Filing the LLC is the part everyone leads with, and the part that matters least. You can be a marketing agency on paper in an afternoon. Whether you're still one in eighteen months comes down to two questions almost no starter guide takes seriously: can you get someone to pay you for a clearly-defined service, and can you keep selling it while you're buried in delivering it. This guide is about those two problems.
Sell the work before you build the company
The most useful thing you can do in week one is not register a business. It's to define one service, for one kind of buyer, and sell it to a paying client. That sale is the only evidence that your agency is a real business and not a plan for one.
Concretely, that means resisting the urge to be full-service. "We do everything for everyone" is how you end up competing with thousands of other generalists on price alone, because a buyer has no way to tell you apart from the next one. A specific wedge — paid social for consumer subscription apps, technical SEO for B2B SaaS, lifecycle email for e-commerce — does three things at once. It lets you charge more. It makes you faster, because you're repeating similar work instead of learning a new industry every engagement. And it makes you referable, because someone can describe what you do in a single sentence.
You don't have to marry the niche. Plenty of agencies start narrow to get traction and widen once they have proof and cash. But starting broad, before you've sold anything, is the reliable way to stay cheap, generic, and forgettable. Validate the wedge by closing one deal — ideally before you quit your job — and let the entity, the site, and the tooling follow the contract rather than precede it.
Price on the math, or the margin quietly disappears
The most common pricing mistake isn't charging too little in absolute terms. It's building the fee around the hours you can see — the actual campaign work — while ignoring the hours you can't: account management, reporting, revisions, the status call that runs long, the software subscriptions. Add those back in and the true cost of servicing a retainer is often two to three times the visible delivery hours. Agencies that price on visible hours are running at a loss they won't notice until they're too busy to fix it.
Work it backwards instead. Estimate the real hours a retainer will consume, with a buffer for scope creep, then set the fee to hit a delivery margin in the region of 55–60%. That one discipline is most of what separates agencies that compound from agencies that grind.
You have three basic models to choose from, and the choice shapes the whole business:
- Retainer. A fixed monthly fee for an agreed scope. You sell once and deliver for twelve to twenty-four months, so your cost of sale per dollar of revenue is low and cash flow is predictable. The tradeoff is that you commit capacity before you know exactly what each month will demand. This is what most agencies should build toward.
- Project. A flat fee for a defined deliverable — an audit, a launch, a campaign build. Good for testing a new client before a retainer, but you're re-selling constantly and carrying higher overhead per engagement. Aim for 50–60% gross and 20–25% net, and treat projects as a doorway to retainers rather than a business model.
- Performance. Fees tied to results — a share of ad spend, a cost per lead, a revenue split. Tempting because it sounds fair, risky early because your income now depends on variables you only partly control. Most agencies that use it blend a smaller retainer with a performance bonus instead of betting the whole fee on outcomes.
The first hire takes you off the critical path
In the beginning you are the entire company: the person who sells the work and the person who does it. That's fine until it isn't, and the moment it stops being fine is specific and predictable. When you get busy delivering, you stop selling — there are no hours left for it. Two months later the projects wrap, you look up, and the pipeline is empty and has to be refilled from zero. So you sell hard, land work, get busy, and stop selling again. That oscillation is the feast-or-famine cycle, and it's the default state of a solo agency, not a sign you're doing something wrong.
The way out isn't heroic effort, it's taking yourself off the critical path. The first contractor or hire usually isn't there to grow the agency — it's to protect the one activity you keep dropping. Some founders hire delivery help so their own hours free up for sales. Others bring in help on the sales and account side so delivery quality doesn't collapse when they're pitching. Either way the point is the same: an agency capped by one person's calendar can't grow past the size of that calendar.
The challenge that actually sinks new agencies
It's rarely the marketing. Founders start agencies because they're good at the craft. What catches them out is the finance and the structure underneath the craft, and it shows up as three interlocking problems.
Cash flow is the one that kills first. In agency surveys it's the most-cited money problem — on the order of 29% call it their single biggest pain, and about a third say it's capping their growth. The mechanics are simple and unforgiving: you do the work in one month, invoice at the end, and get paid weeks after that — assuming the client pays on time. Payroll and your own bills don't wait for a client's accounts-payable cycle. A profitable agency can still run out of money.
Client concentration is the one that kills fastest. It's normal for a young agency to have one or two clients making up most of its revenue. That feels like success right up until the anchor client cuts their budget, brings marketing in-house, or simply leaves — and half your income disappears in a single email. Diversifying away from concentration is a survival task, not a growth luxury.
Founder dependency is the one that caps everything. As long as every deal needs you to close it and every account needs you to run it, the business can only grow to the size of your personal bandwidth. The three problems feed each other: the founder is too busy delivering to sell, so the pipeline thins, so they cling to the big client, so concentration rises, so a single loss becomes existential.
Referrals are the channel that compounds
Every agency's first clients come from the same place: the founder's existing network — a former employer, an old colleague, someone who already trusts their work. That's a great start and a terrible system, because it runs out. Once it does, most founders default to cold outreach, content, and paid lead gen. All of those can work, all of them slowly, and all of them drop you into an attention contest with every other agency running the identical playbook. None of them answer the buyer's real question: why you, over the ten other agencies also sitting in the inbox.
The channel that consistently outperforms is the one that removes that question. A warm introduction from someone the buyer already trusts converts faster, resists price haggling, and arrives pre-vetted. The catch has always been making referrals happen on purpose instead of by luck.
That's the specific gap a referral circle closes. It's a small group of non-competing independent professionals who serve the same kind of client — a fractional CMO, a startup lawyer, a recruiter, a dev agency — and deliberately send work to each other. When a lawyer closing a founder's funding round hears that founder complain their growth has stalled, the marketer in the circle gets the introduction, before that founder ever opens a tab to go looking. It doesn't replace your other channels. It decouples your pipeline from your own spare time, which is the exact thing the feast-or-famine cycle destroys.
Referna matches independent experts — agency founders included — into circles of 8 to 12 non-competing peers serving tech and startup companies. Every member interviews with the full circle and shares a real client success story before joining, moderated by Gordy, Referna's AI community manager, so the leads that reach you carry real trust instead of being a cold pitch you had to make from scratch.