Loading page…
An enterprise web-build shop that refuses any project under $100K, so ten full-time people carry $2M of revenue — and the problems those clients kept repeating became the spec for a product.
about 80% growth the year before that · 10 full-time employees, 12 people in total, which the host works out to roughly $200K of revenue per employee · $100K minimum project value, five or six projects a year, one as large as $850K · the spin-out SEO Central had about 100 companies on its beta list at a planned $140 a year, and had not launched at the time of the interview
"Generally we don't do anything under 100K." Five or six projects a year, ten full-time people, $2M — the price floor is not a policy, it is the entire business model.
Fewer bars = easier, cheaper, or faster for an AI-assisted solo builder. Editorial judgments based on the case details.
Josh Northcott was a web director at a large enterprise software company, and in that job he spent years handing work to the biggest agencies and service providers in the country. What stuck with him was not the work, it was the pattern: "I always wondered why stuff didn't get done on time on budget and all those things, like why is it always a mess, like why does it suck." He and his co-founder Justin left their corporate jobs and started Hounder — he puts it at late 2016 and "about seven years now," while the host's introduction dates the launch to 2015. The firm does large web system implementations: marketing system builds, CMS work, custom design, product applications. The clients he names are Adobe, Shopify, the NHL, UL and the City of Los Angeles, and he describes the platform companies among them as clients and partners at the same time. The numbers are unusually clean for an agency. The first million-dollar year came in 2021. 2022 doubled it to $2M, after roughly 80% growth the year before. The 2023 target is $2.5M. Doing that are 10 full-time employees, 12 people in total, with no army of on-and-off contractors: "we're only growing as the demand grows." Nathan Latka does the arithmetic on air — about $200,000 of revenue per employee. The last thread is the one Josh is hardest on himself about. Seven years of enterprise projects surfaced the same problems again and again, which is the raw material for a product, and he sat on it: "I epically failed by taking us almost seven years to get there." The product, SEO Central, was in beta at the time of the interview.
A $100K minimum has no cold-outreach path to its first client.
The quick-reference card — acquisition channels, replication playbook, risks & traps, and our verdict — is members-only.
Join the Unicorn Club — $5/moCancel anytime, in one clickPaste into Claude Code / Codex and get a working version of this product end-to-end.
Two production-grade prompts per idea — a full build spec and an SEO growth plan — plus 10 tailored build specs a week. Members only.
Join the Unicorn Club — $5/moCancel anytime, in one clickEvery claim traces back to a listed source — open any of them and check the number yourself. The exact words are quoted below, with a timestamp that jumps to the second each figure is said.
“…an agency that broke a million in Revenue in 2021 2 million Revenue last year hoping to break 2.5 this year but they're secretly building an internal SAS tool they…”
0:03 · hear it in the video“…we did it's gonna be uh it's gonna be about only 140 bucks for the whole entire year so annually annual cost of 140.”
13:09 · hear it in the videoData credibility: Founder interview on Nathan Latka's Top Founders podcast, published August 2023, with the guest's figures also listed on the host's getlatka profile for the company. Everything financial is founder-stated on air with no dashboard, invoice or accounting export shown: $1M revenue in 2021, $2M in 2022, a $2.5M target for 2023, about 80% growth the year before, 10 full-time employees and 12 people in total, a $100K minimum project value, five or six projects a year, one project at $850K, and the roughly $200K of revenue per employee that the host calculates live rather than the founder reporting. Two dates conflict inside the same episode: the host's introduction says he left enterprise software in 2015, while the founder says Hounder started in late 2016 and has been running about seven years. The SEO Central product is entirely forward-looking at the time of recording — about 100 companies on the beta list, a soft launch planned for the end of that quarter, and a $140 annual price point that had not been tested on a paying customer. No margin, profit, salary or client-contract detail is disclosed anywhere.
The first invoice in this model is rarely the hundred-thousand-dollar one. A buyer who has never worked with the firm will not hand over a full replatform, but will approve a paid discovery: four to six weeks, $15K to $25K, ending in an architecture document, a content model and a fixed-price build quote.
Two things compound in an enterprise service business, and neither of them is marketing. The first is the logo.
The quiet stretch in project work is structural rather than emotional. Sell $100K, deliver it in four months, and the revenue that looked like a business disappears exactly when the team is too deep in delivery to sell the next one.
There are two products here and only one of them earns money today. The agency builds large web systems for enterprises: marketing system implementations, CMS work, custom design and product applications, for organisations from Adobe and Shopify to the NHL and the City of Los Angeles. Projects start at $100K and there are five or six a year, hand-picked. The second is SEO Central, a spin-out that comes directly from seeing the same problem at every one of those clients. Josh's description of the category it replaces is precise: today an SEO plugin shows you red dots telling you to change a title, rewrite a description, fix a heading. "Imagine a world where you don't have you no longer see red dots... but instead it just does it all for you and optimizes it for you." The plan goes further than generating text — he describes machine learning sitting behind the page watching click-through rate on a ranking result and changing the title and description itself, an autopilot mode for the SEO practitioner rather than a suggestion list. The distribution choice is the part worth studying. It is built for the enterprise content stack the agency already lives in — Drupal, WordPress, Adobe Experience Manager, Sitecore — and sold top-down through the platform relationships Hounder already has, which Josh contrasts explicitly with the product-led, bottom-up motion of the plugin incumbents. At the time of the interview it was a soft launch at the end of the quarter, about 100 companies signed up for the beta, at seocentral.co and seocentral.ai.
