How to Start an SMMA

Almost everything you’ve watched about starting an SMMA was made by someone whose business is selling courses about starting an SMMA — including the 30-day programs, which I’ve read straight elsewhere. This is the other version. I run an agency — Rebel Video, YouTube ads for local businesses — and this is the guide I’d hand my own kid: what the model really is, why most attempts die, and the exact order to build one that doesn’t.

What an SMMA is — and the part the videos skip

SMMA stands for Social Media Marketing Agency: a business that runs marketing — content, ads, customer acquisition — for other businesses. The model is completely real. Local businesses need customers, most have nobody competent running their marketing, and they will pay well for someone who fixes that.

Here’s the part the videos skip: the version being taught is the version that fails. The standard playbook — offer “social media management” to anyone who’ll take a call, post content, charge a few hundred a month — puts you in the most crowded, lowest-paid corner of the entire industry, selling a service no owner can measure. When the client can’t tell whether your work made them money, you’re the first expense cut. That’s the churn treadmill, and it’s where most new SMMAs quietly die.

The one decision that changes the math

Before niches, before outreach scripts, before you buy a domain: decide to sell a result, not a service. “I’ll manage your Instagram” is a service — unmeasurable, interchangeable, priced by comparison shopping. “I’ll put your ad in front of homeowners in your towns who are searching for a new roof right now, and you’ll count the customers it brings you” is a result — measurable, specific, priced by what a customer is worth. Everything below flows from that one decision.

Step 1: Pick one service — and make it one that produces customers

New agencies fail by offering ten things badly. Pick one thing a business will pay real money for, and get genuinely good at it. My case for that one thing being video ads — specifically YouTube ads:

• The result is countable. Views, clicks, calls, booked jobs. You’ll never have to argue about whether your work is working — the numbers say it for you. That’s what keeps clients for years instead of months.
• The client’s budget is small enough to say yes to. A real YouTube campaign runs on $20–30 a day. You’re not asking an owner to bet five figures on you — you’re asking for less than their coffee budget, aimed precisely.
• Almost nobody else is there. Every other new agency is fighting over the same crowded feeds. YouTube — the search engine people talk to like a neighbor, playing on the biggest screen in the house — is still nearly empty of local advertisers. I wrote up exactly why YouTube ads work for small businesses if you want the client-side view of what you’d be selling.

If you want the full map for that specific lane, I wrote it here: how to start a YouTube ads agency.

Step 2: Pick a niche that answers the phone

The best SMMA niche isn’t a secret category — it’s any business where one new customer is worth real money and the owner can count what you send them. Roofers, law firms, solar companies, contractors, med spas. A roofer’s average job is five figures; send him three real customers and he can do the math on you instantly — in your favor. Compare that with a business where a new customer is worth $40 and nobody can trace where they came from. Specific business types, always: “I run ads for roofing companies” opens doors that “I do social media for business owners” never will.

Can you run an SMMA from anywhere?

Yes — and that’s genuinely one of the best things about this business. Everything you do is a phone call, a screen share, and an ad account. I’ve built campaigns for businesses I’ve never physically visited.

But there’s a catch worth naming, because “location independent” gets sold as “you can be vague about where your clients are,” and that’s the version that fails. The businesses that pay real retainers are local businesses — a roofer serving four towns, a law firm serving one county, a solar company serving one utility territory. Their ads are geographically targeted whether you’re sitting in Boston or Bangkok. So you can live anywhere, but your clients need to be somewhere specific, and you need to be able to name it.

The practical version: pick a metro you know, or one you can learn cold — the towns, the seasons, what people complain about. Run your ads there. Where your laptop is doesn’t matter. Where your client’s customers are matters completely.

Step 3: Learn by doing the real thing — for yourself first

Here’s the move almost nobody makes: your first client is you. Before you pitch anyone, run a real campaign advertising your own new agency — real targeting, a real ad, a real landing page, real money at $20–30 a day. Three things happen at once:

1. You learn the craft on the only teacher that matters — a live campaign with your own money in it. No course simulates that.
2. You build your proof. When an owner asks “have you done this before,” you show them the working campaign that reached them.
3. Your clients come to you. The ad that finds business owners searching for marketing help is the same skill you’re selling. Attract clients — don’t buy contact lists, and don’t grind cold DMs into the void. The agency whose own ads work is the agency that gets hired.

