New, Casual and Regular Viewers: How to Read YouTube's Loyalty Mix, and What Each Segment Wants From Your Packaging

Somewhere in your Studio analytics is a number that almost every creator reads exactly once, decides is broken, and never opens again. It sits in the Audience tab, it is labelled regular viewers, and on most channels it is startlingly small — a rounding error against the subscriber count in the corner of the screen.

The number is not broken. It is measuring something the subscriber count has never measured: how many distinct human beings came back to your channel month after month, over a year. Almost everything creators argue about — whether to chase a trend, whether to change the thumbnail style, whether posting weekly matters, whether a viral video was worth anything — has a cleaner answer once you can see your audience split into the people who just met you, the people who drop in, and the people who have made you part of their week.

This is how that split is calculated, what each of the three segments is actually telling you, and what changes in the work depending on which one is missing. It is also, unavoidably, an article about thumbnails, because the three segments meet your packaging in three completely different states of mind.

What the three segments actually count

The starting point is the monthly audience figure. YouTube's documentation on unique viewers defines it as the total unique viewers who watched your content in the last 28 days — deduplicated, so the same person watching on a phone in the morning and a television in the evening counts once. That 28-day pool is then split three ways.

Segment	What YouTube counts	What it is really telling you
New	Viewers who watched your channel for the first time in the selected period.	How much reach you are winning outside your existing audience — and therefore how well your packaging works cold.
Casual	Viewers who watched at least once a month in one to five months of the past year.	People who know the channel exists and choose it occasionally. The largest recoverable audience most channels have.
Regular	Viewers who watched at least once a month in six or more months of the past year.	Your actual core. Habit, not intention — the group that would notice if you stopped.

Two structural details change how the numbers should be read. The first is that the segments are defined over a rolling twelve-month history but reported against a 28-day audience: a regular viewer is someone with a year-long pattern who also happens to have watched you this month. The second is that the split is channel-level. There is no per-video version of it, which means it cannot tell you which upload converted a stranger into a repeat viewer. It can only tell you whether the conversion is happening at all.

Both details push the metric towards one honest use: a quarterly read on the shape of your audience, not a number to check on upload day. YouTube's own guidance on the segments treats them the same way — as an input to content strategy, where the suggested moves are to experiment with new topics when you want new viewers, and to publish consistently on familiar topics in familiar formats when you want regulars.

Why your regular viewer count looked like a collapse

Until 2025, Studio reported new versus returning viewers. Returning was a loose category: it meant broadly that someone had watched you before. When YouTube replaced that pair with the three-way split during July 2025, as reported at the time, the definitional bar moved a long way up. Returning asked whether you had ever been back. Regular asks whether you have been back in half the months of the year.

YouTube was explicit that the new segments are not a direct equivalent of the old metric, and the practical consequence is a number that appears to have fallen without anything happening to the audience. Nobody left. The definition tightened, and the part of your audience that had been sheltering inside a generous word got sorted into casual instead.

This matters beyond bookkeeping, because it changed what the metric is good for. Returning viewers rewarded any repeat visit, including the accidental one — a viewer who saw two of your videos in the same week on Home counted as loyal. The regular segment only counts sustained behaviour, which is the thing creators actually want to know about and the thing no other number in Studio reports. It is a worse vanity metric and a much better diagnostic.

The subscriber count is not this number, and never was

The gap is worth spelling out because it is the root of most misreadings. A subscriber count is a lifetime, cumulative total of people who once pressed a button. It does not decay when someone stops watching, it does not distinguish an account abandoned in 2019 from a viewer who watched you yesterday, and it does not require anyone to have watched a second video.

The regular segment is the opposite in every respect: current, behavioural, and indifferent to whether the viewer ever subscribed. Some of your regulars are not subscribed at all — they arrive through Home or search, month after month, because the recommendation system has worked out that they watch you. Some of your subscribers have not seen a video of yours in a year and are counted nowhere in the monthly audience, because they are not in it.

If the gap between the two numbers bothers you, the delivery mechanics behind it are worth reading in full — we worked through what a subscribe click does and does not buy in why your subscribers don't see your videos. The short version is that subscription is a soft, decaying signal of interest, and the loyalty segmentation is what that signal looks like once you measure behaviour instead of intent.

The three segments meet your thumbnail in three different states

Here is where the metric stops being an analytics curiosity and starts changing design decisions. The same 1280×720 image is doing a different job for each segment, because each segment arrives carrying a different amount of prior knowledge about you.

Segment	Where they usually meet your packaging	What the thumbnail has to do
New	Home, search results, suggested videos beside someone else's content, the Shorts feed.	Work with zero brand equity. Read at phone size in under a second, state the subject, and make the promise legible without any knowledge of who you are.
Casual	Home, subscriptions feed, search — usually recognising the channel before reading the frame.	Be identifiable as yours, then answer a second question: is this one for me? Recognition alone does not earn the click.
Regular	Subscriptions feed, notifications, Home — often looking for you specifically.	Very little. They are clicking on the channel, not the image, and they will forgive a weak frame.

