred behind, hands reaching in. **You know what the food is in under a second,** which is all the time you get in a feed.

Chef Ange thumbnail: How to Grill the Perfect Chicken, pieces of grilled chicken close together on a grill — How to Grill the Perfect Chicken, posted 28 August 2021. **Much harder to read at a glance.** Spread the pieces apart and the texture and the shape come through. Then people know what it is without reading the title.

The next block of work

Retention and revenue are both handled now, and search is climbing. Click-through is the one thing left, and it lifts everything else with it. Every extra click is close to free, because YouTube is already showing your videos to millions of people whether they click or not.

Both thumbnails above are yours, posted four weeks apart. That gap is the whole job from here: make the good one the normal one.

09 · Five years on

You still run the same playbook.

You have passed 176,000 subscribers across 359 videos. The packaging habits this audit asked for are simply how your channel looks now: one subject you can read instantly, the keyword leading the title, and the promise spelled out in full.

Chef Ange thumbnail: Roasted Garlic Hummus — Keyword first, and you can tell what it is in under a second.

Chef Ange thumbnail: BBQ Beans — Two words, full frame, nothing else fighting for attention.

Your help early on allowed me to put some timeless best practices in place that got me to this point.

Angelo Todaro · Chef Ange · 176K subscribers

Every finding above is the one I gave you in 2021, unchanged, and a little of it has aged. Tag fields count for much less now, because YouTube reads your title and description instead.

But matching the words a viewer actually types has never stopped working. Neither has the method behind all of it: watch the work, read the numbers, and fix them in the order that pays you back.

ChefAnge.

How I did this

What this covers