How Many Videos Per Month Does a DTC Brand Actually Need?

Most answers to this question are a number somebody made up. We went looking for the sourced ones, and found that the two largest studies
reach opposite conclusions.

We are a production company. We sell video volume, so we have an obvious interest in the answer being "more." What follows includes the evidence against that, and the point at which the honest answer becomes "nobody knows."

Short answer

There is no universal monthly video number, and the honest split is between paid and organic.

For paid creative testing, one number is defensible: at the ~5% win rate Motion measured across $1.29B of Meta ad spend, roughly 20 creatives per winner on average, or about 45 attempts for a ~90% chance of at least one winner. Motion sells an ad-insights platform to teams testing creative at volume — its home page headline is "Ship more winning ads" — and this is its own dataset — it is our only source for that rate, so the whole paid calculation rests on one commercially interested measurement. For organic posting frequency there is no settled optimum: the two best-powered studies disagree, and consistency is better evidenced than volume.

Answer in one minute

GoalPlanning rangeEvidencePaid creative testing~20 creatives per winner (average)Motion, $1.29B Meta spend~90% chance of ≥1 winner~45 creativesOur calculation from the 5% rateOne TikTok campaign structure9–25 creativesTikTok's 3–5 × 3–5, our multiplicationOrganic Instagram, observed~20 posts/monthSocialinsider, 35M postsOrganic optimum frequencyNot settledBuffer vs Dash Social disagreeCreative refreshBefore performance declinesMeta fatigue research, TikTok guidanceCost, at our floorsAt least $7,000 for 20 UGC videos — a lower bound, not a quoteOur published "starts at $350 per video", 5-video batch minimum

One caveat that runs through all of it: a creative is not a production. Hooks, cuts and aspect variants from one shoot each count separately, so the number of videos you commission is smaller than the creative count.

How reliable is the evidence behind these numbers?

Not equally reliable — so we ranked every source before we wrote a word. This article synthesises nine external sources from seven publishers, plus our own published rates. Eight of the nine are studies with a stated dataset; the ninth is TikTok's advertiser documentation, which is prescriptive guidance with no sample disclosed and is not a study.

TierWhat it isUsed here1Platform primary — official documentation and published researchMeta creative-fatigue research; TikTok advertiser documentation2Original research, sample disclosed—3Vendor research with a disclosed sample and method, cited directly and scopedMotion, Buffer (×2), Dash Social, Socialinsider, Metricool4Platform marketing content, even where it reports researchTikTok/System1 — published on TikTok For Business as promotional content, with footnotes to unpublished proprietary studies. Used only for the finding that cuts against us5Vendor content without disclosed methodology, forumsExcluded

Motion, Buffer, Dash Social and Socialinsider are all commercial vendors. Tier 3 is the honest place for vendor research that does disclose its sample — which all four do.

Three distinctions we held to throughout:

Observational vs experimental. Most of the frequency evidence is correlational — vendors analysing accounts on their own platforms. Two of our sources are experimental, and they point in opposite directions.

Meta ran a two-cell split test on creative fatigue (~26,000 cases, two 7-day phases) which supports refreshing creative. Meta calls it a “simple two cell split test with two phases” and never uses the word randomised; neither do we. TikTok and System1 ran a study that "deliberately induced creative fatigue by exposing consumers to a range of ad frequencies over the course of a single session" — also a manipulated design, and its headline finding argues against volume.

We flag both, because granting causal weight only to the experiment that suits us is the easiest way to mislead with a correctly-cited source.

Denominators. Engagement rates are not comparable across studies unless they divide by the same thing. Where we give a rate, we name the denominator — and where a source does not disclose its own, we say that instead of guessing.

Commercial interest, and a single point of failure. Five of the seven publishers sell software to the audience for their own finding — Motion (ad insights), Buffer (publishing and analytics), Dash Social (enterprise social management), Socialinsider (competitive analytics), Metricool (a full management platform, paid-ads management included) — and the other two, Meta and TikTok, sell the ad inventory the numbers are about. There is no disinterested source in this article. We checked that sentence against each of the seven publishers' own product pages on 21 August 2026 rather than inferring the business from the study, and every member holds.

