Key takeaways
- YouTube's own upload-schedule guidance never gives a number. It asks you to weigh how often you can upload and whether that rate is sustainable long term, then to be consistent about it.
- The largest public dataset — vidIQ's study of over 10.2 million channels — found median growth rising at every step up in cadence, with 12+ uploads a month showing 4.1 times the median monthly view growth of channels posting less than once a month.
- A second study, of 18,000 channels split by niche, found the opposite tail: posting above your niche's benchmark is associated with fewer subscribers, not more. There is a ceiling, and it moves by category.
- Channels that actually broke out were not uploading constantly. Long-form-first breakout channels published a median of seven times in 30 days — about 1.6 a week.
- The binding constraint is almost never the upload slot. It is idea supply, production hours and packaging throughput — how many title-and-thumbnail pairs you can make that are actually worth clicking.
- Raising cadence by one rung and holding it for eight weeks is a real experiment. Doubling output for a fortnight and then vanishing is the failure mode the data punishes.
Most creators ask this question in the wrong units. How often should I post — once a week, twice, daily? Stated that way it sounds like a scheduling problem with an answer somewhere in YouTube's documentation, and the internet obligingly supplies one: post more, post consistently, the algorithm rewards activity.
The trouble is that the three largest public analyses of upload frequency point in three different directions. One says growth rises at every increase in cadence. One says exceeding your niche's normal rate correlates with materially fewer subscribers. The third, which looked only at channels that had recently broken out, found them uploading less than twice a week. None of these is wrong. They are measuring different things, and the gap between them is where the actual answer lives.
What follows is what YouTube says, what the data says, why the studies disagree, and a procedure for picking a number that fits your channel.
What YouTube actually says
YouTube's Help Centre has a page called Upload schedule tips, and it is conspicuous for what it omits. There is no recommended number of videos per week. There is no minimum. Instead it frames the decision as two variables you set yourself: frequency — how often you upload per week or month, and whether that rate is sustainable over the long term — and consistency, meaning whether you are predictable about both when you publish and what you publish. It suggests tying content to specific days of the week, and it explicitly puts your own wellbeing in the same paragraph as the schedule.
That framing is not evasion. It reflects how the recommendation system works. Impressions are served per video, to individual viewers, based on what that viewer has watched and what similar viewers did with your video when it was shown to them. There is no channel-level dial labelled "posts often" that boosts everything you publish. A video that satisfies the people it reaches gets shown to more of them; a video that does not, does not. Your cadence changes how many times a week you enter that lottery. It does not change the odds of any single ticket.
This is why "the algorithm punishes you for not uploading" is a myth with no mechanism behind it. Nothing tracks the gap since your last upload and applies a penalty to the next one. What happens when you stop for months is more ordinary: your videos stop being fresh candidates for the home feed, your audience's habits drift, and the system has to re-learn who wants your work. That is a cold-start problem, not a punishment, and it is why the return video from a long break often underperforms and the third one does not.
The big correlational study: more is more
The largest public dataset on this question is vidIQ's analysis of 10,210,278 channels with at least 1,000 subscribers. They calculated each channel's average monthly uploads from April 2025 through March 2026, counting long-form and Shorts together, then sorted channels into five cadence tiers and compared median monthly view and subscriber growth across them.
Median growth rose at every step up. The headline comparison: channels averaging 12 or more uploads a month had 2.18% median monthly view growth against 0.53% for channels posting less than once a month — 4.1 times as much. Median subscriber growth over the same comparison went from 0.00% to 0.91%. Against a more realistic baseline, channels posting one to three times a month, the 12-plus group still had 2.45 times the median view growth (2.18% versus 0.89%). The middle tier, four to seven uploads a month, sat at 1.32% view growth and 0.39% subscriber growth — above every lower tier.
| Average uploads per month | Median monthly view growth | Median monthly subscriber growth |
|---|---|---|
| Under 1 | 0.53% | 0.00% |
| 1 to 3 | 0.89% | — |
| 4 to 7 | 1.32% | 0.39% |
| 12 or more | 2.18% | 0.91% |
Why a rising median is not an instruction
Read literally, that table says upload three times a week. Read carefully, it says something narrower: among channels that already exist and already have a thousand subscribers, the ones publishing more are also growing faster. Which way the arrow points is the entire question, and a cross-sectional median cannot answer it.
Three explanations fit the same numbers. Uploading more could cause growth. Growth could cause uploading more — a channel whose videos are working gets revenue, motivation and often an editor, all of which raise output. Or competence could cause both: creators who have worked out their format, their audience and their packaging tend to produce faster and grow faster simultaneously. The study's own framing is honest about this, describing the high-cadence tier as an opportunity rather than a guarantee and noting that it holds only while the videos still satisfy viewers. There is survivorship in the sample too — channels below 1,000 subscribers are excluded, and the ones that burned out at four uploads a week and quit are not in the current-growth figures either.
