The One-Person Marketing Department

Picture a company doing $6.2 million a year. Twenty employees. Growing. The marketing department is one person. Call her Maya. Maya is a composite, but the shape of her month is not.

Maya is responsible for Instagram, short video, YouTube, LinkedIn, email, the website, paid ads, reviews, press, sales decks and events. That is eleven places the company has promised to show up. Three outside vendors help, and none of them has ever spoken to the other two.

Her month has 160 working hours in it. In a company built this way they go roughly like this. Ninety-nine hours go to making things: posts, edits, versions, resizes. Twenty-nine hours go to coordinating and approving, which mostly means chasing vendors and reformatting whatever they send back. Eighteen hours go to redoing work that came back the wrong size or with the wrong message.

That leaves fourteen hours for deciding what to make.

Fourteen hours a month is about forty minutes a workday. It is less than two full days out of twenty. In a busy week it is the first thing Maya drops, and it is the last thing her boss notices missing.

The strange part of Maya’s year is that the output went up and the audience did not follow.

That is now the common case. HubSpot surveyed more than 1,500 marketers for its 2026 State of Marketing report. Of them, 83.5% said they are expected to produce more content than before. Fifty-six percent said the internet is already flooded with AI-generated content. Sixty-five percent said consumers are getting better at spotting that content and ignoring it. Semrush’s 2026 research found roughly 30% of marketers reporting less traffic since AI tools went mainstream.

The work goes up. The audience goes down. That is not a discipline problem and it is not a talent problem. Something underneath the work changed.

Adobe has published results from enterprise customers using its generative tools. At Newell, packaging content came together 75% faster and social assets 33% faster. At Lumen, getting a campaign live went from 25 days to 9.

Think about what happened to photography when film stopped costing money. The shooting became free. The choosing became hard. Nobody had to decide which twenty-four frames were worth a roll anymore, so the skill that mattered moved from the shutter to the edit.

Marketing production went through that same shift. Maya’s competitors got the same tools the same week she did, which means speed stopped being an advantage the moment everyone had it. The advantage moved to knowing what is worth making at all.

That is the thing Maya has fourteen hours a month for.

Gartner surveyed 401 senior marketers between January and March 2026, most of them at companies above $1 billion in revenue. Marketing budgets came in at 7.8% of company revenue, up from 7.7% the year before. That is one tenth of one percentage point. Gartner notes the figure sits 18% below where it was four years ago, when budgets ran 9.1%.

Inside that flat number, 15.3% of budget went to AI, and only 30% of those CMOs said their organization was ready to scale it. Fifty-six percent said they lacked the budget to deliver their own plan.

One caveat is worth stating plainly, because this argument usually skips it. Gartner is measuring billion-dollar companies. Nothing in that survey tells you what a $6 million company should spend, and any benchmark drawn from it and applied to a small business is being stretched past what it can carry. What the survey does show is that the squeeze runs top to bottom. If companies with a full department and a chief marketing officer cannot fund their own plan, the company with one marketing employee is not going to out-budget the problem.

Here is the money in the shape it usually takes. The figures below are an illustration, built on published wage and benefit data.

Maya’s company spends about $164,700 a year on marketing. A social media vendor takes $60,000. Podcast production takes $24,000. Website, email and tools take $12,000. Maya herself is a $48,000 salary that costs the company $68,700 once benefits are added. The Bureau of Labor Statistics put benefits at 30.1% of total compensation for private industry workers in March 2026, so that markup is not aggressive.

At the March 2026 median, $164,700 is roughly two and a quarter American jobs, fully loaded. It arrives as three invoices and a payroll line, and nobody is accountable for the total.

Now price the alternative. Employing the four roles it would take to cover eleven channels, at median wages, with benefits and the software each person needs, runs about $647,000 a year. That is more than a tenth of the company’s revenue, spent before a single campaign runs.

She can afford to use a department. She cannot afford to employ one.

Three doors are open to her today.

An agency covers maybe two of the eleven channels and prices as though it covers all of them. Freelancers cover three, and Maya still writes every brief. Software covers none. It is a twelfth subscription, and speed was never the thing that broke.

Then there is the part nobody prices at all. If Maya leaves on a Friday, the passwords go with her, along with the vendor relationships, the brand history, the unfinished projects and the only working knowledge of how any of it fit together. None of it is written down. Replacing her and getting the replacement productive takes four to six months.

The workload is the visible problem. The continuity is the expensive one.

Nobody owns a power plant. A hospital, a school, a corner shop and a family home all draw from one expensive, expert, always-staffed system. Each pays to be connected, then pays for what it uses.

Look at any electricity bill and there are two lines on it. A service charge that does not move, which pays for crews and spare capacity and somebody awake at three in the morning. And a usage line that moves with the month. A hot August costs more than a mild one, and nobody renegotiates.