Every planning number and every price in this article traces back to one figure: Motion's ~5% win rate, from Motion's proprietary dataset, published by a company whose business depends on advertisers testing creative at volume. We have no second source for it. That is the weakest joint in the piece and you should know where it is.

Motion sells an ad-insights platform — software that connects creative elements to performance. Motion's home page headline is "Ship more winning ads," and the platform page promises to "Make better ad creative" and to "turn performance insights into winning creative" (both retrieved 2026-08-21). The interest runs through the product category as well as the customer base, and it points the same way as ours — which matters more here than anywhere else in the article, because this is the source the whole paid calculation rests on.

Sources and their limits

SourceDatasetFindingLimitationMotion 2026$1.29B Meta spend · 578,750 creatives · 6,015 accounts · Sep 2025–Jan 2026~5% of creatives are winnersVendor. Motion sells an ad-insights platform to "performance-creative teams" and agencies, headlined "Ship more winning ads" and promising to "Make better ad creative." Its customers are the advertisers running high creative volume and its published research argues for volume, so its interest points the same way as oursMeta Analytics 2023Two-cell split test, two 7-day phases, ~26,000 casesConversion likelihood associated with ~45% drop by 4th exposure; fatigue guidance improved conversion 8%The 45% is associational; only the 8% is causalTikTok advertiser docsPrescriptive guidance3–5 creatives per ad group, 3–5 ad groupsNo sample disclosedBuffer "How Often Should You Post on Instagram in 2026"2.1M posts · 102K accountsFollower growth rises with frequency, returns diminishingCorrelational; Buffer's own customersDash Social 2025~1.5M feed posts · ~3,000 brands · Jul 2025–Jun 2026Engagement rate peaks at 2–3 posts/weekCorrelational; enterprise skewSocialinsider 202635M posts · 447,613 pages · 2025~20 posts/month; carousels 0.55% per followerFollower denominatorBuffer consistency study Jan 2025>100,000 Buffer users, all major platformsHighly consistent posters saw >5x engagement per post vs sporadic postersCorrelational; Buffer's own users; explicitly did not control for posting frequencyTikTok/System1 May 20258,000 heavy TikTok users · 6 markets · experimental — exposure deliberately varied within a session"High frequency cannot save poorly performing ads"Published as TikTok For Business marketing content; underlying study not released. TikTok sells the inventoryMetricool Jun 2026, with its trends page24,364,803 posts · 375,118 accounts · "between January–February 2025 and January–February 2026"Posting volume up 24% year on year; recommends Reels for reaching new audiences. Its two published rate movements go opposite ways: Reels engagement rate up 24.76%, "the only format growing in that direction," and single-image posts down 45.98% on engagementCited here only for scale and for the recommendation that cuts against our carousel bet. Vendor; analytics customers. The two figures sit on two different pages of one study, so an article citing one page cannot see the other (both pages checked 2026-08-24)

How often should a DTC brand post — and why do the two studies that measure it disagree?

Because the two largest studies that actually relate posting frequency to an outcome disagree, and we cannot reconcile them.

Two caveats on that sentence. Larger Instagram datasets exist — Socialinsider's 35 million posts and Metricool's 24 million dwarf both studies below — but they report what brands do, not what frequency achieves. And a number like 12 or 20 is not automatically unsourced; ours lands near 20 too.

The point is narrower: on the question of whether posting more works, the two best-powered attempts to measure it reached opposite answers.

Buffer analysed 2.1 million posts from 102,000 Instagram accounts in August 2025. Follower growth rose with frequency at every step: +0.12% weekly at 1–2 posts a week, +0.26% at 3–5, +0.44% at 6–9, +0.66% at 10 or more. The study records the taper in its own words — "Even at higher frequencies, there are still meaningful 20-25% incremental gains for posting more frequently, however the returns are diminishing."

Dash Social analysed nearly 1.5 million Instagram feed posts from roughly 3,000 brands between July 2025 and June 2026 and found a peak rather than a slope: "Brands that posted 2 to 3 times per week to their feed saw significantly higher engagement rates than those posting more or less frequently."

The tempting explanation is that they measure different things — total reach versus per-post rate — and it is partly right, which is the most annoying kind of answer. Buffer reports reach per post rising with frequency; Dash Social reports an engagement rate falling.

Both are per-post, so the obvious reconciliation fails, but the two quantities are still not the same thing, and reach per post rising while an engagement rate falls is an ordinary dilution pattern rather than a contradiction.