That last row is the one that quietly distorts channels. Regular viewers click at rates no cold audience will ever match, and they click early — in the first hours after publication, when the video is shown mostly to people who already watch you. If you judge a thumbnail by day-one click-through rate, you are mostly measuring the enthusiasm of people who would have clicked a grey rectangle with your name on it.

YouTube's documentation on impressions and click-through rate says this outright: it is common for videos with fewer impressions and views to have higher click-through rates and higher average view duration, because they have been shown to a narrower, more loyal audience, and a video that reaches the home feed will naturally see its CTR fall as impressions grow. The same documentation is where the only real benchmark comes from — half of all channels and videos sit between 2% and 10% impressions CTR, which we unpack in what is a good CTR on YouTube.

Reading your own mix: four shapes and what each one means

The absolute numbers are close to meaningless across channels — a cooking channel with a 200,000-strong monthly audience and a niche software channel with 4,000 cannot be compared. The proportions are informative, and in practice they fall into four recognisable shapes.

Shape	What it usually means	What to change first
Mostly new, almost no casual	You are winning cold clicks and losing the second visit. Common on trend-chasing channels and on channels whose breakout video has nothing to do with the rest of the library.	The channel page and the next-video path, not the thumbnails. Make the catalogue legible: a trailer, sensible playlists, end screens that point somewhere relevant.
Healthy casual base, thin regular sliver	The normal, functioning shape for a growing channel. People like you enough to come back sometimes; nothing has yet made you a fixture.	Rhythm and series. Recognisable formats, a predictable slot, and a reason to expect the next one.
Mostly regular, very few new	A closed channel. Comfortable, high CTR, flat views. The packaging is working on people who already know the code and failing on everyone else.	Packaging for strangers, and topic range. This is the shape that stagnates for years while the analytics look fine.
New spike, everything else unchanged	One video broke out to an audience with no interest in the channel behind it. A month later the mix looks exactly as it did before.	Nothing, in a hurry. Judge it at the next quarter's read, and resist rebuilding the channel around a video that brought no one back.

Two of those four shapes are diagnosed wrongly almost every time. The closed channel reads its own CTR as proof the packaging is excellent, when the CTR is high precisely because the reach is small. The new-spike channel reads a viral video as a mandate, and spends three months making imitations of it for an audience that was never there. Both errors are visible in the loyalty mix and invisible in the view count.

New to casual is a packaging problem

A new viewer has to decide about you twice: once about the video, once about the channel. The first decision happens in the thumbnail and title, cold, against everything else on the screen. If that decision keeps going your way and the casual segment still does not grow, the leak is at the second decision, and the second decision is made somewhere other than the video.

The practical checks, in the order they usually pay off:

• Does the frame work at the size it is actually shown? Most impressions are small, on a phone, in a scrolling feed. The subject has to survive the shrink, which is a composition and contrast problem rather than a detail problem — the rules that transfer are in thumbnail composition.
• Does a stranger know what the video is about without knowing you? Inside jokes, recurring characters and running gags cost nothing with regulars and cost everything with new viewers. They are the most common reason a packaging style that "always worked" stops growing the channel.
• Where does the video send them next? End screens, a pinned playlist, a channel page that is not a wall of unrelated titles. A new viewer who wants more and cannot find a second thing to watch does not become casual.
• Is the next video recognisably from the same channel? Not identical — recognisable. The distinction is the whole argument of thumbnail consistency, and it is what makes a second encounter feel like a return rather than a first meeting all over again.

Topic range belongs in this list too, uncomfortably. YouTube's own advice for growing the new segment is to experiment with topics, which is at odds with the advice for growing regulars — publish consistently on similar topics in a familiar format. Those two instructions genuinely conflict, and the resolution is not to split the difference on every upload. It is to know which leak you are fixing this quarter.

Casual to regular is a habit problem

Nothing in a thumbnail turns an occasional viewer into a habitual one. What does it is predictability: a format they can anticipate, a slot they learn without being told, and a reason to open YouTube expecting you rather than discovering you.

YouTube's guidance on the segments points in the same direction, and it is short: upload consistent content on similar topics or in a familiar format, use community features — posts, comments, premieres and live chat — and keep branding consistent across the channel. Plainly stated, none of that is a growth hack. It is the unglamorous machinery of becoming a fixture, and it is measured in months because the metric itself is measured in months.