Where the content actually comes from

The fear that stops most people: I’m not a videographer, I don’t have a studio, what am I going to post?

You’re not going to post. You’re going to advertise — and the ad that works for a local business is not a production. It’s the owner, in his shop or in his truck, on a phone camera, saying the true thing his customers already feel. I’ve watched a roofer’s phone-shot video outperform a polished agency spot for less than the polished spot cost to storyboard. Substance over form, always.

So the content pipeline is simple:

• One to-camera video from the owner. Ten minutes of his time, your questions, his answers.
• His actual sales conversations. Every objection he answers on the phone twenty times a week is a video. He’s already written your script — he just doesn’t know it.
• The thing he wishes customers understood. Ask any owner that question and stand back.

You direct, he talks. That’s the whole job. You never need a camera crew and you should never sell one.

How to come up with ad ideas that work

Stop hunting for ideas and start with the four moves. Every ad that grows a business does these in order:

1. Intervention — open with the one message the right person can’t ignore. Their belief, the truth they already feel. The wrong people skipping is the filter working.
2. Assurance — quiet the doubt the intervention raises. Calm authority, proof, and the plain reason it works.
3. Encouragement — move them from “this could work” to “I can do this.”
4. Action — one specific thing to do now. Not “learn more.”

Once you have that structure, ideas stop being the bottleneck. You’re not inventing a concept — you’re finding which of the owner’s true statements belongs in move one.

Do SMMA clients pay for their own ads?

Yes — and it is never worth arranging any other way. The ad spend goes on the client’s card, in an ad account the client owns, and you charge a separate fee for the work. You are paid to build and run the campaign. You are not a media reseller and you should never be one.

New agencies get talked out of this constantly, usually by someone selling a “we handle everything” script. Here is what actually happens when you fold the ad budget into your fee: you are now financing your client’s advertising out of your own cash, at no margin, with the money already spent by the time an invoice goes unpaid. One slow-paying roofer and you are underwater on a job you did correctly. Worse, if you ever have to walk away, the account, the audience and the history walk with you — and a client who loses all of that because of a billing dispute becomes a story other business owners in that town hear.

The clean structure, every time:

• The client’s Google Ads and Meta accounts, in the client’s name, on the client’s card. You get admin access. If it’s a brand-new business, they open the account during the onboarding call while you are on the screen share — ten minutes.
• Your fee is invoiced separately and it never moves with their budget. Bill for the campaign, not for the spend.
• Say the daily number out loud in the sales conversation. “$20–30 a day, on your card, and my fee is X.” Owners relax when the money is plainly separated. The ones who flinch at the ad budget were never going to be good clients.
• Everything you build stays theirs. The account, the tag, the audience you build, the videos. Say that early. It is the opposite of what the lead sellers offer, and it lands.

The one exception worth knowing: some large accounts insist on being billed for everything through one vendor. That is a contract with terms and a real finance conversation, not a starting point for a new agency.

How to introduce your agency to a business

The introduction that works is not a pitch, and it isn’t a template with the business name dropped into slot one. It is evidence you looked at their market before you opened your mouth. Owners can tell the difference in a sentence and a half, because they get the other kind eleven times a week.

Three rules that do most of the work:

• Lead with something only their market would produce. A competitor already running video. A search term their customers are typing that they own none of. What people watch before they buy the thing they sell.
• Name what you do in one line, and make it specific. “I run YouTube ads for roofing companies” opens doors that “I help businesses grow online” never will.
• Ask for a conversation, not the sale. The first exchange has exactly one job: earning fifteen minutes.

Here is the whole email. It is short on purpose — length reads as a pitch:

On the phone it is the same three beats, shorter: who you are, the specific thing you noticed, and a question. Then stop talking. The silence after a specific observation is where the meeting gets made.