We cannot go further than that, because Dash Social's page never states what its engagement rate divides by — we looked, and the word appears nowhere on it. An article whose own rule is to name the denominator has to admit when the source will not.

What actually differs is the population and the method. Buffer's sample is 102,000 accounts; Dash Social's is about 3,000 brands on an enterprise platform. Both are vendor analyses of their own customer base. Neither is experimental, and neither controls for the obvious confound: accounts that post more tend to have more people and more budget, which independently drive growth.

So we will not tell you these are secretly compatible. Two large studies disagree, and the honest position is that organic posting frequency is not settled.

How many ad creatives do I need per month to find one winner?

At a ~5% win rate, about 20 creatives per winner on average — and about 45 if you want to be confident of getting one. On the paid side the arithmetic is forced by testing. Meta and TikTok both publish creative requirements, and one study with disclosed spend gives the win rate that makes the paid testing maths work.

Motion analysed $1.29 billion in realised Meta ad spend across 578,750 unique creatives and 6,015 advertiser accounts between September 2025 and January 2026. Roughly 5% of creatives are winners, where a winner reaches at least 10× the account's median spend and at least $500.

A 5% rate makes 20 creatives the average needed per winner. It does not make 20 sufficient.

At a 5% independent rate, 20 attempts produce at least one winner about 64% of the time — which means roughly a one-in-three chance of twenty creatives and nothing to show for it. For 90% confidence you need about 45.

We have seen that 20 quoted as a target; it is a mean with a large failure tail, and planning to it is planning to be disappointed a third of the time.

TikTok's advertiser documentation, last updated June 2025, is more prescriptive: "between 3-5 different creatives per ad group and 3-5 diversified ad groups per campaign." Our multiplication, not TikTok's recommendation, puts a properly structured campaign at 9 to 25 creatives before you have tested anything.

One caveat that cuts against us: Motion counts unique creatives, which in practice includes hooks, cuts and aspect-ratio variants built from the same footage. One shoot or one UGC video can yield several.

Anyone converting Motion's creative counts directly into a video order — including a production company quoting you — is overstating what you need to shoot.

Why does paid creative stop working, and how often must you refresh it?

Because repeated exposure degrades it — and this is the one part of the question with experimental evidence rather than correlation.

Meta's analytics team published research in 2023 reporting that "at 4 repeated exposures the associated likelihood of a conversion drops by about 45%." Meta's own wording is associated — that figure is a modelled association, not a causal measurement, and we are not going to upgrade it. And read the sentence before it in the same post, because the two do not agree: the model is "a logistic regression that predicts the likelihood of a click," and Meta states the result as "the reduction in likelihood of a click follows the shape of (N+1)^-.43" — then converts it in the next sentence to a drop in "the associated likelihood of a conversion." A 45% fall in click-through is a much smaller claim than a 45% fall in conversion, and nothing in the post bridges them. The inconsistency is Meta's; flagging it is ours, and it cuts against the number this section is built on.

The causal result sits underneath it. Meta ran what it calls a “simple two cell split test with two phases” across “roughly 26 thousand cases” — the first 7 days establish “a distribution of fatigue levels,” the second 7 measure what happens when new creative is added — and found that acting on fatigue guidance "can improve conversion rate by on average 8% for high fatigue cases." Meta reads its own result as causal, with its own hedge on the front: "This suggests that adding in new creative into fatigued ad sets has an overall causal improvement in conversion rates." A controlled experiment, a modest measured effect. That 8% is the strongest evidence in this entire subject. One caveat on the design: the sentence introducing the design ends on a colon and the two cells are never described — in the rendering we can retrieve, nothing follows it but the sample size. So what was held constant in the control cell is not stated, and a two-cell split test whose cells are undescribed cannot be independently appraised. Medium refuses this network on every direct rung, so we read the post through a text-extraction service — and that rendering cannot tell us whether the missing description is a figure. The post announces a figure in so many words exactly twice — "See the notional diagram below:", where the rendering does show an image marker, and "In case that isn't helpful, here a plot of that relationship:", where it shows none and goes straight to the next sentence. One of the two, so an absent marker means nothing. What we can say is that the cells are not described in any text we can retrieve, and no more than that.

