Agency, freelancer, or software for small businesses
Agency, freelancer, or software for small businesses
A small business owner wrote this to me a few weeks ago: "I've already tried an agency and a freelancer. With both of them I ended up starting from zero again."
That describes the problem better than any comparison table. Almost every comparison between agency, freelancer, and software talks to you about price and scope. None of them tells you what happens eight months in, when that person is gone.
And they are gone. Freelancers take a full-time job. Agencies rotate the account executive without telling you. That's when you find out what you actually bought.
What you're actually buying in each case
The three options are sold as if they were the same thing at different prices. They aren't.
| Agency | Freelancer | Software + data pilot | |
|---|---|---|---|
| Monthly cost (CLP) | 500,000 – 3,000,000+ in fees | 250,000 – 500,000 | Low or none to diagnose |
| The same in dollars | US$550 – 3,280+ | US$275 – 550 | — |
| What you get | Multichannel execution, creative, reporting | Execution on 1 or 2 channels | A read of your numbers against your sector |
| Who decides | They do, you approve | They do, you approve | You do, with judgment |
| Ad spend that justifies the cost | Above 1,500,000 a month (US$1,640) | 500,000 – 1,500,000 (US$550 – 1,640) | Any |
| What happens if they leave | A new executive comes in, the context resets | It all stops | Doesn't apply |
| Main risk | Judge and defendant | A single person | You have to make the calls |
The fee ranges come from the same analysis I did in what a marketing agency charges. The rule that matters: your ad spend should be 3 to 5 times the management fee. If you pay a 400,000 fee (about US$440), this only adds up if you're putting 1,200,000 or more (US$1,310) into ads. Below that, you're paying an enormous percentage purely for administration.
The question nobody asks
When you compare these three options, everyone pushes you to ask what it costs and what's included. The question that decides the outcome two years out is a different one:
Where does the knowledge of your marketing live when that person leaves?
With an agency, the knowledge lives in their team. They know which audiences they tested, which creative worked in March, why they turned off that campaign. When they rotate the executive — and they do — that context is lost for them too. You explain your business from scratch again, now paying the same fee.
With a freelancer the risk is simpler and more serious: it's all in one head. If they land a stable job, your marketing goes dark that same day.
With software the knowledge stays in your accounts and in your diagnostic. It's the only one of the three where what you learned is still yours at the end.
Neither of the first two paths is a mistake. What's expensive is signing them without having asked this first, and without putting in writing who owns the accounts. How to negotiate that is something I covered in the real alternatives to an agency.
When each one makes sense
Each has its band and none of them wins everywhere.
An agency makes sense when you invest above 1,500,000 CLP a month in ads (about US$1,640), you need creative production at scale, you operate in several countries, or you have complex integrations. At that volume the fee is diluted against the return. Hire them, but knowing how you're going to measure them from the outside.
A freelancer makes sense in the 500,000 to 1,500,000 ad spend band (US$550 to 1,640), when the problem is execution and not strategy, and when you're clear about what you'll be asking for. It costs less and tends to be faster to coordinate than a small agency. Accept the risk that it's one person, and have a plan B.
Software and running it yourself makes sense below 500,000 in ad spend (US$550), or in any band when what you're missing isn't someone to push the buttons but knowing whether what's already running is any good. If you've never touched an account, how to run your Google Ads without an agency covers it end to end.
And all three fall apart at the same point if you don't know how to evaluate them. That part you can't outsource.
What to ask any of the three before signing
This applies equally whether you hire a big agency, a recommended freelancer, or end up doing it yourself. It's the part that gets discussed least and costs the most.
1. Whose name are the accounts under? The Google Ads account, the Meta Business Manager, the pixel, the domain. If they're in the provider's name, the history isn't yours. You negotiate this on the way in, never on the way out.
2. What number are you going to show me, and what do I compare it against? If the answer is "ROAS," ask what a good ROAS is in your sector. If there's no answer, you don't have an evaluation: you have a number floating in the air.
3. Who builds the report? If whoever executes is also whoever measures, they have an incentive to justify the budget and show what looks good. Not because they're dishonest — because that's how incentives work. It's the same thing that happens when the teacher grades their own exam.
4. What happens the day this ends? Ask them to tell you before you sign. How many days of handover, what documentation is left behind, who holds the access.
5. What money decision will I make differently with this information? If a report doesn't change a single decision, it's decoration. Expensive decoration.
Why I keep pressing on this
I had a consultancy that went under. Four years, good service, happy clients.
The lesson I took from it wasn't about partners or timing. It was that I had built a relationship where the client depended on us to understand their own business. While we were there, it worked. When we stopped being there, the client was exactly where they had started, with four years of money invested and nothing of their own to show.
That's why I find it hard to write a comparison that ends in "hire this one." The important decision isn't who you hire. It's how much you understand by the end of the contract.
Before you decide: know where you stand
Choosing between agency, freelancer, or software without knowing how your numbers look today is choosing blind. You'll end up deciding on price, which is the worst criterion available.
You can scan your site free: in 30 seconds it detects what you're measuring, whether your KPIs are on target, and how you compare to your sector. Nothing to install. It's the baseline you need before sitting down to negotiate with any of the three — and it also works for evaluating the one you already have.
Conclusion
All three options are valid and all three have their band. An agency above 1,500,000 in ad spend (US$1,640), a freelancer in the middle zone, software and your own hands when the budget is small or when what's missing is judgment rather than execution.
What doesn't change is the question. When this ends, what's left that's yours?
Frequently asked questions
What's cheaper, an agency or a freelancer?
A freelancer, almost always. A specialized freelancer's fee in Chile runs from 250,000 to 500,000 CLP a month (US$275 to 550), against 500,000 to 1,000,000 for a boutique agency (US$550 to 1,095) and 1,000,000 to 3,000,000 or more for a mid-size agency (US$1,095 to 3,280). But price alone doesn't decide: a freelancer covers one or two channels and is a single person, while an agency covers more fronts with a team behind it. The honest comparison is between what each one solves, not between the two figures.
How much ad spend do I need for an agency to make sense?
Above 1,500,000 CLP a month in advertising investment, roughly US$1,640. The management fee should be a smaller fraction of what you invest in ads, not a third of your budget. Below that floor the fee takes a disproportionate slice of your total budget, and the return doesn't cover it.
Can software replace an agency?
Not in day-to-day execution — someone has to build campaigns, write ads, and upload creative. Yes in diagnostics, which is where small businesses are most blind: knowing what's being measured, whether the KPIs are on target against the sector, and where it makes sense to move the budget. In practice the most common setup is to combine them: software so you have your own judgment, and someone — you, a freelancer, or an agency — to execute.
How do I know if my current agency is doing well?
You need an external point of comparison. Take the number they show you and check it against your sector's average, not against your own previous month. If nobody can tell you what "good" is in your industry, you don't have an evaluation: you have a report. That gap is what an independent diagnostic fills.
What should I ask for before signing with anyone?
That the Google Ads and Meta accounts stay in your name, with the provider as an invited user. That it's in writing how many days of handover there are at the end and what documentation they hand you. And that they tell you upfront which metric they'll report and what reference it's compared against.