Become a member to get exclusive access to:
Two Venice operators on building an in-house content engine: Kasper Brandi Petersen (LABFRESH) on making 2,000+ videos a year with no agency or influencer fees, and Ash Read on where SEO is heading as buyers start asking AI to shortlist brands.
Kasper Brandi Petersen, Co-founder of LABFRESH, & Ash Read, content operator (ex-Buffer, ex-Wayflyer) · July 8, 2026 · Venice All Hands
Two operators, two halves of the same problem: how a consumer brand produces a lot of content without an agency or a big budget. Kasper Brandi Petersen walked through the LABFRESH engine - more than 2,000 videos a year, made entirely in-house by young creators hired off TikTok, paid a low base with large upside, and organized into "content houses" a few times a year. Ash Read covered the discovery side: where SEO is heading as buyers start asking LLMs to shortlist brands, and the systems that let a small team produce differentiated content fast. The through-line: build an internal habit, not a purchased service, and let people who live on social do the work their way.
Five takeaways for builders:
Two years ago LABFRESH ran the standard playbook - chasing creators' inboxes, shipping clothes, writing elaborate briefs, negotiating constantly - and Kasper felt he was micromanaging things he couldn't control. Today the brand makes more than 2,000 videos in 12 months, all in-house. They pay almost no influencers. The work runs on employee-generated content: last year 28 employees appeared in videos, and showing up on camera is now implied whether you manage the store or the production line.
— Kasper Brandi Petersen, 00:11:03
Kasper's target was to double output to at least 100 videos a month - but with a caveat he stressed repeatedly: volume is massively overrated. Once you have a base of work, you can usually tell before launch whether a video will scale; he reckons he can forecast this 9 out of 10 times. Making a hundred versions of the same hook is wasted effort, because changing only the hook or voiceover doesn't register as a new video to Meta anyway. The old habit of a thousand fresh creatives and static ads with swapped taglines is, in his experience, no longer the best way to operate.
To staff this, Kasper replaced agency-trained marketers with young people who live on social media. Rather than post jobs on LinkedIn or collect CVs, he and his co-founder made TikTok videos - set to The Office theme song, walking around the offices, showing the egg cookers and the nap spots - and invited viewers to apply as paid creators. Within six months he'd hired ten people that way, for both retail and content. The logic is cost as much as fit: the best and cheapest creators are on TikTok, and a video with a good thumb-stop ratio reaches the right applicants for less than a job post.
The content team - five creators, now six - works three-to-four-day weeks, because Kasper doesn't believe anyone can be creative five days a week and still make their own content on the side. In interviews he asks to see candidates' screen time; his last three hires each had at least 10 hours a day. Pay is built around a deliberately low base, with the real upside coming in other forms - trips and equipment rather than cash: a camera, custom grills, two weeks in Switzerland, content houses in Tanzania and Mexico. Keeping that budget free is intentional, so there's room to reward the occasional low-cost video that massively outperforms.
Young and chaotic creatives still run on tight process. Kasper manages the team closely on Asana, where deadlines are hard - if something is due Monday, 99% of the time it ships Monday, even if that means working the night or the weekend. He doesn't track hours, only delivery. Work runs in two-week content sprints, because creative people can't hold a longer deadline without leaving everything too late. The strongest, most organized creator now manages the rest, with a separate lead on branding and a senior "Head of New Things" brought in for project management.
LABFRESH's content splits into storytelling (top-of-funnel, organic-first in feel), UGC/ads, and statics. But Kasper is blunt that reach is pay-to-play - their videos get almost nothing organically now - so he doesn't want the team shooting anything purely for organic. New creators always want to, and always get disappointed. His resolution is to collapse the distinction: whatever you shoot for top-of-funnel paid advertising, with good storytelling, is the same content that would have carried organically. TOFU videos with strong soft metrics - thumb-stop ratio, view-through - get a forced minimum daily budget even when they don't convert immediately.
The organising idea behind the shoots is the "content house": four or five times a year, the team gets an Airbnb - Düsseldorf, Mexico, Tanzania - plus a budget, and goes out to source people to shoot with. It didn't work when marketing staff tried to run it, because they aren't part of that world; a modelling agency ignores Kasper's emails or quotes high, but his team DMs the model directly on Instagram and does a faster, cheaper deal outside the agency. Creators who appear in ads are typically whitelisted and paid around 10% of the spend on their content, with small flat fees. Twice a year the brand runs bigger seasonal campaigns with paid models on flat fees.
For talent, the incentive structure depends on scale. Regular creators get a percentage of ad spend - Kasper can see exactly how much went into ads carrying their name. A commission on sales only makes sense, he argued, when at least half of customers recognize the person's name or face, because only then can you scale the advertising behind them; otherwise you're depending on them to post, which doesn't move enough product. Celebrity deals are the exception: a Danish footballer was paid roughly an 8,000-euro modeling fee plus a percentage of sales on a co-designed capsule - potentially 30–40,000 euros in total - with a German footballer on similar terms the following month.
