Nutra operators and their playbooks
Guilherme Bifi: how a Brazilian affiliate built a US nutra operation on CPA deals, formats and the back end
Guilherme Bifi went from losing money on his own offers to running a US supplement operation with partners and a large affiliate base. This is what he has said publicly about how it works: affiliate deals, offer testing, creatives, VSLs and the back end, and how to apply it honestly.
On this page
- Who is Guilherme Bifi?
- Why Bifi moved from revenue share to CPA
- How he validates an offer before paying for traffic
- Format over copy: Bifi's creative system
- Leads, vertical VSLs and the buy button
- How he runs split tests: 70/30, every day
- The back end: where Bifi says the profit is
- How the operation is organized
- The mistakes he talks about
- Compliance, as he describes it
- How to apply Bifi's ideas to your own offer
- Frequently asked questions
Guilherme Bifi is a Brazilian direct-response marketer who sells supplements in the US market. He became an affiliate in 2022, got his own payout switched from revenue share to a fixed CPA, and in 2024 became an offer owner with partners. In interviews he credits three habits: testing formats before copy, validating offers cheaply before buying traffic, and a back end that keeps selling after the first order.
Everything below comes from what Bifi has said in public, mainly a three-hour podcast interview published in November 2024, plus his own channels and other podcast appearances. Revenue figures he has mentioned are left out on purpose: they are his own unaudited numbers, and they tell you nothing about what your offer will do.
Who is Guilherme Bifi?
His YouTube channel describes him as co-founder and CEO of H&W, working in direct-response sales for the US market (his channel). A podcast in December 2023 introduced him as a copywriter with more than ten years in the market who runs an operation focused on the US and earns as an affiliate while traveling (DMCast #14).
The path he described in the November 2024 interview:
The path he described
What happened, by his account, in the November 2024 interview
- 2012–2013Starts in digital marketingAs an intern; graduates in advertising.
- 2014Works abroad in New ZealandAnd improves his English.
- Mid-2010sOnline sales for a health companyRuns its online sales and moves its field sales to WhatsApp.
- 2020–2021Loses his salaried job in the pandemicA dropshipping attempt fails.
- 2021Enters direct responseThrough a course.
- Early 2022His own first offers lose money
- April 2022Told to become an affiliateA more experienced operator reviews his copy and tells him to become an affiliate.
- 2022–2023Affiliate on ClickBankFirst on revenue share, then on a CPA deal; works remotely from Southeast Asia through a lean period.
- 2023Starts mentoring other affiliates
- February 2024Decides to become an offer ownerAt Affiliate World Dubai, decides to add his own product and become an offer owner with partners.
His second appearance on a Brazilian podcast in July 2025 was about mindset and discipline rather than tactics (Atomicast). The tactics are in the long interview, and they are what this page covers.
Why Bifi moved from revenue share to CPA
On revenue share he received a percentage of each sale, front end and back end, and shared in the refunds and chargebacks. By his account, the offer owner refused to raise his percentage, so he escalated to the network and got a fixed CPA instead: a set payout per new customer.
His argument for CPA starts with the affiliate: the payout no longer shrinks when buyers refund, so the affiliate can buy traffic against a known number. He frames the deal as the affiliate selling a lead to the owner rather than selling a bottle to a consumer, and says owners value that lead highly. What lets an owner pay a high fixed price for it is the back end, which he describes from the owner's side further down.
The math he used makes the point that the size of the payout matters more than the niche. A worked version: at a return on ad spend of 2, a $60 product leaves $30 per sale after ads, while a $180 CPA payout leaves $90. Same efficiency, three times the money per sale, so the CPA offer can afford more expensive traffic and scales further. Here is that worked version side by side:
Same ROAS, three times the money per sale
A worked example at a return on ad spend of 2: what is left per sale after ads
Two cautions if you are on the other side of the deal. Offer CPA only when you can predict refunds and the back end; otherwise you are guaranteeing a payout against revenue you can't see. And watch traffic quality: a CPA deal makes refunds the owner's problem, so track refunds by affiliate and price the deal accordingly.
How he validates an offer before paying for traffic
Bifi tests new offers on email lists before running paid traffic. He rents access to large lists whose readers share the target audience's problems, sends about five or six emails for an offer, and reads the earnings per click. His bar, by his account, is an EPC above $3, which he associates with a VSL conversion rate above roughly 1%.
The logic is cost. An email send costs less than the paid-traffic budget needed to learn the same thing, and a cold list behaves enough like cold paid traffic to kill weak offers early.
