--- title: "When to Outsource Marketing and When to Build It In House" url: "https://marketermagazine.co/insight/when-to-outsource-marketing-and-when-to-build-it-in-house/" author: "Nick Baudoin" published: "2026-09-25" updated: "2026-09-25" --- # When to Outsource Marketing and When to Build It In House Outsource the work you can describe. Build the work you're still figuring out. That's the line I wish someone had handed me before I spent about $200,000 giving the front of my business to three vendors: an outbound marketing provider, an ad agency on retainer, and the ad spend itself. None of them did a bad job, which is the part that took me longest to understand. The engagements failed on a distinction that hardly ever comes up in the pitch, which is whether the thing you're buying is execution or judgment. ### Why the spend looked like the responsible call At the time it was the obvious move. We're an agency. Our people were on client work. Building an outbound function meant hiring, tooling and months of learning something adjacent to what we already did well. Buying it meant a provider with a process, a start date and a line item I could forecast. That reasoning is sound for execution. If you can write down what good looks like, someone else can do it, often better and cheaper than you. Media buying, production, list building, the mechanics of a send. All of that travels. It falls apart for judgment. Judgment is the part where somebody decides which claim to lead with for this specific buyer, what a reply actually meant, and which failure is worth fixing first. That work compounds inside the person doing it. Buy it, and the compounding happens in the vendor's building while you rent the output. ### The tell that you bought the wrong one There's a specific symptom, and it's easy to miss because it looks like success. The reporting stays healthy. Sends go out, impressions accumulate, the dashboard is green, and the monthly call has plenty to talk about. Meanwhile nothing downstream moves. No pipeline you can trace back, no change in who's booking, no shift in what buyers say in the first ten minutes of a call. That gap is the diagnosis. When you outsource execution, you read the numbers and know what to do next. When you've outsourced judgment, the numbers describe activity and you have no way to act on them, because the person who could interpret them doesn't work for you. You end up renewing not because it's working but because you can't prove it isn't. ### Price the exit before you price the retainer The question I now ask before signing anything is what's left behind when this ends. Some spending leaves an asset. A documented process, a dataset, a piece of research, a page that keeps ranking, a system somebody on your team can now run. Some spending leaves a gap exactly the shape of the vendor. Both show up in the same row of the budget and they are not the same purchase. That framing also settles most renewal arguments. If twelve months of a retainer has produced nothing your team could carry on Monday without the vendor, you didn't buy capability. You rented activity, and the price of stopping is that you're back where you started, twelve months later, at whatever the market rate is now. ### What I'd do differently Build the first version yourself, badly, before you buy anything. Not forever. Long enough to learn what good looks like in your specific market, which is usually a few weeks rather than a few quarters. I eventually did this with our own list building. I built the pipeline myself first, then wrote the runbook for how someone else should run it. Once the runbook existed the work became genuinely outsourceable, because I could describe what I was buying and recognise it when it arrived. The order matters and I had it backwards. Buying first means you never develop the standard. Without the standard you can't evaluate the vendor. Without an evaluation you keep paying for another month. The by-product turned out to be worth as much as the savings. Writing the runbook forced decisions I'd been making by instinct into rules somebody else could follow, and rules somebody else can follow are also rules you can automate. We now have software handling most of the recurring parts of that pipeline: it watches the numbers, prepares what it thinks should happen next, and a person approves before anything goes out. None of that would have been possible from a vendor relationship, because the knowledge that made it automatable only existed because we'd done the work by hand first. ### What actually travels well Three things have held up as safe to hand off. Anything with a defined output and an obvious quality check: design production against an established system, development against a written spec, media buying inside a strategy you set. Anything you'd do identically for any company in your category. There's no advantage in owning it and no learning in doing it. Anything you've already run yourself and documented. This is the big one, and it's why the sequence matters more than the decision. What doesn't travel is positioning, the first version of any new channel, and the interpretation of results. Those three are where the compounding lives, and they're the three most commonly included in a full-service retainer. ### The version of this that rarely gets said out loud A retainer is easier to defend internally than a hire. It's reversible, it's a smaller number on paper, and if it fails you can say you tried the professionals. A hire is a commitment with a face attached to it. So a lot of outsourcing decisions are really risk-transfer decisions wearing the costume of capability decisions, and they tend to get made at exactly the moment you can least afford to rent your own judgment. I'm not arguing against vendors. We are one. I'm arguing that "we don't have the capacity" and "we don't know how to do this" are different problems with opposite correct answers, and the second one doesn't get solved by a purchase order. ### A cheap test before you sign Ask the vendor to describe the first failure they expect, and what they'd do about it. An execution vendor answers this easily and specifically, because they've run the process enough times to know where it breaks. A vendor selling judgment answers it in the abstract, or reframes it as a question of giving the strategy enough time. The difference shows up inside two minutes and it costs nothing to run. The follow-up is better still: ask what they'd need from you to make the engagement work. If the honest answer is "access and approvals", you're buying execution and the arrangement can work. If the honest answer involves someone on your side owning the message, the market and the interpretation of results, then you're being told, correctly, that the judgment stays with you. That's the moment to decide whether anyone on your team has the time to hold it, because if the answer is no, the retainer will quietly fill that gap and you'll be back here in a year. ### Where this advice stops Two honest limits. If you're pre-product or pre-positioning, none of this applies cleanly, because there isn't a standard to develop yet and the fastest way to find one may well be to watch somebody experienced work. And if the channel is genuinely specialist, regulated media, technical compliance, anything where a mistake is expensive and non-obvious, build-it-yourself-first is bad advice. Buy the expertise and spend your learning budget on being able to brief and audit it instead. ### Common questions **How long should I run something myself before outsourcing it?** Until you can write down what a good result looks like and what the common failure modes are. If you can't write the brief, you can't evaluate the work, and you're about to pay someone to make that decision for you. **What if we genuinely don't have the internal time?** Then buy execution and keep judgment. Hand over production and media buying, keep the decisions about message and market. The split is the point, not the volume. **How do I know it's working before the contract ends?** Agree one downstream number before you sign, something closer to revenue than to activity, and agree what a normal first ninety days looks like on it. If the only numbers on the monthly call are the vendor's own activity metrics, you've already lost the ability to judge the engagement. --- Nick Baudoin is Founder and President of [Alkali](https://www.byalkali.com/), which builds websites and marketing programs for established B2B companies. Alkali's study of [more than 55,000 US B2B websites](https://www.byalkali.com/reports/state-of-b2b-websites/) looks at what first-time buyers actually run into when they arrive.