For context on how common fatigue already is: "The mean count of user/creative previous exposures across all Meta ad impressions is 4.2 with over 19% of ad impressions having been seen more than five times." Two qualifiers from the same paragraph, because without them that figure is misleading: the counts are "produced over a 30 day lookback window," and repetition is "especially prevalent with ads that optimize for link clicks and offsite conversion events. By way of contrast brand advertisers have dramatically less repetition of this sort in part due to frequency caps that advertisers impose." If you are running brand campaigns with frequency caps, the 4.2 is not your number.

TikTok's refresh guidance is a trigger rather than a calendar: replace "when delivery results exhibit a consistently declining trend, or when daily new users are low."

And Meta names a remedy that costs no production at all. In the same article: "Alternatively to refreshing creative, targeting new audiences could also resolve this problem." It adds the reason that does not fully let you off — "However, the finite nature of audiences to target creates a strong motivation to develop new and differentiated creative assets that counteract this phenomenon" — but the order matters. If your audience is nowhere near exhausted, expanding targeting is the cheaper lever, and a production company omitting that sentence from the section arguing for more production would be omitting the one that costs it money.

This is the real argument for producing more, and it is narrower than the one agencies usually make. It is not that more content performs better. It is that the creative you are running degrades with exposure, and once your addressable audience stops growing, you need replacements ready before performance tells you so.

What does the evidence say against making more videos? Three findings that cut against volume

Enough that we are not going to pretend otherwise — and the sharpest is Buffer's, which measured the thing we are not selling.

Consistency is better evidenced than frequency. Buffer's separate consistency study across more than 100,000 users and all major platforms found "the highly consistent posters saw more than 5 times the engagement per post compared to those who posted sporadically" — while explicitly noting "this analysis didn't take into account the impact of posting multiple times per week." Showing up reliably was measured. Posting a lot was not.

Buffer contradicts its own headline. The frequency study quotes Buffer's resident data scientist Julian Winternheimer saying "However, there is a big caveat that the most important factor probably remains content quality", and heads the section "More does not always mean better".

TikTok says frequency cannot rescue weak creative. From research with System1 across 8,000 heavy TikTok users in six markets: "High frequency cannot save poorly performing ads."

Motion writes that "Winners are statistical outliers. The more ads an account tests, the more chances the account has to produce one" — and on Motion's own page that passage is the caption to a finding about enterprise advertisers shipping 18.8 creatives a week. It explains why volume works: more chances to surface a winner, at per-creative odds that do not change.

Motion's line is the honest frame for the whole question. Volume does not improve your creative. It buys more attempts at an unchanged success rate. If the rate is near zero, twenty attempts produce twenty losses — and the money would have been better spent on one good idea.

How many videos a month should you actually budget for next quarter?

Work backwards from testing rather than from a posting calendar, and set the number by what you can afford to test properly.

If paid is part of the plan, the base rates give you a range: about 20 creatives per winner on average, closer to 45 if you want reasonable confidence, and 9 to 25 to structure a single TikTok campaign — remembering that several creatives can come from one production.

For organic, the honest benchmark is what comparable brands do rather than what performs best. Socialinsider analysed 35 million Instagram posts from 447,613 pages across 2025: roughly 8 Reels, 7 images and 5 carousels a month, about 20 posts.

Their format data is more actionable than their cadence data. Measured as engagement per follower, carousels averaged 0.55% in 2025, Reels 0.52% and images 0.37%. Note that this divides by followers, while Dash Social's engagement rate divides by something its page does not disclose — so the two are not comparable and we are not going to pretend to know by how much. Within one dataset, though, format is a lever you can pull this week without producing anything more.

What does content volume cost to produce — 20 videos a month, or 45?

Our published rates, as a Los Angeles production company, so you can do the multiplication yourself.

Published rateTermsUGC video"On average, pricing starts at $350 per video"Minimum 5 per batch · 14–30 day production · usage rights typically 6 monthsSocial-media video packageSocial-media video packages start at $3,000 per shoot, published on our homepageScope varies; /video-marketing publishes no price, stating only that cost "varies based on the scope of work and deliverables"2D CGI"~$70/sec on average""usually takes around 1 production hour per second"Complex 3D CGI"~$300–1,000+/sec""may require 6–10+ hours per second"

Every figure is a floor or an approximation. None is a flat rate.