That structure is already how businesses buy several expensive things. NetJets and Flexjet sell hours on a fleet, with a monthly management fee and an occupied hourly rate. Amazon Web Services sells a commitment plus what you consume. ADP TotalSource and TriNet sell one HR person to call, with specialists behind them, at a platform fee plus a rate per employee. A fire department is funded all year and used rarely, and nobody calls that waste.

Marketing is close to the last expensive capability still bought as a pile of separate purchases.

The plain version of the alternative is renting a portion of a complete media department instead of hiring the whole thing. The industry word for it is fractional, from the same root as a fraction of a plane. That is the entire meaning of the word.

In 2023 the World Federation of Advertisers and The Observatory International surveyed 45 large multinationals with roughly $60 billion in combined annual ad spend. Sixty-six percent had built an in-house agency, up 16 points from 2020. Another 21% were considering one. Seventy percent held strategic capability inside the company: brand, creative or media.

They kept the judgment. They rented the doing.

The sample is small and the survey is three years old, so read it as direction rather than a headcount. The direction has held since 2020.

Here is the arithmetic that makes this practical, using the illustration above.

Today the company pays $5,000 a month to the social vendor, $2,000 to the podcast vendor, $1,000 for website, email and tools, and $5,725 for Maya. The total is $13,725 a month.

Rebought as one thing, Maya stays at $5,725, now directing the function instead of absorbing it. A defined share of a seven-person team costs $6,750. Vendors and tools sit inside that number. That leaves $1,250 for paid distribution.

The total is $13,725 a month. The change in spend is zero.

Nothing here needs a bigger budget. It needs the same budget bought once.

This is where buyers actually lose money, so the arithmetic is worth being blunt about.

Publicis Groupe finished 2025 with an 18.2% operating margin, the best among the holding companies, and it spends roughly two thirds of net revenue on people. WPP came in at 13.0%, down from 15.0% the year before, alongside a £641 million goodwill write-off. These are the largest and most automated marketing companies in the world, running at a scale no small shop will ever reach.

So when someone quotes $2,000 a month for a full department, that number is not a bargain. Run it against a seven-person unit and it is a loss of roughly 68%, somewhere near $19,000 a month. Nobody absorbs that for long.

From the buyer’s side, sustained loss follows a pattern. The first quarter brings a full team, fast replies and everything delivered. The second quarter gets quieter, later and more junior. By the third quarter one overloaded person is holding the account. By the fourth quarter the emails stop.

Twelve months on, the company is hiring again, a year behind where it started. If a cheap vendor has ever gone silent on you, that is the reason.

The test is frequency. A share of a team pays off only if you would use it every month.

Buy a project instead if you need one thing built once, if your marketing happens in two bursts a year, or if you have one channel and it is working. Buy nothing at all if nobody inside your company can approve anything, because in that case the bottleneck is approval, and no amount of outside capacity fixes it.

There are three things worth doing, and two of them are free.

Add up the real bill. Every invoice, every subscription, the salary with benefits included. Then take it as a share of revenue, and look at what you have actually been spending.

Count the channels. List every place your company is expected to show up, then circle the three you would defend if you had to drop the rest.

Then ask any operator five questions. How many clients can your team hold. Who is my backup, by name. What is not included, and what does it cost. How do your people earn a bonus. What happens when I want to hire my own team. A number with the workings behind it is a good answer. “As many as we need” is not.

Here is the falsifiable version of all this. By the end of 2027, expect capacity to show up on marketing proposals the way commitment tiers showed up on cloud pricing: reserved hours, a stated ceiling, a published rate for anything above it. If the standard proposal eighteen months from now is still one monthly figure and a list of deliverables, this argument was wrong.

The last strategy at Maya’s company was written fourteen months ago. It is still on the drive, still current as far as anyone knows, because the fourteen hours a month that would have replaced it went to resizing an image for Instagram.

Sources: Gartner 2026 CMO Spend Survey (401 marketing leaders, January to March 2026). HubSpot 2026 State of Marketing (1,500-plus marketers). Semrush 2026 content marketing research. Adobe enterprise case studies, Newell and Lumen. US Bureau of Labor Statistics, Employer Costs for Employee Compensation, March 2026. Publicis Groupe FY2025 results. WPP 2025 preliminary results. World Federation of Advertisers with The Observatory International, 2023. Company-level figures for Maya’s employer are an illustration, not a reported case.

Michael Wildes is the founder and CEO of Drive Phase Holding Company, a permanent-capital firm focused on building category-defining companies across business, media (owner of Massif & Kroo), aviation, and impact. After leaving a career as a professional pilot, he spent a year as Business Editor at FLYING Magazine, writing 330+ articles on aviation's transformation. Now he builds permanent-capital companies focused on long-term trends that compound over decades. Based in Arlington, Virginia.

Connect: mikewildes.com | [email protected]