Hire Spiky, Not Well-Rounded: How to Build a Lean DTC Marketing Team That Punches Above Its Headcount

Picture the org chart a founder slides across the table to me in week one. Five neat boxes under "Marketing". A marketing manager. A social person. An email person. Someone who "owns" the website. A junior who does a bit of everything. All smooth-edged, all reasonable, all hired to be safe pairs of hands.

And the whole thing is doing about 1.4x blended ROAS on ~$40k a month of spend, which is to say it's barely moving.

Here's the thing I've noticed across dozens of these charts. The teams that punch hardest aren't the ones with the most boxes. They're the ones with one or two genuinely sharp people and a deliberate decision about everything else.

So I want to make the case for hiring spiky, not well-rounded. And then I'll tell you which spike to hire first at each stage, because that order matters more than founders think.

The river stone problem

There's a phrase I keep coming back to, borrowed from how some of the best operators describe their own teams. You can either be a river stone or a pointy object.

A river stone is smooth all the way round. Decent at everything, the edges knocked off, nothing that cuts. A pointy object is sharp in one specific place and frankly a bit useless everywhere else, but where it's sharp, almost nobody on the planet is sharper.

Most "marketing managers" you hire for a 7-figure DTC brand are river stones. Not because they're bad. Because the role itself is defined to be smooth: a bit of paid, a bit of email, a bit of organic, a bit of reporting, a bit of agency-wrangling. You've written a job description that rewards roundness.

And roundness, in ecom, is where performance goes to be average.

My take is simple. At small scale you cannot afford average in the one or two areas that actually decide whether you grow. You need a spike there. Everywhere else, you need coverage, not brilliance. Coverage you can buy. Spikes you have to hire and protect.

Why one spike beats five generalists

Run the maths on the five-box chart. Say those five hires cost you, very roughly, $28k-$35k a month all in once you load on the on-costs. For that money you've bought five people who are each operating at maybe 60-70% competence in their patch, because each patch is too broad for anyone to get truly sharp at.

Now picture the other version. One spiky operator on, say, ~$8k-$10k a month who is genuinely world-class at the one thing that's currently capping you. And a partner agency covering the rest as a system, not as five separate part-time brains.

You've spent less, and the one decision that was actually holding back growth is now being made by someone operating at 95%, not 65%.

That's the trade. And the reason it works is something worth saying out loud: return on time spent is a different metric to return on money spent. You can always find a bit more budget. You cannot find more hours in your operator's week. So the question isn't "can we afford another generalist?" It's "where is the one place a sharp person changes the trajectory?"

Everything that isn't that place is a candidate for the agency to absorb.

The "rest of the team" model

This is, honestly, how I think most lean brands should be built, and it's how we tend to slot in alongside the brands we work with.

The brand hires one in-house spike. The agency becomes the rest of the team. Not "the Meta vendor" sitting in a silo, but the strategy, the creative production engine, the testing roadmap, the retention flows, the reporting layer. The connective tissue between the boxes you didn't hire.

The in-house spike owns the thing only an insider can own: the brand's voice, the product roadmap context, the gut-feel on what's on-brand and what isn't. The agency owns the machinery that turns that into spend and revenue.

What you get is a T-shaped team without paying for a T-shaped headcount. One deep vertical (your spike) sitting on top of a wide horizontal (us, or whoever your "rest of the team" is). The founder stops being the glue holding five contractors together, which is usually where their week was quietly disappearing.

Which spike to hire first

Here's my actual playbook on the order. It's not a law, but it's held up across a lot of org charts.

Up to roughly $50k/month in revenue: hire the creative spike.

At this stage, nothing matters more than the rate at which you can produce ad concepts that don't look like ads. Account structure is mostly solved. Email is a few flows. The thing capping you is creative volume and quality, full stop. So your first real hire is a creative operator: someone who can write hooks, brief and direct UGC, and turn out concepts every week. Pair them with an agency running the media and the testing discipline, and you've covered the whole front end with one sharp person.

I'd take one obsessive creative lead over a generalist marketing manager every single time at this size. It's not close.

Roughly $50k to $150k/month: keep the creative spike, and let the agency carry retention while you grow into a second spike.

Now a second problem appears. You're acquiring fine, but your returning customer rate is, say, sitting under 20%, which means you're filling a leaky bucket. The instinct is to hire an "email manager". My push-back: at this stage, retention is more system than spike. Flows, segmentation, a sensible send calendar. That's exactly the kind of horizontal an agency should be running for you, so you don't burn a precious headcount slot on it yet.

So the second hire I'd argue for is whichever spike compounds hardest for your specific brand. For most it's still more creative firepower, because creative is the one input that decays and has to be constantly refed.

Roughly $150k/month and up: now hire the retention or brand spike in-house.

Above this line, retention stops being a set of flows and becomes a genuine profit centre worth owning internally. A returning customer rate climbing from ~18% to ~30% on a brand doing ~$180k a month is a serious amount of margin, and it deserves a sharp in-house person whose whole identity is loyalty, lifecycle and LTV. This is also where a brand spike (someone who owns positioning and the deeper story) starts to earn their seat.

The agency doesn't disappear here. It shifts. We go from "the rest of the team" to "the part of the team that scales the machine while your two spikes go deep". Same logic, more layers.

What this looks like in practice

Picture a homewares brand sitting at ~$60k a month, ~$55 AOV, plateaued. The founder is about to post three job ads: marketing manager, social, email.

What I'd do instead: cancel two of those three. Hire one creative lead who lives and breathes the product and can ship concepts weekly. Hand the media buying, the testing roadmap and the email flows to a partner who runs them as one system. Same monthly outlay, possibly less, but now the one input that was actually capping growth (fresh creative that converts) has a sharp owner instead of being one of five jobs nobody had time to do properly.

That's the whole idea. Stop buying roundness. Buy one spike, rent the rest, and protect that spike's time like it's the scarce resource it actually is. Because it is.

A quiet test before you post the job ad

If you're about to add a box to your org chart, here's a question worth sitting with first: of everything that's currently capping your growth, how much of it is a "we don't have the right sharp person" problem, and how much is a "the system around our existing people is messy" problem?

They look identical from the founder's seat. They are not the same fix, and hiring a generalist to paper over a systems problem is the most common money-leak I see on lean teams.

If you can't tell which one you've got, that's usually the most useful thing to find out before you spend a dollar on either. Map your account, your creative output rate, your retention numbers and your real unit economics against where you actually want to be in twelve months, and the gap tends to point straight at the one spike worth hiring. If you'd value a second read on where that gap sits before you commit a headcount to it, that's exactly the kind of thing a Signal/Noise Audit is built to surface.