Agency or In-House? The Honest Decision Tree (From an Agency That Tells Clients When to Leave)

Picture two versions of the same brand, both at roughly A$4m a year.

Brand A hands the whole account to an agency. Media buying, creative, landing pages, the lot. Eighteen months later they want to bring a piece in-house and realise nobody on their own team actually knows how the account works. They're starting from zero, and the agency owns all the muscle memory.

Brand B brings on an agency too, but tells them on day one: "We want to learn this, and in a year or so we plan to run media buying ourselves." The agency builds the program, the brand watches closely, and when the handover comes it's clean. Same spend, same agency, completely different outcome.

That gap is the whole question. Not "agency or in-house" as a yes/no, but which functions you own, which you rent, and how you sequence the move between them.

I run an agency. We're a Meta-ads-led growth shop for Shopify brands. And I'll tell clients to internalise functions and walk, because I think the brands that plan for it are better clients while we have them. So here's the honest version of how I'd decide, stage by stage.

Start with the core-competency test

Before you think about budgets or retainers, answer one question: what is the thing your brand has to be world-class at to keep growing?

For most DTC brands I work with, that's the creative-and-acquisition flywheel. The ads, the angles, the landing pages they point to. If you're a paid-social-driven brand, that engine is your business. It's not an offshoot of an agency. It's the reason you exist.

Here's my take on the split:

• Own what's core. If acquisition is how you live or die, you should eventually own media buying, have some level of creative production happening in-house, and be able to design and test landing pages without raising a ticket with an external partner. Not all of it, not on day one, but the spine of it.
• Rent what's specialised or seasonal. Email and SMS retention, Amazon, a TikTok Shop launch, a Super Bowl-style TV buy - these are great things to bring an agency in for. They're either not your core skill set, or the channel isn't big enough yet to justify a full internal team.

The test isn't "is this important". Retention is important. The test is "is this the thing customers fall in love with us for, and the thing that breaks if it's a black box to us". If yes, own it over time. If no, a specialist will almost always do it better than a junior internal hire you stretched to afford.

A quick gut check on the rent side. There's nothing wrong with deciding email isn't a core competency and handing it to a retention agency. You don't need to own everything. You need to own the things that, if they went dark tomorrow, would sink you.

The early stage - rent almost everything, but learn loudly

If you're under roughly A$2m a year, you probably can't afford to build a real internal team, and the best people already have jobs. That's just true. So an agency is the smart move - not as a way to offload and forget, but as a way to buy a working program fast.

Two things matter here.

First, treat the agency like an internal employee, not a vending machine. Give them every bit of data, every learning, every bad review and angry-customer screenshot. The brands that get nothing from agencies are usually the ones who filled out the onboarding form in five minutes and went quiet. To put this in perspective: a creative team that's six levels deep in your competitor's reviews and Reddit threads will outperform one working off a thin brief every single time. You control how good that brief is.

Second, learn out loud. I'd tell the agency from the start what you're hoping to internalise later. The right partner says "cool" and takes you on the journey. That's not shady. That's just extracting the value you're paying for. A good agency knows the lifecycle and would rather have an honest client for a year than a resentful one who churns and won't say why.

What I'd recommend at this stage: pick one channel to start understanding deeply yourself, even if the agency runs it. Sit in the meetings. Ask why. You're not trying to take it over yet. You're building the base layer so that when you do, it's a handover and not a cold start.

The growth stage - internalise the core, keep specialists around it

Somewhere in the A$2m to A$10m range, the maths flips on your core function. You're doing enough media buying, building enough creative, testing enough landing pages, that owning it starts to make sense.

This is where I'd start moving the spine in-house. A first media buyer. A creative strategist running a small internal team. The signal is volume: when you're doing the same thing over and over, week after week, that's when an internal owner beats a retainer.

But - and this is a big one - internalising the core doesn't mean firing everyone. The best-run brands I see at this stage have a strong in-house team with specialist agencies around it, not instead of it. You always need to outsource some creative. You cannot generate enough variety internally on your own, and outside eyes bring angles your team would never reach.

Here's the thing about how relationships end at this stage. An agency being a great fit at one moment and a worse fit a year later isn't a failure. It usually means the brand built a better competency. That's a good thing. It's not a knock on anyone.

The way to do it well is transparency, set early. When you bring a creative agency on knowing you'll build the flywheel internally in a year, say so. "We love working with you, we want to learn from you, here's how we see this evolving - we'll own these parts, you'll keep these." Not being upfront isn't kind. It feels easier in the moment, but it's a bit selfish. The honest version is the generous one. It lets everyone win for the time you're together, and it means the people grinding on your account aren't blindsided.

The scale stage - own the engine, but watch for the insulation problem

Past A$10m or so, you've probably internalised the functions that define you. Media buying in-house. A real creative team. CRO and web owned, because the website is too central to keep raising tickets for.

And now you hit a problem almost nobody warns you about. The longer your team sits inside one brand, the less it knows about the outside world.

I've watched in-house leaders go six months without touching a new ad account or a new tool and quietly fall behind on things they used to be the expert on. The best part of a good agency is that it sees a lot of brands. It has data and patterns you simply cannot get from inside one company. When you bring everything in-house, you lose that window.

My thought on solving it: keep a thread to the outside open on purpose. That might be a senior operator on a light retainer who pops into the account once a month - not executing, just advising, just bringing fresh eyes. It might be connecting your team directly to teams at non-competing brands so your media buyer is comparing notes with someone else's media buyer. It might be a paid community. The point is you've decided to internalise, and you've decided not to go blind doing it.

This is also the stage where I'd lean hard on performance-aligned partners over pure retainers. A retainer with no skin in the game and a fully internal team is a recipe for a comfortable, insulated account. An advisor or specialist who only keeps the work by earning it keeps everyone honest.

What this actually means for you

Run the decision in this order:

1. Name your core competency. The one engine that has to be world-class. For most DTC brands, that's creative and acquisition.
2. Match the stage. Under A$2m, rent almost everything but learn loudly. A$2m to A$10m, start internalising the core while keeping specialists around it. Past A$10m, own the engine and deliberately protect against insulation.
3. Sequence the handovers and be transparent about them. Set the in-housing plan with your agency from day one. The good ones respect it. The ones who don't were never the right partner.

One more thing, because it's the most expensive mistake I see. Don't run a multi-million-dollar acquisition business with the engine fully outsourced and no internal knowledge of how it works. If that agency gets bought, raises its minimum, or just decides to give you thirty days' notice, you've got nothing to stand on. The opportunity cost of never building any in-house muscle is the scariest line on the page, even if it never shows up in a spreadsheet.

If you're staring at your own account and you genuinely can't tell which functions you should own next versus keep renting, that's usually a sign the picture's gone fuzzy. A Signal/Noise Audit is built to make exactly that call clear - we map what's core to your growth, what's actually working in the account, and which functions are ready to come in-house versus the ones better left to a specialist. No pressure to work together after. The point is just to hand you the decision tree filled in for your numbers, not a generic one.

So which function would you bring in-house first if you had to choose this quarter, and do you actually have the base layer to do it cleanly?