And the version that beats all of it: have your own ad running. An owner who found you because your ad reached him is not an introduction at all — he called you. That is why the first campaign you build is your own, and it is the entire structure of how I teach this.

Step 4: Price on the outcome, not the hours

The $500-a-month generalist agency is a job with extra steps — a dozen anxious clients, endless deliverables, and churn eating everything you sign. The durable shape is the opposite: a few clients at real retainers, each paying for a result they can count. When your campaign sends a law firm cases or a roofer jobs, your fee is measured against revenue, not against what a cousin’s kid would charge for posting. Fewer, better clients is not just more money — it’s the only version of this business you’ll still want to run in year three.

Step 5: Keep clients by reporting the number they care about

Retention is where agencies are actually built, and retention is a reporting habit: every week, in plain language, what a new customer cost and where it came from. Not impressions, not engagement — what the phone ringing actually cost. The client who always knows their number never wonders what they’re paying you for. Give the campaign two weeks to learn before judging it; then judge it ruthlessly, out loud, together. Honesty about a slow week buys you years of trust.

Do you need an LLC — and the legal side nobody covers

You do not need an LLC to start. You can sign your first client as a sole proprietor in every state in the country, and plenty of working agencies ran their first year that way. Form the entity when there is revenue to protect and a contract to sign under it — not as step one, and never as a substitute for having run a campaign. I practiced law for ten years before I built this, and I have never once seen a business owner ask an agency for its formation documents. They ask what you did for somebody like them.

What actually matters, in the order it will come up:

• A written agreement, even a one-pager. Scope, fee, term, and how either side ends it. Most agency disputes are not fraud — they are two people who never wrote down what “managing the ads” meant.
• Who owns what. Say it in writing: the ad accounts, the tracking tag, the audiences and the footage belong to the client; your process is yours. This is a selling point, not a concession.
• Claims you make about results. Advertising claims have to be true and substantiated, and that applies to the ads you write for a client and to the ads you write for yourself. No invented numbers, no borrowed screenshots, and testimonials have to be real and typical — the FTC updated its rules on exactly this in recent years and enforcement is not theoretical.
• Your client’s own rules. Contractors have licensing and advertising requirements that differ by state, attorneys have bar advertising rules, and medical and financial clients have their own. You are putting words in a licensed professional’s mouth — ask what their rules are before you write, and put the compliance sign-off on them in the contract.
• Calling and texting. The moment your work produces phone numbers that get dialed or texted, TCPA consent language is your problem too. Get it on the form, in writing, before anybody calls anyone.
• Money and paperwork. Separate business bank account from day one, keep the ad spend off your books entirely, and expect to hand a W-9 to any client who pays you more than a few thousand dollars.

None of that is a reason to delay. It is a reason to do the boring hour of setup properly and then go run a campaign. This is general information from my own experience, not legal advice for your situation — entity and contract questions should go to a lawyer licensed in your state.

The mistakes that kill new SMMAs

• Offering everything. Ten services means mastery of none and a pitch nobody remembers. One service, one niche.
• Selling activity instead of outcomes. If the client can’t count it, the retainer is always one budget meeting from cancellation.
• Buying contact lists and blasting cold outreach. You’re a marketer — if your own client acquisition is a purchased spreadsheet, that tells prospects everything. Attract, don’t buy.
• Staying free forever. One or two proof projects, briefly, fine. But a business that doesn’t charge isn’t a business, and owners respect a price.
• Building the brand before the skill. The logo, the fancy site, the LLC paperwork perfected — while zero campaigns have ever run. Run the campaign. Everything else is decoration.

So what should you actually do this week?

Pick the one service (I’ve made my case: video ads). Pick one niche where a customer is worth real money. Then get a real campaign live — for yourself — and learn the craft with skin in the game. You can grind that out alone, or do it inside a structure where the first campaign goes live in week one with someone who runs this every day looking over your shoulder:

Own a local business and want the ads for yourself instead? That’s the in-person tour — Vegas Sep 5 · Santa Barbara Sep 19.

Frequently asked questions