Congrats on the revenue. Sorry about your money.
with Adam, Pentane
Congrats on the revenue. Sorry about your money.
Show Notes
There is a scenario playing out inside thousands of e-commerce companies right now. Revenue is up. ROAS looks great. The ads team is celebrating. And the founder is quietly losing money every single month - because nobody ever stopped to ask how much should be spent before the first dollar went to ads.
Adam knows this scenario firsthand. He took Bottle Keeper - a single product with a single original patent - from zero to $60 million in sales across multiple channels. He filed 42 continuation patents off that one idea. He spent $500,000 on IP litigation in year three. And along the way, he made every expensive mistake an e-commerce operator can make. Now he has built Pentane: software that encapsulates everything he learned and tells other founders the number they actually need to know - whether they are profitable, and what to change next.
This episode is about the gap between gross revenue and real profit, why AI agents trained on operator expertise are different from generic chatbots, and why consumers are beginning to reject AI-generated content in favor of anything that feels genuinely human.
Why High ROAS Can Still Mean You're Losing Money
ROAS - return on ad spend - measures how much revenue a dollar of advertising generates. A 4x ROAS sounds like a win. But if your fixed costs are too high, your variable costs are miscalculated, and your advertising budget was set by feel rather than math, a 4x ROAS can still produce a net loss at the end of the month.
Adam's insight is deceptively simple: most founders set their advertising budget after the fact, as a percentage of what is left over. Pentane flips that. You set the foundation first - what you should be spending on fixed operations, what your variable costs look like per unit sold - and advertising becomes a lever you can actually control, not a guess dressed up as a strategy.
The Three-Bucket Model: Not All Expenses Are the Same
Adam separates every dollar a business spends into three categories. Fixed expenses are the costs of running the company regardless of sales - rent, salaries, subscriptions. Variable expenses scale with revenue - payment processing fees, shipping, packaging, returns. Advertising is pulled out of both categories entirely and treated as fuel: money you deliberately pour in to generate more sales.
The reason advertising lives in its own bucket is that it behaves differently than the others. You can dial it up or down in real time. You can measure its direct output. You cannot do that with your warehouse lease. When you lump advertising in with general operating expenses, you lose the ability to see what it is actually doing - and you lose the ability to make intelligent decisions about how much more fuel to add.
Charlie: What Happens When You Train an AI on Operator IP
Adam built an AI agent called Charlie - a web application connected to Pentane via API - that a founder can ask questions like "How do I hit 5% net profit this month?" and receive a specific, data-informed answer in under ten seconds. Charlie does not answer like a generic language model. It answers like Adam, because it was trained on Adam's transcript, his frameworks, his operator experience from $60 million in e-commerce sales.
He built Charlie using Open Claw - a tool that used to sit on a year-or-two roadmap. He built it in an hour. This is what the current AI environment makes possible: a domain expert can now build a digital version of their own knowledge base and deploy it as a product, not just a chatbot.
The underlying model is the same one you have access to. The difference is the context it sits inside. An agent trained on deep, proprietary operator experience produces fundamentally different output than a model prompted with generic instructions. That context gap is where the real product value lives.
42 Patents from One Idea: The CIP Strategy
Bottle Keeper started with one patent on a beverage insulation sleeve. From that original filing, Adam's team filed continuation patents - each one extending, modifying, or specifying a different aspect of the original invention. By year three, they had 42 patents. They also had a $500,000 legal bill and a strategy for dealing with Amazon infringers that most founders have never considered.
Amazon will not act on DMCA takedowns or vague IP claims from small brands. What Adam discovered is that Amazon does respond to federal court orders. The strategy: sue small infringers not for damages but for consent judgments - agreements to stop infringing that produce a court order number. Submit that number to Amazon. The listing comes down. It costs money upfront, but it creates a system that scales without requiring you to win every case on the merits.
The Human Content Correction
As AI-generated content floods digital channels, Adam sees consumers beginning to knee-jerk back toward anything that feels genuinely personal and human. The same dynamic that drove early social media adoption - proximity, authenticity, the feeling that you are watching a real person - is resurging as a correction to synthetic content saturation.
