Essay · 01
Why I'm building one vertical-AEO brand before a horizontal SaaS
· 9 min read
In November 2024 I had a working prototype of a horizontal AI-visibility tool sitting on my hard drive. It scraped the public AI Overviews on a Google SERP, parsed the citations, and produced a one-page report telling the buyer which sources the AI engines were quoting in their category. It worked on landscapers in Charlotte. It worked on driving instructors in Manchester. It worked on dog groomers in Denver. The problem was that I had no idea who to sell it to, because the answer was "everyone with a website," which is the same as "nobody in particular."
I parked the prototype for six weeks and came back to a different question. Not "what does this tool do," but "whose specific Tuesday morning does it fix." When I stopped being able to picture that person, the prototype stopped feeling like a product and started feeling like a feature in search of a buyer. That is the moment I decided the next two years of work would be one vertical-AEO brand at a time, not a horizontal SaaS, and this essay is the case for that decision.
The thing horizontal tools quietly outsource to the customer
Every horizontal SaaS I have admired up close — Notion, Airtable, Zapier, Linear in its early years — does one thing well and then makes the customer do the integration work of figuring out which of their own problems it solves. That works when the customer is technical, has a real budget, and can spare a week to wire it into their workflow. It works less well when the customer is a 52-year-old landscaper in Asheville who is already irritated that Angi sold the same lead to four of his competitors this week.
The horizontal AI-visibility category right now has Ahrefs Brand Radar, the various "GEO" dashboards being bolted onto SEMrush and Surfer, a wave of YC-funded plays like Profound and Athena, and roughly thirty Twitter accounts promising to track your ChatGPT citations for $99 a month. They are real products. Some of them are good. None of them tell a small-business owner what to do tomorrow morning in a sentence he would repeat to his wife at breakfast. They tell him his "share of voice" went up two points in Perplexity, which is meaningful to a marketing analyst at a Series B company and gibberish to the guy who fixes lawns.
The translation layer is the product. In a horizontal tool, the customer pays for the translation. In a vertical tool, the operator pays for it once and amortises it across every customer in the niche. That is the whole thesis, and everything else in this essay is a consequence of it.
What I got wrong about this in 2017, and again in 2023
Through the back half of the previous decade I ran a one-man white-hat SEO consultancy. Those were the years when a lot of the loudest voices in the field were openly running private-blog-network schemes — buy expired domains with residual authority, prop them up with thin scraped content, point the link-juice at client sites until Google noticed, which Google always does. I never ran that play. I told the prospects who asked for it that I would not, and I lost the work to people who would. What I underestimated at the time was that refusing the shortcuts was not enough on its own. My positioning — "SEO for ambitious small businesses" — left every client engagement starting from zero. New industry, new signals, new map. I never built compounding expertise in anything, because the next client was always a different shape from the last. When Google's algorithm updates and a wave of self-serve SEO tools compressed the middle of the market in 2016 and 2017, I had no specialist moat to retreat into. The mistake was not the tactics I refused; it was the horizontal positioning that gave me nowhere to specialise in the first place.
I forgot that lesson in 2023, when the first wave of generative-AI tooling arrived and I sketched a plan for a horizontal "AI workflow consultancy." Same shape as the 2017 consultancy, different decade's hype cycle. I did not ship it, mostly out of laziness, and I am now grateful for the laziness. By early 2025 the market for that exact positioning had filled up with people running the same play, and the ones doing well had narrowed to a single vertical — legal AI, accounting AI, recruiting AI — within twelve months of opening their doors. The horizontal pitch was a six-month on-ramp to a vertical pitch they had not planned for.
So this is the second time I have watched the same lesson land. I would rather not need a third.
The boring economics of one vertical at a time
The argument for going horizontal first is usually a TAM argument. There are roughly 33 million small businesses in the US. Even if a horizontal AI-visibility tool reaches one tenth of one percent at $50 a month, that is $20m a year, which sounds better than any single vertical can plausibly produce. The argument is correct on the spreadsheet and wrong in the world, because horizontal tools do not reach one tenth of one percent of US small businesses from a standing start by a solo operator. They reach it after a Series A and a 20-person sales team, and the version of me that ships that company is not the version of me writing this sentence.