Q: Do the creators edit the videos too, or is that done elsewhere? (Gaia)
They do all the editing themselves. Kasper is considering an offshore editor in the Philippines, as some other Venice members do, but earlier attempts in the Philippines and India felt too slow. To keep output fast he constrains tools deliberately: some videos must be cut in CapCut or on a phone rather than Final Cut, and the fancy camera isn't always allowed.
Q: Is it all ads, or organic too - and only Instagram? (Jennifer Bailey)
For menswear they run Instagram and Facebook; for the women's-wear brand it's Instagram only. TikTok was shut down because it was too distracting to upload daily, but they plan to relaunch it. On reach: Kasper has seen the same drop everyone has - LABFRESH gets almost nothing organic now, though their newer women's-wear brand pulled around 2.5 million organic views in its first three months. He suspects Meta throttles brands it knows will pay to play - while adding he has no basis for the accusation.
Q: Do you use gifting as a track - seeding, then affiliates, then creators? (Rasmus Weber, SkyPAD)
No. They ran it about three years ago, sent product to roughly 150 people in the first three months, and used zero of the resulting videos at scale - wrong sizes, bad styling, poor lighting. So they cut it and pay creators instead. Kasper framed it as fashion-specific: sizing and style problems make seeding unreliable for apparel, even where Rasmus finds it highly effective (seeding 100–200 per month) in a less competitive category.
Ash's thesis on AI content: it raises the floor - anyone can produce average content fast - but humans are still needed to raise the ceiling. Just because you can one-shot a thousand landing pages doesn't mean you should. His cautionary example was a roughly 10,000-page fashion site that peaked at about 180,000 page views a month (he hedged it might have been sessions); about three months after the peak, Google hit it, and while it still did around 50,000 pages a month, revenue fell to roughly 10% of peak. AI-driven volume tends toward short-term wins that can harm a brand long-term.
The gap between SEO and AEO comes down to how people search. Where you used to rank for "best trail running shoes" or "waterproof hiking boots," LLM queries now run 10–15 words and arrive threaded - someone asks for a trail shoe, then what's best for a specific use, then waterproof, then a price - and the answers come back far more personalized than Google's. What LLMs reward is information gain: net-new material that isn't already all over the internet. Proprietary research, customer data, real technical specifics. A supplement brand should add the trade-offs, the blood-work results, why it dosed a formula the way it did; a cookware brand, why it chose its materials - all net-new, all raising the chance of being cited.
Ash's most portable tactic: every brand should keep a reference bank - one store of everything about the brand for LLMs to pull from. Podcast transcripts, documents on how and why the product is made, specifications, proprietary research, case studies, tone-of-voice docs. He hosts it on GitHub and keeps adding to it. Then whether you're writing a blog post, an email, or social copy, you draw from one place, so the output actually knows the brand. On top of the bank he builds pipelines - his SEO pipeline runs about 12 steps, with a human giving feedback at each - and closes with a learning loop in Claude that studies the human edits and retrains the workflow so it improves each time.
Off-page work is shifting. Backlinks still matter for SEO, but for LLMs, mentions matter just as much - where the brand appears and the words around it are what the model scans. That raises the value of digital and traditional PR, industry publications, reviews, forums, and communities. Reddit in particular is heavily cited by LLMs, though it has dipped a bit lately. The goal is to get the brand named alongside the keywords you want to own.
Ash's first move is Google Search Console: filter for queries over 10 words (a regex - he has Claude write it) to see the long-tail, AI-style questions already sending people to your site, then build content around them, whether blog posts or extra FAQs on product pages. He's skeptical of AEO tools like Profound - most of that data, he said, is "kind of BS," because there's no reliable way to know what people type into LLMs; run the same query 100 times and you get different answers. LLM traffic is still around 1% of overall traffic for most people, but it's rising as Google puts AI overviews in more results - and it tends to be low-volume, high-intent. On Shopify, agentic commerce works fairly well out of the box.
Kasper Brandi Petersen is co-founder of LABFRESH, the Amsterdam-based menswear brand known for stain-, odour-, and wrinkle-resistant premium cotton shirts, which he built with co-founder Lotte Vink after it became one of the most-backed European apparel projects in crowdfunding history. He is Venice member number one. Drop him a message in Slack.
Ash Read is a content operator who previously led content at Buffer and Wayflyer, and built a homewares content site to 400k+ monthly pageviews. He now builds AI-assisted content pipelines focused on SEO and AEO. Drop him a message in Slack.
Continue reading with these related insights and strategies.
Tom Gozney built Gozney from a £5,000 loan and a homemade garden oven into a ~$100M live-fire cooking brand. In this Venice AMA he breaks down brand-building over revenue-chasing, scarcity and drop culture and distribution discipline.
True Classic President Adam Eisenstadt on taking the brand from a $40M COVID launch to ~$280M - the international playbook: go broad to learn, run a country-by-country P&L, and keep operations simple.
George Sanderson (Honest Dog Co.) and Ollie Groombridge (8hours) walk through the dashboards and agents already running in their businesses — built in hours, costing hundreds, not thousands.
Start exploring the Venice community and discover new opportunities to grow your business.
Browse Members