From “MacGyver Marketing” to Autonomous Marketing: How SMBs Evolve

I have had a version of the same conversation more than three hundred times. Different founder, different industry, different country. Almost always it starts with a variant of this sentence: “We have been doing marketing for years, but I don’t know if it’s actually working.” By minute ten I can usually tell which of four stages they are stuck in. By minute twenty the founder can too, and the relief is visible. Naming the stage is half the work.

This is the framework I use. It is not borrowed from a McKinsey report or a Gartner pyramid. It is the pattern I see when I sit with founders. Four stages. Each with its own symptoms, its own costs, its own revenue band where it usually appears, and its own exit ramp.

Before I walk through it, one honest note. Most founders, including me five years ago, assume they are further along than they actually are. If while reading you feel a small sting of recognition, that is the framework working. Stay with it.

The broader context for this piece: the AI marketing playbook for SMBs covers the tools layer, the agency vs DIY vs AI comparison covers the execution options, and the deep dive on the FastStrat agents covers the Stage 4 machinery. This piece sits above all of that: where are you, and where should you go next.

Stage 1 — “MacGyver Marketing”

The founder duct-taping tools.

Typical revenue range

0 to roughly 1 million USD ARR. Often pre-product-market-fit or just past it. One to ten employees.

What it looks like

The founder is the marketer. They have a Canva account, a Mailchimp account, a half-built Instagram, a website that works most days, and a spreadsheet where campaigns go to die. Every tactic is a reaction. A competitor runs an ad; the founder runs an ad. A customer asks if they have a newsletter; the founder starts a newsletter. Somebody reads an article on TikTok; the founder tries TikTok for three weeks.

MacGyver was a television character in the 1980s who defused bombs with a Swiss Army knife, a rubber band, and whatever was in his jacket pocket. The marketing equivalent: founders who ship, constantly, from whatever materials they have at hand, with zero time for planning and zero tolerance for anything that does not produce visible movement in the next 48 hours.

Symptoms

• No written plan. If pushed, the founder can describe “what we are doing” but not “why, in priority order.”
• The tool stack grew by accretion. Nobody decided on Mailchimp; it was free when they started.
• Attribution is vibes-based. “Facebook kind of works.” “LinkedIn brought us a big one, I think.”
• Content is irregular. Two posts one week, silence for six.
• The founder is the bottleneck. Everything waits for their 30 minutes.

Hidden costs

The founder’s time is the biggest cost and it is invisible. Let’s say the founder spends 8 hours a week on marketing and their time is worth 150 USD per hour. That is 62,400 USD per year, unpaid, buried. Add the software spend that nobody audits (usually 200-500 USD per month of tools the founder is not using). Add the opportunity cost of everything the founder is not doing because they are marketing instead.

The deepest cost is compounding. Stage 1 rarely produces cumulative assets. Each post starts from scratch. Each campaign re-learns what the last one learned. There is no flywheel.

How to know you’re stuck

You are stuck in Stage 1 if: (a) you cannot in one paragraph describe your ICP, channel mix, and this quarter’s priority; (b) when a good week happens you cannot explain why; (c) when a bad month happens you cannot explain why; (d) hiring your first marketer feels expensive but every other option feels like more of the same.

How to evolve

Three moves, in order:

1. Write the plan. Not a slide deck. A document you would hand to a new employee on day one that says: who we sell to, why we win, what channels we run, what success looks like this year. The annual marketing plan guide covers the structure.
2. Pick one channel to be good at. Not five, badly. One, deeply. This is where most Stage 1 founders resist: the fear of missing out is strong. The channels you choose depend on your ICP, which is why step 1 has to come first.
3. Set up measurement before you need it. UTM tags, a CRM field for lead source, a simple weekly metric review. Without this, Stage 2 or 3 will feel as random as Stage 1.

Exit signal: you can answer “why is this quarter better or worse than last quarter” with a specific answer, not a shrug.

Stage 2 — “Agency-Dependent”

Outsourced, cannot scale it, no internal capability.

Typical revenue range

1 to 10 million USD ARR. Usually between 10 and 40 employees. The founder has handed marketing off to an agency because they have to, and because a friend said so.

What it looks like

An agency runs the work. Monthly retainer, somewhere between 2,500 and 25,000 USD per month. Weekly or biweekly status calls. Decks. A shared Slack channel. The deliverables are real, usually competent, sometimes very good. The founder is relieved to not be doing marketing themselves.