The boring vertical numbers, by contrast, work for one operator. There are roughly 600,000 landscaping and lawn-care businesses in the United States. If GreenRank — the first sub-brand in this portfolio, live at greenrank.garden — reaches one half of one percent of them at the $49/month tier, that is 3,000 customers and roughly $150,000 a month in revenue. The conversion math to get there is brutal but legible: I know what a landscaper Googles, I know which forums they read, I know what they pay Angi per lead, I know what their slow season looks like. I do not know any of those things about "small businesses" in general, and I am suspicious of anyone who claims they do.
The trade-off is real, and I want to name it rather than gloss it. The vertical bet caps the upside. If GreenRank works exactly as planned, it never becomes a billion-dollar company. It becomes a steady cash-flow business that funds the second sub-brand, which funds the third. That is fine by me. I am building a portfolio of small, durable, owner-operated brands, not a venture-sized platform. If you are reading this because you are deciding between vertical-first and horizontal-first for an outcome with a different shape — a fund-returner, an acquisition, a category-defining IPO — the answer for you is probably different and this essay is not advice for your situation.
What a single vertical lets you do that a horizontal tool cannot
The thing nobody warns you about until you ship a vertical product is how much faster the feedback loops get. When the buyer is always a US landscaping owner-operator with two employees, every support email is a data point about the same person. Within a month I knew that the word "visibility" lands and the word "citations" does not. Within two months I knew that a $99 audit is the right price point and a $199 audit is the wrong one, not because of a pricing experiment but because three different people in a row used the same phrase — "that's a tank of diesel" — when I asked about the cheaper number. You cannot run that experiment in a horizontal tool because the comparison set is too noisy.
The same goes for the content side. AEO — getting your business cited in AI answers when somebody asks one of these engines a question — has a different shape in every industry, because the AI engines pull from different sources for different categories. For landscapers in the US, the citations are concentrated in a handful of trade directories, a few homeowner-forum threads, the local newspaper review roundups, and the company's own service-area pages. For wedding photographers it would be Pinterest, The Knot, Reddit, and Instagram captions. For commercial HVAC it would be Reddit, BBB, and specific industry publications. A horizontal tool either flattens those differences and becomes useless, or surfaces them all and overwhelms the buyer. A vertical tool only ever has to know one map.
Patrick McKenzie wrote a line years ago that I keep coming back to: "concrete beats abstract." His example was a SaaS founder who could not get traction selling "appointment scheduling software" and then sold the exact same product as "appointment scheduling for bingo halls" and immediately had a business. The product did not change. The map of the customer's world did. I am betting the same dynamic exists in AI-visibility tooling, and I am betting it harder than is comfortable because I am only building one of these at a time.
What changed in the work because of this
Three concrete things changed when I committed to vertical-first.
First, I deleted half the prototype's features. The horizontal version had a "compare across categories" view that took six weeks to build and answered a question no single buyer was asking. Gone. The vertical version has a "what your three closest competitors are cited for that you are not" view that took four days to build and is the thing every landscaper I have shown it to immediately leans toward the screen for. Same underlying data pipeline. Different surface area. The horizontal version respected the engineer in me. The vertical version respects the buyer.
Second, the pricing got simpler. A free score, a $99 one-off audit, a $49 monthly service. No tiers. No "contact us for enterprise." No usage-based meter. The horizontal version had four pricing tiers because I genuinely did not know which buyer to design for. The vertical version has three prices because I know exactly who the buyer is and what their next twenty dollars are already committed to.
Third, the writing got better. This essay exists because the vertical bet generated a real point of view, and a real point of view is the only thing worth publishing under a name. The horizontal version of the same essay would have been three thousand words of throat-clearing about "the changing search landscape." The vertical version can name a number — 600,000 US landscapers — and a competitor — Angi — and a price — $49 a month — and let the reader decide whether the argument lands. If you want to see what the bet looks like in production, it is at greenrank.garden; the rest of the portfolio is documented at /#sub-brands and will grow one brand at a time, on purpose.
The next essay in this series is about the specific research path that picked landscapers as the first vertical out of roughly thirty candidates I had on a whiteboard. The one after that is about the first $99 audit, what shipped, and what I would do differently. If you have a question about any of this that you would like me to answer in writing rather than on a call, the contact form is the right channel.