Three things make the difference more often than the rest:

1. Series over one-offs. A numbered or named format gives a casual viewer something to follow rather than something to encounter. It also gives the thumbnail a job it can do well: signal which series this is, and which instalment.
2. A rhythm they can learn. Not frequency for its own sake — a predictable rhythm you can sustain. The throughput trade-offs are worked out in how often should you post.
3. Surfaces that are not the upload. Community posts, comment replies, live chat and premieres are the parts of the platform that create contact between videos. YouTube is building more of this deliberately: the 2026 announcements included members-only community spaces, group messaging, richer comments and AI-assisted moderation, which we catalogued in Made on YouTube 2026.

How to read the mix without over-reading it

The segmentation is coarse by design, and three habits stop it being misused.

Read it quarterly, against a comparable window

A metric built on twelve-month watch patterns cannot respond to a change you made on Tuesday. Look at it every three months, compare like with like — a month against the same month last year rather than against December — and expect movement in the mix to lag the work that caused it by a season.

Pair it with traffic sources and subscription status

The loyalty mix tells you what happened; the other two reports tell you where. Traffic sources separate the surfaces that bring strangers from the ones that bring your existing audience, which is the subject of traffic sources explained. Advanced mode then lets you break a report down by subscription status, so you can see whether a thumbnail is earning clicks from people who already know you or from people who do not. That single breakdown answers more packaging questions than the loyalty mix ever will.

Do not tune packaging to the segment you already have

Thumbnail tests are decided by whoever is watching while the test runs, and early in a video's life that skews towards regulars. If you are testing to grow, the result you care about is the one that survives contact with cold traffic — which means letting a test accumulate impressions beyond your core before you believe it. The mechanics, including how the three-variant test allocates traffic, are in Test & Compare.

The Shorts complication

A channel that grows through Shorts will usually see a new-viewer-heavy mix, and the reason is structural rather than a failure of the content. The Shorts feed delivers impressions in volume to people who did not choose the channel, and it does not require anyone to visit a channel page on the way through. Enormous numbers of first-time viewers, very little of the second decision.

That does not make Shorts worthless for loyalty; it makes the bridge the thing to engineer rather than assume. The evidence on what actually crosses from short-form to long-form, and what does not, is in do YouTube Shorts help your channel. Read against the loyalty mix, the question becomes concrete: if three months of Shorts have added new viewers and moved the casual segment not at all, the Shorts are working and the bridge is not.

What the metric cannot tell you

It is worth being clear about the limits, because the segmentation invites over-interpretation.

It cannot attribute. No video is credited with converting a new viewer into a casual one, so the mix will never hand you a list of your best entry points. The nearest proxy is a video's share of new versus returning traffic in its own reports, plus the subscription-status breakdown, plus whether the channel page and end screens actually offered a second thing to watch.

It cannot tell you about people who left. Viewers who lapsed outside the 28-day window are simply absent, not counted as a loss. A channel can watch its regular segment hold steady while its casual segment quietly empties, and the only visible symptom will be flattening views.

And it says nothing directly about recommendations. Being a regular viewer is not a ranking signal you can farm; it is a description of behaviour that the recommendation system has its own, much richer view of. The reason to grow the segment is not that YouTube rewards the label. It is that a channel with a real core audience survives the months when nothing breaks out.

A quarter's worth of work, in order

If the mix has told you something uncomfortable, the sequence below is roughly the order in which the fixes pay.

1. Write down the shape. Monthly audience, and the three shares, today. You need the baseline before you change anything, and the metric moves too slowly to reconstruct later.
2. Name the leak. One of them, not both. New-to-casual and casual-to-regular need different quarters.
3. If it is new-to-casual: audit the last ten thumbnails for legibility to a stranger, fix the channel page and end screens, and pick one recognisable packaging system you can hold for a quarter. A full pass on the rest of the channel is laid out in how to audit a YouTube channel.
4. If it is casual-to-regular: commit to one series with a name, a slot and a thumbnail pattern, and use community posts between uploads so the channel exists in the weeks you do not publish.
5. Leave the metric alone for ninety days. Then read it again, in the same window, next to traffic sources and subscription status.

None of this is fast, which is the point of a metric built on watch-months. The value of the loyalty mix is not that it tells you what to do this afternoon. It is that it is the only number in Studio that distinguishes a channel getting bigger from a channel getting busier — and those two things feel identical from the inside for a surprisingly long time.

Where packaging fits in that picture is narrow and important. It cannot make anyone a regular, and it does not need to. It is the only thing that works on a person who has never heard of you, which makes it the entire first step of a funnel whose later steps are made of rhythm, format and contact. If the cold half of that job is where your channel is leaking, Thumblore exists to make the frame that a stranger can read in under a second — and the arguments behind what that frame should contain are in the high-CTR rules and the best colours for thumbnails.

Check the mix once a quarter. Fix one leak at a time. And when the regular number looks small, remember what it is asking of a stranger: six months of choosing you, on a platform built to offer them something else every few seconds.