The problem is not that the agency is bad. The problem is that the capability lives at the agency, not in the business. Fire the agency and marketing stops on Tuesday.

Symptoms

• Nobody internal could describe the current strategy in detail without looking at an agency deck.
• When a campaign wins, the lessons stay with the agency.
• When a campaign loses, the agency explains it as “algorithm changes.”
• The agency’s work is competent, but you have no independent way to judge whether it is competent for your business specifically.
• Scope creep in one direction (more billable hours) and scope starve in another (things you actually need).

Hidden costs

The visible cost is the retainer. The hidden costs are three.

First, translation tax. Everything you want to do has to be explained to the agency, translated into their internal systems, and translated back out to you. The translation loses 20-30% of intent each pass.

Second, calendar tax. Agencies work on agency calendars. If you need something in a week, the answer is often three weeks.

Third, dependency tax. If the agency relationship ends, you pay for the knowledge transfer (often badly), pay for the replacement search, pay for the ramp of a new agency. Agency switches cost most SMBs a lost quarter.

For more on the economics, how much a small business should spend on marketing shows where agency spend fits in total marketing budget.

How to know you’re stuck

You are stuck in Stage 2 if: (a) you feel captive to the agency even while believing they are good; (b) you have had the same “we need to bring some of this in-house” conversation for three consecutive quarters without moving; (c) the idea of not having the agency feels dangerous even though you cannot name the specific thing they do that you could not replace.

How to evolve

The move is not firing the agency. It is selectively reclaiming capability.

1. Own the plan. The annual plan must live with you, not with the agency. You can hire the agency to produce it, but the document lives in your drive, the decisions are yours, and an outside party could read it and understand your business.
2. Own measurement. Your CRM, your analytics, your KPI dashboard. The agency can execute against these but cannot own them. If they own them, you are hostage.
3. Reduce agency scope to their genuine edge. Most agencies have one or two things they do better than you ever will (high-end creative, specialized paid media, PR). Keep those. Move strategy, planning, personas, competitive research, calendar, and KPI work in-house, possibly with AI augmentation. The 60-minutes-vs-3-months post walks through which deliverables shift easily.

Exit signal: you can fire the agency tomorrow without losing marketing capability for more than two weeks.

Stage 3 — “Tool-Stacked”

Has tools but no strategy. Roughly 52% of SMBs live here.

Typical revenue range

2 to 50 million USD ARR. 20 to 200 employees. A small marketing team of 1 to 5 people. Enough budget to buy tools and enough ambition to try most of them.

What it looks like

HubSpot. Mailchimp (still, somehow). Webflow or WordPress. Google Analytics 4. Meta Ads Manager. Google Ads. A social scheduler like Later or Buffer. Maybe Semrush. Maybe Ahrefs. Maybe Hotjar. Maybe ChatGPT Plus. Maybe Jasper. Maybe Notion for content calendar. Maybe Asana. Maybe a BI tool that nobody opens.

Every tool does something. Not every tool does something necessary. Not every necessary thing is being done. The team is busy but not strategic. Dashboards exist but nobody has agreed on what they mean.

Why 52%? I am working from an aggregate of HubSpot, Salesforce, and US Chamber of Commerce SMB surveys cross-checked against our own user base; the figure moves between 48% and 55% depending on the study, and the midpoint is roughly right. This is, by a large margin, the most common stuck-point for US SMBs in 2026.

Symptoms

• More tools than people to use them. Tool audit reveals two or three subscriptions that nobody has opened in 60 days.
• Work is happening, but tied to nothing. A lot of posts. A lot of campaigns. Unclear relationship to revenue.
• The team talks about “what we are doing this week” but cannot describe “what we are doing this year.”
• Every tactic has a champion internally who advocates for their tool.
• Meetings about dashboards outnumber meetings about decisions.

Hidden costs

Software spend alone for a Stage 3 SMB is usually 1,500-8,000 USD per month. But the bigger cost is that the stack pretends to be a strategy. Tools are not strategy. The most expensive failure mode in Stage 3 is believing that because you have the tools, you have the thinking.

The second cost is team burnout. Marketers in Stage 3 companies are the busiest unhappy people I know. They work constantly and cannot explain why things are or aren’t working.