The study that found a ceiling
AIR Media-Tech ran a study of roughly 18,000 English-language channels tracked from March 2025 to March 2026, cut by 11 niches and four channel-size segments. Their top-line conclusion agrees with vidIQ — posting more mostly helps — but with a condition that changes the advice completely: "more" is defined per niche, and going past your niche's benchmark is associated with worse outcomes, not better ones.
The spread between niches is enormous. Gaming and kids content sit at 20 to 40 videos a month. Travel sits at 2 to 4. Most other categories peak somewhere between 2 and 10. And the penalty for overshooting is not subtle: in health and fitness, channels posting 10 to 20 times a month had 55% fewer subscribers than channels posting 2 to 5 times.
Why the optimum moves by niche
Cadence benchmarks track two things: how fast the underlying material renews, and how much production each video needs. A gaming channel can turn a session from last night into a video with a day's work, and its audience's appetite refreshes as fast as the games do. A travel video requires travel. A fitness video competes with a library of near-identical videos that already rank, so the marginal upload adds little and dilutes a lot. When you exceed the rate your niche supports, you are not adding videos — you are adding thinner versions of videos you already made.
What the channels that broke out were doing
A third analysis approaches it from the other end. OverseerOS examined 168 channels that had recently produced breakout videos and asked what their upload rate looked like. Long-form-first channels published a median of seven times per 30 days — roughly 1.6 a week — and 59.6% of them uploaded eight times or fewer. Shorts-first channels ran at a median of 23, about 5.4 a week. Higher-frequency channels did accumulate more breakout hits in total, but not proportionally more: the extra uploads bought extra chances at a lower rate per upload.
That last detail is the one worth sitting with. If doubling your uploads doubled your hits, cadence would be a straightforward lever and everyone should pull it. It does not. Somewhere between the first video a week and the fourth, the quality of each additional idea, script and package falls faster than the count rises.
Reconciling the three
Put them together and a coherent shape appears. At the bottom of the range, cadence is genuinely the binding constraint. A channel publishing less than monthly is barely in the recommendation system's working set, it has almost no data to learn from, and the creator has no iteration loop — eleven data points a year is not enough to learn anything about your own audience. Going from monthly to weekly is close to free, and every dataset agrees it helps.
In the middle of the range, cadence and quality trade off against each other, and where the crossover falls depends on your niche, your production burden and how much of the work is you. The vidIQ curve measures channels sitting at various points on that trade-off; the AIR study measures what happens when you push past it. At the top, volume becomes a different business — a team, a pipeline, or a format cheap enough to reproduce. That works, and it is what the gaming and kids benchmarks reflect. It is also where the failure modes cluster, because a pipeline built for volume is exactly what YouTube's monetisation policies now scrutinise.
The four things that actually cap your cadence
Almost nobody is limited by upload slots. The real constraints are these, in roughly the order they bite.
Idea supply
Every channel has a rate at which it generates ideas worth a viewer's twelve minutes. It is usually lower than people admit, and it is the first thing to break when cadence rises. The symptom is a second video each week that is a variation on the first, or a topic picked because it was Thursday. If you cannot name your next four videos and say why someone would click each one, your idea supply — not your editing speed — is your limit.
Production hours
Do the arithmetic with your own numbers rather than someone else's averages. Take the hours a finished video costs you end to end — research, filming, editing, packaging, uploading — and multiply by your proposed weekly count. If that exceeds the hours you can give the channel in your worst week of the month, not your best, the schedule is fiction. A cadence you can only hit while nothing else is going on is not a cadence.
Packaging throughput
This is the constraint people forget, and the one that most directly caps returns. A video's ceiling is set by its title and thumbnail before anyone watches a frame of it: no amount of upload frequency compensates for a package nobody clicks. Doubling your output means doubling the number of genuinely distinct, genuinely clickable packages you can produce — and packaging is the part most creators do last, tired, in the twenty minutes before publishing. A second upload each week packaged in that state does not add a second chance. It adds an impression sink.
The iteration loop
You learn from published videos only if you look at them. Every upload should produce a reading: what the click-through rate did, where retention broke, which thumbnail won. If you publish faster than you can absorb what the last one told you, you are not running four experiments a month — you are running one experiment four times without reading the result. The shape of the retention graph is where most format decisions actually come from.
Does a second video cannibalise the first?