For founders building brands in e-commerce or content-driven categories, this is a real competitive signal. The question is not whether AI can write your copy or script your ads. It can. The question is whether it should be the only thing doing that - or whether the humans in your story are the differentiation.
Frameworks from This Episode
These frameworks have been added to the AI for Founders Frameworks Library. Filter by E-Commerce or Adam to find them.
- The Profit Foundation First Framework - Set how much you should be spending before touching advertising. ROAS is meaningless without knowing your true cost structure.
- The Three-Bucket Expense Model - Separate Fixed, Variable, and Advertising into distinct categories. Advertising is fuel, not overhead - manage it like a lever, not a line item.
- The Expert Agent System - Train an AI agent on your own IP, frameworks, and operator transcript to answer domain questions at scale. Context is the product.
Tools Mentioned
These tools have been added to the AI for Founders Tools Directory.
- Pentane - pentane.com - Profitability software for e-commerce operators. Organizes expenses into three buckets and tells founders what to change next to hit their profit targets.
Glossary
Terms from this episode have been added to the AI for Founders Glossary. Filter by Adam to see them all.
- ROAS - Return on ad spend. Revenue generated per dollar spent on advertising. High ROAS does not indicate profitability if fixed and variable costs are out of line.
- Contribution Margin - Revenue minus variable costs. What remains to cover fixed expenses and generate profit. The number that tells you whether each unit sold is helping or hurting.
- Fixed Expenses - Costs that remain constant regardless of sales volume - rent, salaries, software subscriptions. Must be set and understood before advertising spend is determined.
- Variable Expenses - Costs that scale with revenue - shipping, payment processing, returns, packaging. Often underestimated by early-stage operators.
- Advertising Expense Bucket - A distinct expense category for paid media spend, separated from fixed and variable costs. Treated as fuel - a deliberate, adjustable lever rather than a line item to minimize.
- CIP (Continuation in Part) - A patent filing strategy that extends a parent patent by adding new claims or embodiments. Adam used CIP filings to grow from one original patent to 42, covering different product variations and use cases.
- Consent Judgment - A court-ordered agreement in which an infringer agrees to stop infringing without a full trial. Produces a court order number that Amazon honors for listing removal - a scalable IP enforcement strategy for small brands.
Q&A: What Founders Ask After This Episode
How do I know if my advertising budget is right?
Set your profit target first. Work backward through your variable costs and fixed costs to determine your contribution margin per unit. Only then calculate how much you can afford to spend acquiring customers and still hit your target. If your ad spend was set before you did that math, it was set by feel - not by a foundation.
What is Pentane actually doing differently from a spreadsheet?
A spreadsheet requires you to know the right questions. Pentane was built to encode the operator knowledge behind the questions - the same logic Adam developed across $60M in e-commerce. The software tells you what to look at, what to change, and in what order. The AI agent Charlie answers natural language questions with specific, data-informed guidance drawn from that operator experience.
Is the patent strategy Adam used only available to product companies?
CIP filings and consent judgment strategies apply to any IP-defensible product - physical or digital. The key insight is that federal court orders are what Amazon and other platforms respond to, not internal DMCA claims. For any brand operating on marketplace channels with protectable IP, this approach is worth understanding before infringers appear.
What does Adam mean when he says consumers are returning to human content?
As AI-generated text, images, and video flood every channel, content that signals genuine human presence - a real face, a personal story, unscripted delivery - stands out by contrast. Audiences are beginning to treat AI content as noise and human content as signal. For founders building content-driven brands, this is an opportunity: your authentic story is a competitive asset that AI cannot replicate.
How did Adam build Charlie in an hour using Open Claw?
Open Claw allowed him to connect his operator IP - his transcript, frameworks, and Pentane's API - to a conversational interface without a full engineering buildout. The speed is real, but the value is in what he brought to the tool: decades of operator experience encoded as context. The tool is a wrapper; the intelligence comes from the domain expertise inside it.