I lost $950M... then built a $200M brand with David Beckham

Why does it matter? IM8 doubled its ad spend and its CAC went down.

Danny Yeung watched $950M in market cap evaporate when COVID testing dried up — and then built a $200M supplement brand in 18 months with David Beckham as co-founder. The engine behind that run isn't what most DTC founders think it is.

• Doubling monthly ad spend can actually lower your CAC — if offline brand-building is doing invisible work your dashboard can't see
• The celebrity brand model fails more often than it succeeds; the co-founder model, with genuine accountability, is a different animal
• Meta now does the targeting for you — the only remaining edge is creative volume and diversity
• The founders who build lasting companies know when to be delusional and when to be brutally honest — and they're never stuck in one mode

Never delusional = employee. Always delusional = failed founder. The skill is knowing when to switch.

Shaan frames it mid-conversation, almost as an aside, and Danny doesn't push back — because it describes exactly how he operates. The art of founding isn't sustained belief or sustained skepticism. It's toggling between them at the right moment.

"Somebody who's never delusional, they end up being an employee. And somebody who's only delusional ends up being a failed founder. It's the person who's able to shift gears between the two and know when you need to be in each mode. That's sort of the art of the whole thing."

Danny's whole run is a case study in this. He went into supplements with zero category experience, his entire board against him, and a public market watching every move. That required pure delusion — the conviction to push forward when smart people, with legitimate points, were telling him not to. But before launch, he ran clinical trials. Spent a week in a factory testing 20 or 30 different flavor profiles until one landed. Recruited scientists from Mayo Clinic and NASA. Pure realism mode.

The failure mode isn't being too confident or too cautious. It's getting locked in one gear. The board members who called him crazy weren't wrong about the risks — Danny just knew that this was the moment to be delusional, and later would come the time for rigor. Before any big decision, the useful question isn't 'am I right?' It's: which mode am I running in right now — and does this moment actually call for that?

IM8 doubled ad spend from $17M to $33M in a quarter — and their CAC dropped.

Q1 CAC: $305. Q2: marketing spend doubled to $33M for the quarter. Conventional wisdom says scaling spend raises CAC. IM8's fell to $301. Then July: spending over $10M a month — CAC dropped to $239.

Danny's explanation isn't a paid media optimization story. It's 200-plus offline events. Scientists and formulators in rooms with real customers. Seven events in seven days in New York. None of it shows up clean in a dashboard. "I don't have quantifiable data to say this is all offline," he admits, "but I'm going to continue to double down."

The mechanism makes sense even without the attribution. Once someone has stood in a room with a Mayo Clinic doctor who formulated the product — and then sees your retargeting ad — they're not seeing an ad. They're seeing a reminder. The conversion cost on that impression is a completely different number than cold traffic. The in-person touchpoint does something to paid acquisition efficiency that no amount of audience optimization can replicate.

Most DTC brands are optimizing their landing pages while this lever sits untouched, because offline is harder, can't be A/B tested, and doesn't scale cleanly. That's exactly why it carries disproportionate value. If your CAC is stuck or climbing, ask when the last time was that a real human being tried your product in front of another real human being — and whether you've actually invested in that at any real scale.

AG1 took 10 years to hit $100M. IM8 did $200M in 18 months — because Beckham isn't their ambassador.

Grüns hit $100M in 22 months and got acquired for $1.2B after three years. IM8 hit $100M at 11 months and $200M annual run rate at 18. AG1 — arguably the defining supplement brand of the last decade — needed the full 10 years to reach $100M.

The difference Danny points to isn't distribution or ad budget. It's the structure of the Beckham deal. His pitch was deliberately unusual: "David, I don't want you just as ambassador. I want you as co-founder — but this brand has to live beyond you. I don't want this to be another celebrity brand."

Beckham had been approached for 30 years and turned everyone down. That framing — real skin in the game, not a check for a face on packaging — is what got the yes. And co-founder accountability changes everything downstream. The product actually has to work, because the downside for Beckham isn't losing a brand deal — it's his reputation. That alignment is real in a way that PR can't simulate.

The board told Danny exactly what the data says: more celebrity brand failures than successes, full stop. He agreed with the data and bet that the co-founder structure made the comparison invalid. The tell that he was right? Aryna Sabalenka's team emailed inbound after her nutrition coach quietly put her on the product. Three months later, measurable recovery improvements for the world number one in tennis. That email wasn't solicited. Every major ambassador after that was inbound too.

Ad targeting is a commodity now. Two to three thousand ads live at once — that's the actual edge.

2,000 to 3,000 ads live at any given time. 500 to 800 creators tested every week. A proprietary creative fatigue score that flags when an ad is about to die — before the numbers confirm it — so the team can rotate before ROAS degrades.

"Ad buying is like a commodity right now," Danny says. "Before, you had to be really good at it. But now Meta does all of that for you. You don't need to do the targeting. All you need to figure out is how do you get really good creatives and really good creative diversity — because now the creative does the targeting."