Partly, and it matters more than the frequency debate usually allows. Impressions are personalised per viewer, but your subscribers are a finite pool with a finite appetite, and the subscription feed, notifications and home-feed slots you occupy are shared between your own uploads. Publish twice in three days and the second video is competing with the first for the same attention on the same people, at the moment the first is still accumulating the early signals that decide how far it travels.
The effect is smallest when your videos reach genuinely different audiences — different topics, different search intents, different formats — and largest when they are near-duplicates aimed at the same viewer. A thumbnail test also needs enough impressions to resolve, and a video's first week is when most of its discovery momentum is decided; splitting a fixed audience across more uploads in that window makes every one of those signals noisier. So if you are moving from one video a week to two, space them, and make them different enough that a viewer would plausibly want both. Two near-identical videos three days apart is the worst configuration available.
Shorts and long-form do not share a cadence
Mixing formats changes the question, because Shorts are cheap to produce and are distributed through a feed with a different appetite. AIR's companion study of the same 18,000-channel dataset found the optimal Shorts share sitting between 25% and 40% of total uploads, with the benefit varying sharply by niche: in fitness, food, business and education, channels mixing Shorts with long-form had more subscribers than long-form-only channels, while in gaming, entertainment and science the long-form-only channels showed higher medians.
Size matters too, and it reverses. Among food and cooking channels, mixing in Shorts correlated with 31% more subscribers at smaller sizes — 104,000 against 79,400 in the 100K–1M monthly views band — but at 10M+ monthly views the long-form-only channels were 47% ahead. Shorts look most useful as a discovery instrument early, and least useful as a long-form substitute later.
So do not count your uploads as one number. Count them as two schedules with different economics and different production costs. The question of whether Shorts help your channel at all is worth settling before you use them to inflate a cadence figure.
What consistency buys, and what it does not
Consistency is the most repeated advice on this topic and the most casually justified. It is worth being precise about the mechanisms, because two of the three commonly cited benefits are real and one is not.
What consistency genuinely does: it builds an audience expectation, so your subscriber base becomes a habit rather than a list; it makes production sustainable, because a fixed slot forces a process instead of a sprint; and it gives you comparable data, since videos published under similar conditions can be measured against each other.
What it does not do is earn a ranking bonus. There is no schedule-adherence signal. Publishing every Tuesday at 9am does not make Tuesday's video rank better than the same video published on a Thursday; it makes it easier for the people who already like your work to find it. That is a real benefit, and it is a distribution benefit rather than an algorithmic one. If you want to think carefully about timing itself, the mechanics of when to publish are a separate decision from how often.
The ceiling YouTube enforces
There is a hard limit on the volume strategy, and it is written into monetisation policy. In July 2025 YouTube renamed its "repetitious content" rule to "inauthentic content" and clarified what it covers: material that follows a template with minimal variation, is easy to reproduce at scale, and lacks meaningful author input. YouTube framed it as a clarification of long-standing requirements for original and authentic work rather than a new restriction, and for most creators nothing changed. But it draws the line clearly for anyone whose cadence plan depends on a reproducible pipeline.
The distinction that matters is not how fast you produce, or whether AI is involved in the process. It is whether each video carries something a person added. A high-volume channel where every upload is a genuinely different idea, researched and packaged deliberately, is ordinary YouTube; one where the template is the product is what the policy is aimed at. If your route to a higher cadence runs through making each video more like the last one, you are heading toward the ceiling rather than the growth curve.
Choosing your number
A procedure, rather than a recommendation, because the recommendation depends on facts only you have.
- Find your niche's benchmark. Take the eight to twelve channels you actually compete with — similar size, similar topic, currently growing — and count their uploads over the last 90 days. The median of that set is your benchmark, and it beats any published figure because it is measured on your competitors rather than on the average of all YouTube.
- Compute your honest capacity. Hours per finished video, multiplied by the cadence you are considering, against the hours you can commit in a bad month. If it does not fit, the number is wrong regardless of what the benchmark says.
- Take the lower of the two, then subtract one if your last three videos were rushed. A cadence you hit with something left over is a cadence you can hold for a year.
- Move one rung at a time. Monthly to fortnightly, fortnightly to weekly, weekly to twice weekly. Hold the new rate for at least eight uploads before judging it.
- Judge it on the right metric. Not total channel views, which almost always rise when you publish more, but median views per video and subscribers gained per video. If those fall by more than the count rose, the extra upload is costing you.
| Situation | Sensible starting cadence | What to fix before raising it |
|---|---|---|
| New channel, no format settled | Weekly long-form | Finding a format that gets past 40% average view duration |
| Format works, growth is flat | Weekly, plus packaging rework on the back catalogue | Click-through rate on the last ten videos |
| Solo creator, videos take a full week | Weekly, or fortnightly with better ideas | Production time per video, before anything else |
| Small team or cheap format | Two to three a week | Whether video two is genuinely distinct from video one |
| Gaming, kids, news, reaction | Niche benchmark, which may be daily | Whether each upload still has an author's input in it |
What to do with the slot you decided not to fill
Choosing not to add an upload is only sensible if the hours go somewhere. There is a long list of work with a better return per hour than a rushed extra video.