The agency's model is stated plainly: no project under $100K, five or six of them a year, plus ongoing services for a subset of clients. Some engagements are far larger — he mentions one at $850K running across a year, and projects that approach a million dollars over several years. That mix produced $1M in 2021, $2M in 2022 and a $2.5M goal for 2023, off 10 full-time employees. The cash-flow problem inside that model is the one Latka pushes on: complete a $100K contract in four months and you have to replace that revenue from scratch. Josh's answer is ongoing services plus month-over-month tracking of what comes in against what it costs, with a 20% annual increase as the planning target even in years when actual growth came in at 80% or 100%. The product prices very differently. SEO Central was planned at about $140 for an entire year, annual rather than monthly. Our read: that number only works as a land-and-expand or a volume play, because $140 a year means roughly 14,000 paying sites to match what ten people already generate in services — and a top-down, enterprise-partnership sales motion is the most expensive way in the world to sell a $140 subscription. Either the price rises for the enterprise tier or the beta list has to convert into something much bigger than a plugin licence.
Headcount is the cost structure, and Hounder runs it tighter than the category norm. Ten full-time employees, twelve people in total, and — when Latka assumes there must be a contractor army behind $200K of revenue per head — a flat denial: "no, we we kind of only we're only growing as the demand grows." The team is people Josh had worked with personally in previous companies, which is how a shop this small takes on Adobe-scale implementations without a bench. The only tool he names for running the business is SEMrush, which he credits as "huge" for content marketing. The book he names is Clone Yourself. He sleeps about six hours. Our read: no margin figure is given anywhere in the interview, so treat the economics as inferred. A services firm with $200K of revenue per employee and no contractor markup is operating at the healthy end of agency benchmarks, where fully loaded delivery cost typically eats 50 to 60 percent of project revenue. The exposure is the same one that makes the number look good — with no bench, a single large project slipping a quarter or a key developer leaving removes delivery capacity that cannot be replaced quickly, and hiring ahead of demand is the very thing this firm has decided not to do.
Copy the price floor first. Almost everything good in this business follows from refusing work under $100K: five or six clients a year instead of thirty, ten people instead of forty, and enough margin per project that the firm never needs a contractor bench. The floor is also what makes the client list possible, because enterprises are not looking for the cheapest supplier, they are looking for one that will not collapse mid-implementation. Copy the origin story too. Josh sold to the exact buyer he used to be — he was the web director commissioning this work, so he knew what the procurement process looked like and why projects failed. That is a repeatable move in any vertical with expensive software implementations. Change the seven-year delay. Josh's own harshest line is about how long it took to productise what the agency had learned: "I epically failed by taking us almost seven years to get there." The pattern data — the same problem seen at twenty enterprise clients — is the asset, and it decays as platforms change. Our read: change the price point as well. A $140-a-year plugin sold top-down through enterprise partnerships is a mismatch between a consumer-grade price and an enterprise-grade sales cost. If the wedge is the trust Hounder already has with Adobe- and Sitecore-scale customers, the product should be priced per site or per seat at enterprise levels, with the cheap tier existing only to seed the market.
Client concentration is the structural risk. Five or six projects a year means one client is on the order of 20% of revenue, and losing two at renewal is a layoff rather than a bad quarter. The same tight team that produces $200K per head has no slack to absorb it. The product risk is the one Latka hammers on and never gets a fully satisfying answer to: "why doesn't yoast just launch a plug-in for those things, why don't they, like how do you beat them, what's your mousetrap?" Josh's answer is speed to market plus the size of the clients who already trust them, going top-down where the incumbents are bottom-up. That is a real advantage with a short shelf life — Yoast, Semrush and Ahrefs all have the distribution, the data and the engineering to ship an autopilot mode, and by 2026 automated on-page optimisation is a feature inside several of them rather than a category. The execution risk sits in the gap between beta and revenue. About 100 companies on a beta list is interest, not a business, and the interview describes a soft launch that had not happened yet. Our read: the third risk is attention. A ten-person agency with $2.5M of committed delivery does not have spare engineers, so every hour spent on the product is an hour not spent on the revenue that funds it — the classic reason agency SaaS spin-outs stall at the beta stage rather than dying outright.
The four founder playbooks — first customers, first dollar, the flywheel, the silent stretch — are in the Unicorn Club, on this idea and on all 1,435 others. The story, the numbers, the sources and every later section stay open to everyone.
Join the Unicorn Club — $5/moCancel anytime, in one click