What 20 videos a month costs at our LA production floors

Run the numbers and the shape of the problem appears. Twenty UGC videos at the $350 floor is at least $7,000; forty-five is at least $15,750.

Two forces pull the real number away from those, in opposite directions, and we are going to name both because the first one flatters us and the second does not. Downward: one video does not equal one creative — hooks, cuts and aspect variants from a single production each count as a separate test asset, so you may need to commission fewer videos than the creative count implies. Upward: $350 is a floor, not a rate. It is what the cheapest qualifying video costs, and our page says "on average, pricing starts at $350 per video". Anything with more scope costs more, and there is a five-video batch minimum underneath it all.

So $7,000 is a lower bound, not an estimate, and we cannot narrow it without knowing your scope. Any agency — us included — that multiplies a floor by a quantity and presents the result as a budget is showing you the best case and calling it the price.

We do not publish a monthly retainer price or a standard monthly asset count. We should, and it is on our list. Until then we would rather show you per-unit rates — two floors and two averages — than quote a package we have not defined.

If the multiplication is pushing you toward hiring instead, that is a fair conclusion to reach and we have costed it out honestly in in-house social team vs agency — including the payroll loading most comparisons leave out.

Did volume actually work for a real brand? A 4.5x view result, with four caveats

For one brand, alongside several other changes — and we cannot tell you how many assets it took, because we have not published that figure.

Over 2025 we produced organic and paid content for a wellness brand, and published the engagement as a case study. Instagram views rose 350% — a 4.5x increase. Instagram shares rose 94%. On TikTok, likes rose 577.32% to 236,756 and shares 474.84% to 11,422. Within the first two weeks a reel hit 200,000 views.

Four caveats, because a case study without them is an advertisement.

The asset count is not published. This section asks whether volume worked and cannot fully answer it. That is a gap in our own reporting, not a rhetorical device.

It was not volume alone. The engagement included a 72-slide strategy deck and ran on a stated format thesis — "edutainment," entertain before educate — with carousels as the deliberate bet. We ran the audit before commissioning production, but that ordering is our account of the engagement: the published case study lists the deliverables without stating a sequence. Take it as our claim, not as documented fact. We have written that method up in full as a separate deep-dive on the format strategy, including where the independent benchmark data splits on it — Socialinsider's per-follower figures are consistent with the carousel bet, while Metricool recommends Reels for reaching new audiences instead and publishes a format split rather than a direction: Reels engagement rate up 24.76% against single-image posts down 45.98% on engagement.

A material share of it was paid. Our own pages put the ad contribution at 38.6%, 40% and 41% in three different places and we have not reconciled them, so we will not quote a single figure — call it roughly two fifths bought. This was never an organic-only result, and any volume argument resting on it carries that caveat, ours included.

Two things on our own pages do not reconcile, and we are not going to hide either. Instagram engagement growth is published as 58.31% on the case study and 58.41% on our homepage. This article does not rest on either figure — we have not used it above and we are not using it now — but you should know the pair exists before you quote one of them at us.

Separately, our case-study headline gives an absolute start-and-end view count whose ratio does not match the published growth rate — the two cannot both be right. We have therefore withheld that pair from this article entirely and used only the relative figure our two pages agree on.

Both are on our own fix list. An article about evidence standards does not get to quietly round its own numbers.

The honest answer: how many videos a month does a brand actually need?

There is no universal number, and anyone giving you one without naming a metric, a sample and a source is guessing.

What the evidence supports: on paid, plan for 20 creatives per winner as an average and 45 for confidence, with replacements ready before performance declines. On organic, consistency is better evidenced than frequency, and format choice is the cheapest lever available. Organic posting frequency itself remains genuinely unsettled: the two best-powered studies that relate frequency to an outcome reach opposite answers.

The question worth asking is not "how many videos." It is "how many attempts can we afford, and is our hit rate good enough to justify them." If you cannot estimate your own hit rate, that is the thing to fix first — and it costs nothing to work out.

Working out your number for next quarter?

Send us your paid spend and your current creative count and we will run the win-rate arithmetic with you — including the cases where the answer is that you are already producing enough. We produce in Los Angeles and work with DTC brands across the US.

Talk to us → · UGC rates · Video production · Paid creative · CGI