A brand’s Meta dashboard reported a customer acquisition cost of $27. The actual number was $274. Adam Wellington, founder and CEO of Neon Flux and modern.ai, says the gap was not a platform anomaly, but a symptom of how many DTC brands now operate: with fragmented data, broken attribution, and dashboards they cannot fully trust.

The problem often starts with basic tracking infrastructure. An employee sets up a pixel incorrectly. A media buyer inherits a corrupted baseline. A lead is registered as a sale. Third-party attribution tools then reproduce the same mistake because the underlying event data is already wrong.

“That situation is more common than it is not,” Adam says. “I don’t blame Facebook.”

Data loss compounds the issue. Adam notes that AppLovin deletes ad performance data after 60 days, while Meta and Google retain it for roughly 37 months. When a brand switches agencies or changes its CRM, historical performance data can disappear. “In a world of AI, you need it,” Adam says. “If you’re in business for four years and want all the history of your performance, you just lose it.”

Adam has spent 20 years in performance digital advertising, beginning with campaigns on MySpace at 16. He built UpSurge into an Inc. 500 agency that generated more than $150 million in revenue for clients, including Netflix and Experian, then helped build Neon Flux, a Miami-based brand accelerator managing consumer products across TikTok, Meta, Amazon, and Shopify.

In March 2025, he took the company’s internal infrastructure to market through Modern Agency and modern.ai, a unified commerce dashboard logging approximately $2.5 billion in commerce volume across more than 160 brands. The platform runs server-side attribution, UTM matching, and pixel tracking in parallel, giving operators three reference points rather than one, and preserves historical performance data on paid tiers.

“I’ve been solving our operational problems,” Adam says. “I had this realization that if we built a product from this, it could be a much bigger opportunity.”

His central argument is direct: brands are attributing revenue to channels that did not earn it, missing downstream sales from platforms such as TikTok, and making growth decisions on incomplete numbers. TikTok has become the clearest test case for that dysfunction, not because the platform is uniquely opaque, but because brands keep arriving at it with the wrong frame.

TikTok Is Not Facebook

For 20 years, Adam treated Meta as the dominant performance channel for DTC. Last Black Friday, for the first time, his team moved budget from Meta to TikTok. The return on ad spend was better.

“That’s the first time in 20 years I said that,” he says.

Adam frames TikTok today not as a social platform with advertising tacked on, but as the leading top-of-funnel commerce channel in the United States, with a massive audience and extremely high session time. TikTok Shop has grown into what he describes as one of the world’s largest emerging commerce marketplaces. “That’s basically primetime TV,” Adam says. “That’s where the eyeballs are.”

The conversion rarely happens on TikTok itself. A consumer sees a product there, then buys on Amazon or searches on Google. That downstream halo is what most TikTok reporting misses entirely, and what the integrated data in modern.ai is built to capture. IGK, a hair care brand in the Modern Agency portfolio, used that cross-channel visibility to scale to a 52% ROAS increase and 151,000 new customers in under six months, Adam says.

Creative misalignment compounds the gap. Polished, brand-controlled content that works on Meta fails on TikTok. “What works on TikTok is raw authenticity,” Adam says. “Brands that try to fight that don’t succeed as much.” TikTok-native content translates reliably to Meta, he adds. The reverse does not hold.

The 90-Day Creator Program That Actually Converts

Most brands expect TikTok Shop to move product quickly after launch. Adam puts the realistic window at 90 days, and says the decisions made in that period determine whether the channel ever performs.

“If you’re a new brand, you can’t just show up, offer creators 10% commissions, and all of a sudden you’re successful,” he says.

The approach he advocates starts with direct deals: 20 to 40 creators, upfront cash payments, generous organic commissions, and competitive rates on paid content. The goal in the first 90 days is to build enough shop traction that TikTok’s algorithm surfaces the program to a broader affiliate pool. Without that initial activation, open affiliate plans produce little.

On the strongest programs, top affiliates generate $50,000 to $60,000 per month in a given brand’s TikTok Shop. Adam describes the model as solving two problems simultaneously. “You get content, and you get people on the street talking about your product,” he says. “You don’t need to build an MLM pyramid scheme for it.”

Transparency with creators is the variable most brands underweight. Sharing what is converting, what competitors are paying, and what the brand needs from content is, in his view, the difference between a program that produces impressions and one that produces sales. “Making it as easy as possible for them to show up and do the rest of the job,” Adam says, “is what makes the most successful programs.”

The Channel Mix Nobody Is Running Correctly

Adam runs a roughly 50/50 budget split between TikTok and Meta for brands entering the market. Meta converts faster on a small budget. TikTok requires a ramp-up period but compounds over time, generating content and awareness that drives purchases across every other channel in the stack.

“Meta and TikTok are where you go to get your top-of-funnel sales,” he says. “Both are still in the mix.”

Amazon, in his assessment, is non-negotiable for brands that can list there. Avoiding it on margin grounds is not protecting margin; it is losing sales to customers who intended to buy. Modern.ai integrates Amazon Vendor Central and Seller Central alongside Shopify and ad platforms, producing a blended customer acquisition cost with the Amazon halo included.

Live commerce occupies a different category. Adam has been tracking TikTok’s live shopping feature for nearly two years and believes it will eventually become a major factor. The current data does not support prioritizing it for most U.S. brands. “Shorts are still where all the money is,” he says. “Brand-produced lives haven’t yielded outsized returns compared to just investing more in affiliates and shorts.” 

One Dashboard Instead of 12

The data problem Adam spent years solving internally was structural. Finance teams used QuickBooks or NetSuite. Marketing teams used Triple Whale or Northbeam. Operations ran on project management software. No single view integrated all of them with the business context that AI needs to produce useful outputs: cost of goods, seasonal goals, and shipping lead times.

“As a business owner who wants to see everything in one dashboard,” Adam says, “it just didn’t exist.”

Modern.ai was built on a structured data layer assembled before large language models were commercially available. That architecture is what Adam says makes the AI layer genuinely useful: the platform has context, not just metrics. The core dashboard is free. Paid tiers unlock five AI agents covering acquisition, retention, customer service, inventory, and lifecycle. A proactive recommendation feature called “Radar,” in development, will surface alerts about affiliates due for a bonus and high-value creators who have gone dormant.

Adam is also adding banking, lending, and debit card functionality to the platform, so brand operators can transact inside the same environment where they analyze performance.

What Operators Are Still Getting Wrong

Adam draws a distinction between brands that are adapting to TikTok’s commerce model and those still importing assumptions from older channels.

The most underappreciated gap is not in content format or targeting, he says, but in AI adoption itself. “I’ll still get a call from someone. They ask me a question, and I’ll say, ‘Did you ask Claude? It might have the answer faster than I picked up the phone.’” The larger error is evaluating platforms based on tests conducted months earlier. “Things are changing daily,” Adam says. “You have to be really open-minded for the next three years.”

AI-generated content moderation remains an unresolved problem on TikTok, in Adam’s view. Regulatory pressure on synthetic content will increase. The brands with clean data and full-channel visibility will be better positioned to adapt than those depending on fragmented dashboards. He frames the current period against 20 years of watching digital marketing cycles turn.

“It feels like the birth of the internet all over again,” Adam says. “It’s all so new. It’s all moving so fast.”

Subscribe to Our Newsletter

Check Out Our Podcast