Last week I wrote down what I'd do to grow in 2026. Twelve things. A list of tactics is easy to write and hard to staff. So here's the other half of how to actually go do it.

Worth saying where this comes from. My whole background is marketing at startups, pre-PMF through Series E and an exit, and all of it under 200 people. If that's not you, this probably isn't either.

I've also kept my teams small the whole way, on purpose. That used to be a preference. Now that AI absorbs a chunk of the work, a bloated marketing team is just easier to see.

If I was building a marketing team from scratch today, here's where I'd start.

The first four roles

Head of or VP, and a player-coach. If they haven't shipped something themselves in the past week, they're not a fit. Ask for the last three things they published and what happened. Their other job is getting rid of the approval layer. If your VP is the bottleneck on a tweet, you hired a manager.

GTM engineer. Data, systems, signals. They tell you which message worked before you spend money on it. Good ones build the thing that makes the collateral instead of making the collateral. Share them with sales if you want, but don't split the reporting line.

Brand and content. One person early. Narrative, storytelling, the founder's posts, the newsletter. Hire a publisher, not a writer. Same person picks the story, writes it, ships it, and watches what happens. Everything they ship gets distributed. And they're comfortable going after the incumbents by name — not "legacy tools," actual company names. Video is a separate skill, so buy the production. This splits in two somewhere around $15M and not before.

Demand. Owns paid, landing pages, and everything between a click and a pipeline. Early on the job is running small experiments until you know which message works where. After that it's putting real money behind the two or three that did. Keep it in house. An agency learns your funnel and then leaves with it.

The rest

Customer stories and data. The part of content everyone skips because it's slower. Case studies and customer video on a schedule rather than whenever somebody remembers. And if your product sits on numbers nobody else has, that's a report a year and the cheapest attention you'll ever buy. Brand and content owns this until there's a second person.

Lifecycle. Onboarding, product updates, the newsletter. Write the job as "email marketer" and you'll get someone who builds campaigns and never asks why customers leave. Write it as a retention job.

Community and partnerships. Ambassadors, affiliates, integrations, marketplace listings, the Slack or Discord. One person early. It splits once partnerships makes real revenue.

Product marketing. Owns the customer calls. The skill is hearing the ask behind the ask and turning it into a line the market repeats back to you. Everything else gets easier once that line exists.

Design and video. Contract it. Hire when turnaround is the bottleneck, not quality.

Events. Depends entirely on what you sell. High ACV, long cycles, six people who have to agree? Field marketing is a real hire and probably an earlier one than anything else in this section. Self-serve? A booth is an expensive way to meet people who won't buy. Either way, dinners and small rooms beat conferences at almost every size. Brand and content runs it until it's someone's whole job.

SEO and AI search. Don't hire for this yet. Most people selling AEO are selling a 2019 SEO deck with the labels swapped. It's three jobs you already have. Brand and content owns what gets published, demand owns the site and the pages people land on, and the GTM engineer owns whether you show up in the answers.

Three questions worth asking to any candidate

  • •Show me the loudest thing you've made. Something that traveled further than the size of the company behind it.
  • •Show me something you built with AI. An agent, a workflow, a prompt library. Ask for a demo, not a description.
  • •Show me what you personally wrote. Not what the team shipped. Not what the agency delivered.

Years of experience I'd mostly ignore.

Two things to check yourself against

Marketing usually lands around 5 to 10% of headcount in B2B. Higher in PLG, lower in enterprise. If you're way off, you either know something I don't or you're about to find out why the number exists.

And eight roles chasing everything at once is how teams end up busy and flat. Pick the three things you're trying to win this year. Staff those so they can't fail. Run the rest at maintenance and stop feeling bad about it.