The $40,000 Wake-Up Call
It was 3 AM on a Tuesday in March 2024, and I was staring at a spreadsheet that made no sense.
Our Meta Ads agency — the "premium" one we were paying $40,000 a month — had just sent their monthly report. According to them, we were crushing it. ROAS was up. Impressions were up. CTR was up. Everything was green arrows and congratulatory language.
But when I pulled the actual numbers from Shopify, our revenue was flat. Our return rate had spiked. And our customer acquisition cost had quietly crept up 25% over the past quarter. The agency's beautiful report and our bank account were telling two completely different stories.
I fired them the next morning. Then I fired our email agency. Then our analytics consultant. Not out of anger — out of exhaustion. I was tired of paying six figures a month to people who didn't understand my business as well as I did, and whose incentive was to keep billing me, not to make me more profitable.
That was the moment Reeve was born. I just didn't know it yet.
Running a $60M Business With Duct Tape
Let me give you context. By 2024, I'd been running a direct-to-consumer ecommerce business for eight years. We'd grown from nothing to $60M in annual revenue. We were running massive ad budgets across Meta and Google, processing thousands of orders a day, and managing a support team that handled hundreds of tickets daily.
On paper, it was a success. In reality, it was held together with duct tape.
We had thirteen different software tools. Shopify for the store. Meta Ads Manager and Google Ads for advertising. Klaviyo for email. Gorgias for support. Google Analytics and a half-configured PostHog instance for analytics. Triple Whale for attribution (which disagreed with Google Analytics constantly). A Notion workspace full of SOPs nobody read. And Slack, where all the real work happened in a chaotic stream of messages.
Every decision required me to mentally stitch together data from three or four sources. "How did that campaign actually perform?" wasn't a question I could answer by opening one tool. It required twenty minutes of tab-switching, export-comparing, and gut-feeling.
I kept thinking: there has to be a better way. An engineer friend told me what I was describing was basically a "data integration problem." Sure. But it was also a people problem. I didn't just need connected data — I needed someone (or something) that could look at the data, understand the context, and do something about it.
The Aha Moment
The idea crystallized during a long flight from LA to New York in June 2024.
I was reading about the latest developments in AI — not the hype, but the actual capabilities. Language models that could reason about complex situations. Systems that could call APIs and take actions. Agents that could handle multi-step workflows autonomously.
And I thought: what if an AI could do what I spend four hours a day doing?
Not just dashboarding. Not just alerting. Actually doing the work. Pausing the campaigns that are bleeding money. Adjusting bids based on real-time performance. Answering the straightforward customer support tickets. Sending the emails at the right time. Connecting all the data into one coherent picture and acting on what it sees.
I started sketching the architecture on my laptop. By the time we landed, I had the bones of what would become Reeve.
Why a Rooster?
People always ask about the name and the mascot. Here's the real story.
When I was building the first prototype, I was working out of a house in the countryside. Every morning at 5 AM, my neighbor's rooster would start crowing. Every. Single. Morning. Without fail.
At first, it was annoying. Then it became a running joke with my co-founder. "At least the rooster's reliable," he'd say, after yet another API broke at 2 AM.
But it stuck. Because that's exactly what we were building — something that's always on, always alert, always the first one up. Something that watches over your business the way a rooster watches over the flock.
We named it Reeve — which is an old English word for a steward or overseer. Someone appointed to manage an estate on behalf of the owner. The rooster became our mascot because, honestly, it was funnier than another abstract tech logo. And because the metaphor works: Reeve crows when something needs your attention.
The Hard Part Wasn't the AI
Everyone assumes the hardest part of building Reeve was the AI. It wasn't.
The AI part — connecting language models to ad platform APIs, building the reasoning engine that decides when to act versus when to alert, training the system to understand ecommerce context — that was hard, but it was tractable. We had good models, good APIs, and a clear set of problems to solve.
The hard part was building something that operators would actually trust.
Because here's the thing about giving an AI system access to your ad account: you're handing it the keys to something that can burn $10,000 in a day if it makes a bad decision. That's not a theoretical risk. That's a "wake up to an empty bank account" risk.
So we built Reeve with trust as the core design principle. Every action is logged and explained. You can see why Reeve made every decision. You can set guardrails — spend limits, CPA ceilings, approval requirements for big changes. You can start in "advisor mode" where Reeve recommends but doesn't act, and graduate to autonomous mode as you build confidence.
The goal was never to make operators obsolete. It was to give operators superpowers. To take the 80% of their work that's monitoring, adjusting, and reacting, and automate it — so they can focus on the 20% that's actually strategic.
What Reeve Is Today
We launched the first version of Reeve in late 2025. Today, it's grown into something beyond what I originally imagined.
The core platform connects to Shopify, Meta Ads, Google Ads, TikTok Ads, Klaviyo, Google Analytics, PostHog, Stripe, and a growing list of other tools. It handles ad management, customer support, email campaign optimization, revenue analytics, social media intelligence, and even recruiting.
But the thing I'm most proud of isn't any individual feature. It's the integration. When Reeve spots a spike in support tickets about a specific product, it cross-references with that product's ad campaigns and can recommend pausing ads for a product with quality issues — before you waste another dollar driving traffic to something that's going to generate refunds.
That kind of cross-functional intelligence was exactly what I couldn't get from agencies, and exactly what I couldn't build with thirteen disconnected tools. It's the thing I wished existed when I was running my own business.
Building in Public
We're building Reeve in a way that I would have appreciated as an operator: transparently.
We publish our changelog. We write about our thinking. We share what works and what doesn't. This blog will be a running commentary on the product, the market, and the lessons we learn along the way.
We're not pretending to have everything figured out. Reeve is still early. There are features we haven't built yet, integrations we need to add, and edge cases we haven't handled. But the core product works. It's managing real ad budgets, handling real support tickets, and driving real results for real businesses.
The Future
My vision for Reeve hasn't changed since that flight two years ago. I want to give every operator — whether they're doing $1M or $100M in revenue — the growth team they deserve.
Not the growth team they can afford (which usually means one overwhelmed generalist or a mediocre agency). The growth team they deserve — one that knows their business inside and out, works 24/7, never drops the ball, and gets smarter every day.
We're just getting started. And every morning, the rooster crows.