How to know you’re stuck

You are stuck in Stage 3 if: (a) your team can describe the tools but not the strategy; (b) tool evaluation and renewal take more meeting time than actual marketing decisions; (c) when asked “what are the three biggest things we are doing this quarter and why,” you get three different answers from three team members.

How to evolve

The evolution from Stage 3 to Stage 4 is the most counterintuitive one. The instinct is to add another tool (the one that will finally unify everything). The correct move is the opposite: add a plan, and let the plan tell you which tools to keep.

1. Write or generate the plan. Same artifact as the exit ramp from Stage 1, but now you have data and team to inform it. For the methodology, the annual plan guide. For a real case, the SAGA Audiovisual case study.
2. Audit the stack against the plan. Every tool must earn its place by supporting a specific part of the plan. Cut the ones that do not. You will usually find 30-40% of current spend is removable without loss.
3. Introduce an orchestration layer. A plan plus a remaining stack still leaves you with the coordination problem. This is where agentic platforms earn their keep: they sit above the tools and execute against the plan. Not to be confused with another point tool. It is the layer that makes the other tools work in concert.

Exit signal: every recurring marketing meeting starts with “relative to the plan, here is what happened” instead of “here is what is in our tools.”

Stage 4 — “Autonomous Marketing”

Plan-anchored, AI-executed, measured.

Typical revenue range

Surprisingly, no hard lower bound. I have seen 2 million USD ARR companies in Stage 4 and 40 million USD ARR companies in Stage 3. The determining factor is not revenue; it is whether the company has built plan-anchored execution as an operating system.

What it looks like

There is a current, written annual plan. Everyone on the team can describe the ICP, the channel priorities, the current quarter’s goals, and the KPIs. The stack is deliberately small. Marketing work is executed through a combination of human judgment and AI agents, with clear approval gates between recommendation and action.

Content is produced at volume but tied to pillars defined in the plan. Campaigns are launched with clear hypotheses and closed with clear learnings. Budget moves based on performance data, not gut feel or political pressure.

The team is smaller than it used to be. The output is larger. The founder is not in the critical path.

Symptoms (healthy ones)

• When a new team member starts, the plan is the onboarding document.
• Quarterly reviews are about “did the plan hold” not “what did we do.”
• The team can explain both wins and losses in terms of the plan’s assumptions.
• Tool adds are rare. Tool cuts are regular.
• AI agents do the research, drafting, analysis, and recommendation work that used to require two more headcount.

Costs

Stage 4 is not cheap, but it is cost-efficient. Typical Stage 4 SMB runs 2,000-12,000 USD per month in combined marketing software plus AI platform spend, replacing 25,000-75,000 USD per month in combined agency plus under-leveraged tool spend at Stage 3. The agency that remains (if any) is scoped narrow and deep.

How to know you are here (honestly)

You are in Stage 4 if: (a) an external reader could pick up your plan and your KPI dashboard and understand what you are doing and why; (b) new hires reach productivity in days, not months; (c) performance conversations are about adjusting the plan, not about whether there is a plan.

You are not in Stage 4 if you have AI tools but no plan. That is Stage 3 with a newer tool. The distinction matters.

What keeps companies stuck short of Stage 4

Three things, almost always.

• Founder identity. Some founders enjoy being the marketing hero. Stage 4 requires letting go. Some never do.
• Agency relationships. Some agencies actively resist their clients reaching Stage 4 because Stage 4 compresses agency scope.
• Tool lock-in. Some stacks are expensive to unwind. Some teams have political attachments to specific tools.

How to evolve into Stage 4 (if you are in Stage 2 or 3)

1. Anchor on a living plan. Not a PDF, a living document. Revisited quarterly, updated on material change, owned internally.
2. Layer AI at the orchestration level. Agents that work against the plan, not tools that work in isolation. This is the architectural choice FastStrat was built around; other platforms are possible. The principle is what matters.
3. Shrink the stack deliberately. Fewer tools, used deeply. If a tool does not support a plan priority, cut it.
4. Keep humans in the loop at action boundaries. AI recommends; humans approve. Do not let anyone sell you fully autonomous action for SMB marketing yet. The category is too early.