This flips the DTC playbook from two or three years ago, when audience segmentation and bid optimization were where smart teams competed. That battleground is closed. Meta's algorithm erases the gap between good and great targeting. What it can't do is generate creative on your behalf at volume and diversity. The teams still investing headcount in audience-side optimization are fighting the last war.

The new constraint is throughput: how many angles you can test, how fast you identify winners, how quickly you retire fatiguing ads before they drag performance. Volume and diversity. The 55-60% statics, 40-45% video split isn't arbitrary — it's what you get when you're pulling from a thousand-plus ambassador network generating content daily across personas. Building that pipeline is the work. The targeting takes care of itself.

IM8 ran clinical trials and got certified by two independent labs before a single ad ran.

Before launch. Not as a marketing add-on after the product was already selling — before a single dollar of paid media ran. Eurofins certification plus NSF for Sport, which verifies clean from 280 banned substances. A scientific advisory board: Dr. Don Mu Salem from Mayo Clinic, James Green former chief science officer of NASA, Suzanne Dekoda director of microbiome at Cedars-Sinai. All involved in building the formulation.

The logic was airtight: "Everyone knows David is not a scientist. I'm not a scientist." The attack vector on a celebrity supplement brand is obvious — so Danny sealed it before anyone could throw it.

Pre-launch science investment in a trust-deficit category isn't just integrity. It's a moat that neither celebrity partnerships nor ad spend can replicate, because it's the thing that converts the skeptics who check before they buy. The Sabalenka story is the proof: her nutrition coach recommended IM8 independently, her team tracked her on it for three months measuring recovery several times daily, and then emailed Danny unprompted. That's what happens when the product holds up under the scrutiny of people whose literal job is to optimize elite athletic performance.

In wellness, finance, or any category where the default consumer posture is skepticism, the brands that invest in third-party validation before launch aren't just doing the right thing — they're building inbound demand from the highest-credibility users, who then do the conversion work on everyone else.

When COVID ended, Danny cut 2,000 people fast — and that's why IM8 exists at all.

The competitors didn't move. Q Health: $3 billion market cap, a billion dollars in US revenue. Lucira: another $2 billion company. Both went bankrupt within 18 months of COVID ending, because they kept building — manufacturing facilities, at-home testing lines — at peak headcount after demand had already left.

Danny's framing on his own decision strips out the drama: "Why would I have 2,000 people if there's no testing? They're just going to stand around — which is not good for them. Everyone has to move on with their life."

The companies that went bankrupt had every resource advantage. What they lacked was willingness to ask the simple question: what's the right size for conditions right now? They built their institutional identity around being a $3 billion testing company and couldn't shed it fast enough when the tailwind disappeared. That's how you go from a billion in revenue to bankruptcy in 18 months — not from lack of resources, but from the wrong mental model at the inflection point.

The cut was brutal and the right call simultaneously. It preserved the capital and credibility that let Danny pivot to IM8. The lesson isn't just 'cut fast' — it's to build the habit of asking the right-size question into every operational review, especially when growth is still happening and the inflection isn't obvious yet.

The marketing director hung up on him. Danny called the director's boss — and got 30 restaurants in one meeting.

Hong Kong, early days of the group-buying company. Danny cold calls a big restaurant group, gets the director of marketing, gets hung up on — rudely. Most people close the file. Danny asks himself who this guy's boss is.

"He hung up on me. And then I like found who this guy's boss is."

He tracks down Sandeep — owner of 30-plus restaurants — calls the front desk, asks for him directly. Sandeep takes a 20-minute meeting. Tells Danny the model probably won't work in Hong Kong. And then says he'll partner anyway: "Because you called me after my director of marketing hung up on you, and we're still here. I'm going to give you all of my dirty restaurants."

That one yes filled 30 days of quota in a single conversation. And once Sandeep was in, other groups followed — because now Danny had a credible name to mention.

The principle compounds: gatekeeper rejection is a routing signal, not a verdict. The decision-maker above often rewards the persistence their employee couldn't appreciate, because they've built something themselves and they recognize what it looks like. Most sales processes treat a no as a closed door. Mapping the org chart above it multiplies conversion rates in a way that no amount of scripting or follow-up cadence can touch.

The unconventional moves are becoming the only moves that matter.

The thread across Danny's career is that the standard playbook — audience targeting, ambassador deals, post-launch science — is now the floor, not the ceiling. The teams compounding fastest are treating the counterintuitive as mandatory: offline events when everything is digital, co-founder deals when everyone else does endorsements, clinical validation before the first sale. What looks like extra work is actually the moat. The next DTC advantage won't come from what you optimize — it'll come from what you refuse to skip.

---

Topics: supplements, DTC, celebrity partnerships, Meta advertising, founder mindset, COVID business, entrepreneurship, CAC optimization, offline marketing, product-market fit