A worked example with illustrative numbers: a send produces 2,000 clicks and $6,400 in sales. EPC is $6,400 / 2,000 = $3.20, above his bar, so the offer goes to paid traffic. A second offer gets 2,000 clicks and $3,600, an EPC of $1.80; it goes back for a new lead or a new offer before anyone spends on ads. In RevvyScales the VSL analytics report EPC (earnings per CTA click) per VSL and per variant, so the same comparison works on any traffic source (analytics).
Format over copy: Bifi's creative system
His central claim about creatives is that copy has become a commodity and attention is the scarce asset. The first seconds decide whether anyone hears the message, and the format (the edit, the concept, the packaging) decides the first seconds. So a proven message in a new format usually beats a new message in a tired format.
He sums it up as a formula: a winning creative equals a validated creative, multiplied by a new format, plus hook testing. In practice he keeps the validated structure (a curiosity promise, the pain, the mechanism, the proof, the call to action) and changes the package around it. The formats he named include reaction videos, podcast-style clips, motion edits, kinetic typography with a different font for each word, and gossip-style selfie videos. He also treats organic virality as a cheap pre-test: if a format travels on its own, it is worth paying to distribute.
Three ways to iterate on a winner
He described three directions for building new creatives from a winning one:
Lateral, horizontal, vertical
The three directions he described for building new creatives from a winning one
Lateral
Keep the ad and test new hooks and angles.
Most of the testing happens here, on the opening rather than the body.
Horizontal
Keep the message and change the format, for example a reaction video becomes a podcast-style clip.
Riskier, but it can reach parts of the auction where every competitor runs the same look.
Vertical
Study what makes a trusted expert credible in the niche (tone, bluntness, a contrarian streak) and build those traits into your creative.
He also warned about the shortcut many juniors take, copying a competitor's winning ad with a new hook: networks protect creatives, and plagiarism can move the commission to the original owner. Iterate on your own winners. The VSL hook examples and the hook rate guide cover how to write and measure openings.
Leads, vertical VSLs and the buy button
Bifi expands winning creatives into new VSL leads. He described running about a dozen leads per offer, each a different opening for the same VSL body, so a creative that wins in the feed gets a lead that continues its story.
He said vertical VSLs outperformed horizontal ones in his tests, which fits traffic that is mostly on phones. On the Growth plan and up, RevvyScales lets you attach a 9:16 companion cut to a VSL so phones and tablets get the vertical version and desktops the landscape one, with one URL and one set of analytics (portrait video).
Two small mechanics he mentioned: a mini-hook shortly before the call to action (telling the viewer what is coming in a moment, so they stay for it), and a gamified order step, where a spin-to-win wheel replaced the plain buy button and offered free bottles or discounts. He reported a large drop in the cost per checkout started.
The wheel has an honest version and a dishonest one. If every spin wins the same discount, present it as a guaranteed reward, not a game of chance; a rigged game misleads buyers about what they won, and the FTC treats design tricks like that as dark patterns (FTC report). If the outcomes really vary, sweepstakes and promotion laws may apply. Get advice before you ship it.
How he runs split tests: 70/30, every day
Bifi described daily funnel tests with 70% of traffic on the control and 30% on a single changed variable. He doesn't start at 50/50. When a variant wins, it moves to 50/50, then takes the larger share, and finally becomes the new control at 100%. In steps:
How a winning variant takes over
The daily funnel tests Bifi described
- 70 / 3070% of traffic on the control, 30% on a single changed variable.
- 50 / 50When the variant wins, it moves to an even split.
- The larger shareThen it takes the larger share.
- 100%Finally it becomes the new control.
The logic is risk. A 70/30 split protects most of the day's revenue while the variant proves itself, and changing one variable at a time means you know what caused the result. A worked example: with 10,000 visitors a day, the variant sees 3,000. If the control converts at 1.5% and the variant at 1.8%, the variant produces 54 orders a day where the control would have produced 45 from the same traffic. That difference shows up within a few days, while 7,000 visitors a day stay on the proven page.
He also tailors upsells to the first purchase, so a buyer who took a large bundle sees a different upsell from one who bought a single bottle, and he ties seasonal promotions to offer variants. The general method is in the VSL A/B testing guide. In RevvyScales, folders split traffic between whole VSLs by weight on every plan, and tests inside one VSL (headlines, posters, offers) are on the Growth plan and up (folders and A/B tests).
