Hi friends!
Last week of September, and every Q4 plan that was a draft in August is a budget now.
This week, YouTube quietly repriced every view based rate card in the market, creators send four in ten clicks to the wrong page, and Meta’s biggest spenders are posting its worst returns.
Let’s dig in.
September 29th
📺️ YouTube changed what a view is, and nobody renegotiated
YouTube changed how it counts views in August. A view now registers the moment playback starts, where it previously needed 30 seconds of watch time. Only 60% of public views clear that old bar, so public counts run about 1.67 times higher than engaged views. Feed autoplay accounts for 77% of the gap.
Nothing was inflated. A word was redefined, and every rate card denominated in that word repriced overnight. A $30 CPM agreed on public views is a $50 CPM on engaged views. Royalties and Partner Programme eligibility still run on the engaged number, so the platform kept the strict metric for the money it pays out. Ask which count a quote uses before you sign anything this quarter.
🛍️ Creators send 40% of clicks to a product page. Marketers send 66%.
Bitly surveyed 1,006 consumers and 501 marketers and creators between July 27 and August 5. Consumers expect a link to land on a product page 61% of the time. Marketers deliver that on 66% of links. Creators manage 40%, routing 24% of their clicks into a multi link directory instead.
This will be written up as a trust gap. Ignore that, because trust is an attitude you will not move this quarter. The 40% is a routing decision you can fix this week, and the 79% who abandon after a generic homepage redirect prices what it costs. Your affiliate leak may be the destination URL, not the creative. Bitly sells link management, so weigh the framing.
🍪 Meta’s biggest spenders are getting its worst returns
Metricool looked at 44,355 advertisers and 628,969 campaigns across $471.5 million of spend, January to June in both 2025 and 2026. Meta sales campaign ROAS fell from $5.48 to $4.57. Spend in the sample more than doubled, from $62.9 million to $151.8 million. Purchase value grew too, just slower.
The tier split is the part worth keeping. Advertisers spending over $10,000 a month posted the worst return of any group, 3.97 this year against 4.88 last. US advertisers averaged $2.42, against roughly $12 in Norway. Scale is supposed to buy efficiency and here it buys the opposite, which is the arithmetic behind every Q4 conversation about moving budget into channels that price on outcome.
🎤 Seller Spotlight
A ten year old signed a master toy deal, and reach was not the reason
Bonkers Toys has signed The Rizzler (1.9M followers on TT) the ten year old Staten Island creator Christian Savasta, to a master toy deal. Mystery plush, figures and accessories reach US retailers in fall 2027, with Target and Walmart among the buyers and distribution across more than 70 countries. Bonkers built Ryan’s World into the top selling preschool toy brand for roughly five years, so it knows the route. Head of brands Dan Meyer on how the company picks: “Our consumer need is fan engagement. Do the fans want product from the creator they’re following?”
The signal is demand, not reach: follower count never entered the decision. Fan behaviour did.
The line carries rare chaser variants: the same blind box scarcity we covered with JapanLA on September 15th, now moving through Target.
@itztherizzler Iceman🧊Crashout Edition!!😂🤣
The risk sits in the calendar. Twelve months from signing to shelf means Bonkers is betting on a ten year old still mattering in autumn 2027. Every creator product deal carries that bet. Most brands price it as though the audience is permanent.
🧠 Brand Move of the Week
Sephora opened on TikTok Shop with a product it will not sell itself
Sephora’s TikTok Shop went live this month, opened by a Saturday livestream with Mikayla Nogueira, co founder of POV Beauty and 17 million TikTok followers. The hero product was a POV Beauty twelve piece holiday set, $265 for $327 of stated value, exclusive to TikTok Shop and absent from sephora.com. Monthly drops are planned. TikTok Shop is now the fourth largest health and beauty retailer in the US, past $4.4 billion in beauty since its 2023 launch.
The move is not that a prestige retailer joined TikTok Shop. It is what it chose to withhold from its own site. Sephora put a discounted bundle into a channel where 50% to 70% coupons are normal, and kept its own storefront clean. The channel is doing the discounting the brand does not want on its own domain.
Blaine Nielsen of Rithum named the unresolved part: “Is this really just a one time exploratory purchase, or can you turn this into a recurring customer of yours?” Acquisition through someone else’s livestream is rented. Nobody has published what the second purchase looks like.
🔭 Trend Watch
What are you actually buying?
Four numbers in this issue describe one problem. YouTube public views run 1.67 times engaged views. Creators route 40% of clicks to a product page against 66% for marketers, and 79% of people abandon after a generic redirect. Ryan Trevis, who runs the street interview agency Stradvert, estimates 40% to 50% of the passersby in his ads are paid actors, in a format sold entirely on being unscripted. And a new class of campaign now sells creator distribution priced per thousand views across accounts with no audience data and no control over what the clip sits beside.
None of this is fraud. It is definition drift. The creator channel was sold on the premise that its units are more honest than paid media’s: a view is a person watching, a street interview is a real opinion, a creator link is a recommendation. Each of those definitions has loosened while the price stayed attached to the old one.
Trevis is worth quoting against himself. “People in the street have got no reason to lie about the products. Whereas an actor or an influencer or a UGC creator, people watching, they know they’re being sold to.” He is describing exactly what makes the format valuable while telling you half the cast is hired.
The practical version is dull and worth doing anyway. Every rate you agreed this year is denominated in a unit, and at least one of those units changed meaning in 2026. Buy one campaign priced on engaged views and one priced on public views, then compare the invoices. You will know what your CPM actually bought. Nobody is running that test, which is exactly why it is cheap to run.
📰 Other news you should look at for this week
Promo codes beat every other holiday purchase driver. Sprout Social’s Q3 pulse, fielded by Glimpse across 2,286 users in the US, UK and Australia, puts promo codes first at 30% and targeted social ads last at 18%. 78% plan to use social for gifting at or above last year. 📎 Read more
YouTube is handing the algorithm to the viewer. From October, users can build their own feed tabs by describing what they want in a prompt box, powered by Gemini. Discovery stops being one algorithm to optimise for and becomes thousands of user written ones. 📎 Read more
That 97% creator churn figure is not a payments number. Trolley surveyed 450 independent earners. 97% would consider leaving a platform for any reason. The payments specific figure is 41%. Both are useful. Check which one a deck is quoting you. 📎 Read more
Ok, see you next week. In the meantime, Sell! Sell! Sell! 🫡