Managing Amazon PPC In-House or Through an Agency: How to Decide

Most Amazon sellers I speak to have already made this decision before they've really thought it through. They either hired an agency because it felt like the grown-up move, or they kept PPC in-house because handing it over felt like giving up control. Neither of those is a good reason. And both routes have a specific way of going wrong.

So let me try to give you a clearer framework than "it depends", because I'm not sure that answer actually helps anyone. And I'll tell you what happened to us, because it's probably the most useful thing I can share.

What in-house management actually requires

Managing Amazon PPC yourself, or with someone on your team, sounds straightforward. You have access to Seller Central. You can see the numbers. How hard can it be?

Harder than it looks, in my experience, and not for the reasons most people expect. The technical side, understanding match types, structuring campaigns, reading search term reports, is learnable. It takes time, but it's learnable. The harder thing is the judgement layer on top of that.

Good PPC management on Amazon is not really about the bids. It's about knowing when to spend aggressively and when to pull back, which is a decision that has to be made in the context of your margin, your inventory position, and where you are in the season. If your stock is running low and your lead time is eight weeks, scaling spend right now could leave you winning the Buy Box on a listing that goes out of stock before the orders convert properly. That's a mess. If you've just landed a new shipment ahead of peak and your organic ranking is soft, that's exactly when to push. The data in Seller Central doesn't tell you any of that. You have to bring it yourself.

So the real question for in-house management is not "can someone log in and adjust bids?" It's whether that person has enough context about the business, the margins, the stock situation and the competitive position to make the right call on a Tuesday afternoon when the numbers look a bit odd. That's a high bar, and a lot of businesses are running PPC in-house with someone who can operate the tool but not quite reach that bar. The spend is going out. The ACOS looks acceptable. And the contribution margin is quietly getting squeezed.

What agencies actually offer, and where the model breaks down

A good agency brings real expertise: someone who has managed a lot of accounts across a lot of categories, has seen the patterns, knows the platform shifts early, and can run a proper campaign architecture that most founders would take months to build from scratch. That's genuinely valuable, and I don't want to pretend otherwise.

The problem is structural, and it affects most agencies even decent ones. Your account is one of many. The senior person who sold you the retainer is probably not the person managing your campaigns day to day. The person who is managing them may be good, but they're working off your ACOS target and your ad budget, without a full picture of your unit economics, your landed cost, or what's actually happening with your inventory. They're optimising the account. They're not running the business.

For a brand turning over several million pounds with a complex catalogue and serious ad spend, a specialist agency can absolutely be worth it. The account is big enough to warrant real attention, the fee is proportionate, and there's enough going on that dedicated expertise pays for itself. For a smaller brand, a £30,000-a-year agency retainer on £400,000 in revenue is a lot of money to spend on someone whose senior staff will rarely, if ever, look at your account.

What we actually tried, and what it cost us

Here's the thing I keep coming back to when I think about this: Martian Made is a sleep brand. A consumer products business. We're not a technology software company.

That sounds obvious when I say it out loud. But for a period, we didn't act like it was obvious. Around 2025, PPC costs on Amazon started shifting noticeably. Amazon was squeezing harder on ad spend, costs were rising, and it was well documented across seller forums and communities. By late 2025 and into early 2026, I could feel it clearly in our own numbers, and I was hearing the same from other sellers. The platform was getting more expensive and more complicated to manage.

So we decided to build our own in-house PPC management system. We spent months building it. Then we spent months running it. In total, it took six months of a downward spiral in PPC performance and an inability to keep TACOS under control before we accepted what should have been obvious from the start.

We were never going to build something as good as a dedicated software company. Not because we weren't trying, but because that's not what we are. A software company has engineers whose entire job is to build and refine those systems. We have a lean team whose job is to run a product business. We could not out-engineer people whose whole existence is engineering those tools. Six months to figure that out. Real money, real time, and a period of making a loss to learn something that, in hindsight, we should have seen coming.

On top of that, Amazon has been expanding into AMC and DSP, and for DSP specifically, you need an agency with a dedicated seat to access it. You cannot get there in-house. So if you're scaling and you want access to those fuller, more top-of-funnel solutions, you will eventually have to go through an agency anyway. That's not something I'd really internalised until we were already deep in the problem.

Which option suits whom

If your ad spend is still modest, your catalogue is manageable, and you or someone on your team can genuinely give PPC the attention it needs, not just access to the tool but real thinking time connected to real business context, keep it in-house. In the early days, Amazon Ads Console is enough to get started. Learn it properly. The understanding you build will be worth more than the efficiency you'd gain from outsourcing.

But know where that ceiling is. Once managing PPC starts becoming a chore, once it's getting twenty minutes on a Friday afternoon because everything else is pulling at you, that's the signal. Either outsource it to an agency that will put someone serious on your account, or hire a dedicated person in-house. Both have their drawbacks, as I've covered. But drifting without proper attention is worse than either.

And the thing I didn't fully appreciate until we'd burned through six months of it: when you work with a good agency, you're not just buying software. You're buying a team. Your point of contact has colleagues. They can ask questions of people who've seen the same problem fifty times across fifty different accounts. One founder's understanding of PPC, however good, is not the same as that depth of institutional knowledge. And as the platform gets more complicated, that gap only widens.

My position now is that I should have moved to an agency sooner. That's not me telling you to do the same at whatever stage you're at, because context matters. But if you're at the point where PPC is genuinely complex and genuinely important to your revenue, trying to out-engineer a dedicated software company with a spreadsheet and good intentions is probably not the answer.

The mistake I've seen from in-house management is usually slow drift: small inefficiencies that compound quietly over months. The mistake from agencies is usually misalignment: the account is being optimised for the wrong thing, because the agency doesn't know what you actually need it to do. Different failures, same result.

What's your PPC actually optimised for right now?