The back end: where Bifi says the profit is
He described three levers after the front-end sale:
The back end, lever by lever
The three levers he described
More units up front
Multi-bottle buyers are the most valuable customers, so he scores leads and gives high-volume buyers the fastest service.
Operations
A welcome call within about five hours of the order, follow-up calls and bundle offers, run with call-center partners. He also described a subscription that starts with a free first month and then ships product on a schedule.
Day-one offers
Upsells on the day of purchase: more supplements, information products, a training program and an AI follow-up chatbot.
With that back end, he said, his company accepts a loss on the front end: sales from his biggest affiliates leave it about $20 down per sale, and the back end recovers that within about six months. That is a real model, and also the riskiest one in this business: it needs cash to survive the gap, honest reorder data and refund rates that stay low. The nutra marketing guide works through the math.
Subscriptions need extra care. A free first month that turns into recurring charges falls under the FTC's negative-option rules: clear terms before billing, express consent and a simple way to cancel (ROSCA). The upsell funnel guide covers the honest way to sell after the first order.
How the operation is organized
By his account the company splits into production and management. Production runs in squads: a senior copywriter leads junior copywriters, editors, designers and page developers, with a project manager per squad. Management covers finance and testing, and new offers are validated by email before paid traffic.
He brings in outside copy consultants for structure and persuasion, and runs live training for his affiliates on copy and traffic. He also pays affiliates enough that they stay affiliates rather than becoming rival offer owners, which he sees as part of protecting the business.
The mistakes he talks about
He was open about two expensive lessons. His own first offers lost money for months before a mentor pushed him toward affiliate work, where he could learn on proven offers. And later, the company put a large sum into a display network for direct response and, by his account, lost it within about a week. He called it a decision driven by ego instead of evidence. The lesson is the one every media buyer relearns: test a new channel small, with a stop rule, before you commit real budget.
Compliance, as he describes it
Bifi said networks require compliant wording and that adjusting his language, for example avoiding direct naming of a disease in ads, let him scale further than his earlier, stronger claims did.
One caution on that example. Under FDA rules, implied disease claims count as disease claims (21 CFR 101.93), so a euphemism that still promises to treat the disease doesn't make the claim compliant. The durable version is to make claims you can substantiate, in structure/function language, with the required disclaimer.
How to apply Bifi's ideas to your own offer
Tick these off for your own offer; the ticks stay in this browser:
Applying Bifi's ideas
0 of 8 done
Frequently asked questions
Who is Guilherme Bifi?
Guilherme Bifi is a Brazilian direct-response marketer who sells supplements in the US market. He worked in advertising and online sales, entered direct response in 2021, became an affiliate in 2022 and an offer owner with partners in 2024. His YouTube channel describes him as co-founder and CEO of H&W. He also mentors affiliates.
Why does Guilherme Bifi prefer CPA over revenue share?
On revenue share the affiliate shares refunds and chargebacks, so the payout shrinks after the sale. A CPA deal pays a fixed amount per new customer, so the affiliate can buy traffic against a known number. He frames it as the affiliate selling a lead to the owner. He argues the payout size matters more for scaling than the niche.
What does format over copy mean?
It is Bifi's view that copy has become a commodity and attention is scarce. The format of an ad, such as a reaction video, a podcast-style clip or a motion edit, decides whether people watch the first seconds. So he keeps a proven message and changes the packaging and the hook, rather than writing new copy for an old format.
How does Guilherme Bifi test a new offer?
He validates offers on rented email lists before paid traffic. He sends about five or six emails per offer and looks for an earnings per click above $3, which he links to a VSL conversion rate above about 1%. Offers that pass go to paid traffic; offers that fail get a new lead or are dropped before ad money is spent.
What is a 70/30 split test?
A test where 70% of traffic sees the current control and 30% sees a variant with one change. It protects most of the revenue while the variant proves itself. Bifi promotes a winner in steps, from 30% to 50% and then to the majority, until it becomes the new control. Testing one variable at a time shows what caused the change.
Sources
- Guilherme Bifi — YouTube channel
- O Método para Escalar como Afiliado nos EUA em 2024 | Guilherme Bifi — DMCast #14 (Dec 2023)
- Guilherme Bifi — Atomicast S02E01 (Jul 2025)
- Bringing Dark Patterns to Light — FTC staff report
- Restore Online Shoppers' Confidence Act — FTC
- 21 CFR 101.93: Certain types of statements for dietary supplements — eCFR