The one-page version of the framework

Stage	Who runs marketing	Typical revenue	Signature symptom	Exit move
1. MacGyver	Founder alone	0-1M USD	No plan, reactive tactics	Write the plan; pick one channel
2. Agency-Dependent	Agency	1-10M USD	Capability lives outside the business	Own the plan and measurement; compress agency scope
3. Tool-Stacked	Small internal team	2-50M USD	Tools without strategy; 52% stuck here	Plan first, audit stack, add orchestration layer
4. Autonomous	Plan + AI + small human team	No hard range	Plan-anchored execution as operating system	(Maintenance: quarterly plan reviews)

Common patterns I see that do not fit neatly

Two patterns deserve a note because they break the neat progression.

The “Stage 2.5” trap. Some SMBs do not fully graduate from Agency-Dependent; they add tools and internal headcount but keep the agency on retainer doing overlapping work. Result: paying twice for similar capability, with neither side having full accountability. Usually the fix is to make the agency responsible for one narrow, measurable outcome and stop paying them for strategy overlap.

The “Stage 3 with AI window dressing” misdiagnosis. Companies that added ChatGPT Plus and Jasper to an already over-stacked toolset often feel like they are in Stage 4. They are not. AI as another tool in the stack is Stage 3 with novelty. AI at the orchestration layer, tied to the plan, is Stage 4. The distinction is architectural, not feature-based.

A founder’s honest confession

I spent more than a year in Stage 1 with my first venture, roughly 18 months in Stage 2 with my second, and watched two companies I advised live in Stage 3 for the better part of three years each. I did not recognize the stages while I was in them. I recognized them looking backward. The framework came from noticing how similar the stories were once I started hearing them from other founders.

Which is the small disclosure worth making: I am biased toward Stage 4 because I built a company in it. FastStrat is a Stage 4 enabler. That is not a secret and it is also not the point of this post. The point is that the stages are real whether you use FastStrat or anything else, and the exit ramp from each stage is the same regardless of which tool you pick at the end.

Where to go from here

Pick the stage that honestly describes you. Run the exit move for that stage. If you are in Stage 1, the annual marketing plan guide is the first read. If you are in Stage 2, the 60-minutes-vs-3-months post walks through what to bring in-house first. If you are in Stage 3, the agency vs DIY vs AI comparison and the FastStrat agents deep dive cover the orchestration-layer question. If you are in Stage 4, the AI marketing trends for SMBs in 2026 and the marketing case studies are the kind of reading that sharpens an already-mature practice.

For a live example of a company moving through the stages on camera, the SAGA Audiovisual case study. For where budget fits regardless of stage, how much a small business should spend on marketing.

FAQ

Can you skip stages?

Partially, yes. A well-funded startup with strong leadership can jump from Stage 1 to Stage 4 without passing through Agency-Dependent. Most SMBs, though, move one stage at a time because each stage teaches a lesson the next stage requires.

Can you regress?

Yes. I have seen Stage 4 companies regress to Stage 3 after a leadership change, a layoff, or a pivot that invalidated the plan. The cure is to rebuild the plan. Regression is not failure; it is the system telling you the plan is stale.

Is Stage 4 always the goal?

Functionally yes, for any SMB that wants to scale marketing without scaling headcount linearly. The alternative is either indefinite agency dependency (expensive) or indefinite tool-stacking (busy without results).

How long does it take to move one stage?

Fastest I have seen: Stage 1 to Stage 2 in 3 months, Stage 2 to Stage 3 in 6 months, Stage 3 to Stage 4 in 6-9 months. Typical: roughly double those. The Stage 3 to Stage 4 transition is the hardest and most rewarding.

What if our category is too complex for Stage 4?

Some categories (deep-science B2B, regulated healthcare, certain enterprise SaaS) genuinely require more human judgment than a standard Stage 4 system handles. The right adaptation is Stage 4 architecturally (plan-anchored, AI-augmented, measured) with a higher ratio of human strategists and specialists than a typical SMB needs.

Does Stage 4 replace marketers?

No. It replaces the parts of marketing work that were always mechanical (research compilation, first drafts, calendar generation, KPI compilation) and frees marketers to do the parts that are actually human (taste, judgment, relationships, creative direction).

Next step

If you want to see the Stage 4 toolkit in motion, meet the FastStrat AI team, read the FAQ, or see pricing.

---

About the author. Walter Von Roestel is CEO of FastStrat. He has been building and advising SMB marketing operations since 2019, from agency side to product side, across hundreds of companies in the US and LATAM.