- Repackage the videos you already have. A video with good retention and a poor click-through rate is a finished asset with a broken cover. Swapping the thumbnail on an old upload is a few minutes of work against a video that already proved people finish it — the mechanics of changing a thumbnail after upload including what it does to a video's existing performance, are worth knowing before you do it in bulk.
- Test the packaging on the next one properly. Two or three real thumbnail options, compared, rather than the first thing that exported.
- Fix your opening thirty seconds. Retention losses compound across a channel; a hook fix improves every future video rather than one.
- Build the sequences. Playlists, end screens and a channel page that hands a new viewer a second video turn one good upload into a session.
Breaks, and the constraint nobody puts in the spreadsheet
Cadence advice tends to assume an infinitely available creator. A survey by Tasty Edits that reached out to 29,000 content creators found 79% saying they had experienced creator burnout, with 22% describing themselves as stressed "very often" or "constantly". Those numbers are the real reason the upload-frequency question deserves care: the most common way a channel stops growing is not a bad thumbnail but a creator who quietly stops.
YouTube's own guidance treats this as part of the scheduling decision rather than a separate wellbeing footnote, which is the correct framing. A break costs you the cold start described earlier — a return video that underperforms, an audience that takes a few uploads to re-engage. That is a real cost and a recoverable one. Choosing a cadence you cannot hold, and discovering it eight months in, costs you the channel.
If you need to step away, the version that costs least is a planned one: say so, give a date, and come back with a video aimed at your existing audience rather than a stranger. Announcing nothing and disappearing is the expensive version, and it is the one most people choose.
Common questions
Does uploading daily hurt your channel?
Not inherently — several niches have benchmarks at or above daily. It hurts when daily output means each video is thinner than the one you would have made weekly, when the videos are near-duplicates competing for the same viewers, or when the pipeline that makes daily possible strips out the human input that monetisation policy requires. Daily is a production capability, not a strategy.
Will the algorithm punish me for missing a week?
There is no penalty mechanism for a missed week. What you lose is momentum with your own audience and one entry in the recommendation lottery. A single missed week is close to invisible; a pattern of unpredictable gaps costs you the habit your subscribers would otherwise form.
Should I delete or unlist my underperforming videos to raise the average?
No. Recommendations are decided per video, and the videos that did poorly are not dragging down a channel-level score. A weak old upload occasionally reaches someone; deleted, it reaches nobody. Repackage it or leave it alone.
Is it better to batch four videos and publish them weekly, or publish as they finish?
Batch and schedule. It smooths your worst weeks and means the packaging on video four is not being done at midnight. YouTube's own scheduling tools exist precisely so that consistency does not require you to be at a keyboard on publish day.
Do Shorts count towards my upload frequency?
In the studies above they were counted, which is part of why high-cadence tiers look so healthy — a channel at 12 uploads a month is often publishing one long-form video a week and a couple of Shorts. Track them separately. A Short and a twelve-minute video are not interchangeable units of output, and treating them as one number is how creators convince themselves they are publishing three times a week when they are publishing once.
The answer, stated plainly
For most channels: one long-form video a week, published on a fixed day, held for at least two months before changing anything. It is above the tier where cadence is clearly the constraint, it is within reach of a solo creator with a job, it matches what breakout long-form channels were actually doing, and it leaves enough time that the packaging on each video gets made while you are awake. Add Shorts if your niche rewards them, at a quarter to a third of your uploads. Raise the long-form rate only when the last eight videos were finished early rather than late.
The reason so much of this comes back to packaging is that cadence multiplies whatever your average video is worth, and the title and thumbnail set that value before the video gets a chance to. A weekly channel with packages people click will outgrow a thrice-weekly channel whose thumbnails were an afterthought, every time, and it will do it on a third of the work. If packaging is the thing bottlenecking your schedule rather than filming or editing, that is a solvable problem — it is the reason Thumblore exists, and it is why the option worth reaching for is usually a better cover on this week's video rather than a second video.
If you want the deeper version of the packaging half of this, the complete thumbnail playbook covers the design decisions in order, and the thumbnail preview tool will show you what this week's upload looks like at the sizes people actually see it — which takes about a minute, and is worth more than the extra video you were thinking about squeezing in.