In-House vs Agency vs Consultant: What a Store Actually Needs
For a direct‑to‑consumer store spending between $5,000 and $50,000 a month on ads, a single experienced consultant is usually the right answer, and it is usually right for an unglamorous reason: at that size the work is roughly one competent person’s attention, and both of the other options sell you something other than that. An agency sells you a team, most of which never touches your account. A hire sells you all of somebody’s time, including the hours you do not need. Above roughly $100,000 a month the answer changes, and I will say where.
That is the verdict. Now the qualification, because the sentence above is true for a band of businesses rather than all of them, and the three options each genuinely win something.
On this page
- Which should you choose: in-house, an agency, or a consultant?
- What does each one actually cost?
- The comparison, on the criteria you decide on
- Why is every other page on this question a binary?
- When is an agency clearly the right answer?
- When is in-house clearly the right answer?
- What flips the answer
Which should you choose: in-house, an agency, or a consultant?
It depends on one number more than any other: how much of a competent specialist’s week your marketing actually needs.
A store at $20,000 a month in ad spend running Google, Meta and a feed needs somebody looking properly a few times a week and thinking hard once a month. That is not a full time job and it is not a five person team. It is the shape a consultant fits, and it is why the middle option keeps winning in this band despite almost nobody writing about it.
Two things change the answer. If the work stops being one specialism — paid media plus email plus creative production plus a content calendar plus a site rebuild, all at once — you need capacity, and capacity is what an agency sells. If marketing becomes the core competitive asset rather than a channel, you want the knowledge inside the building, and that is a hire.
What does each one actually cost?
Here are real numbers, which is more than any page currently ranking for this question offers.
- In‑house. A salary, plus payroll taxes, benefits, software and management time. Whatever you budget for the salary, the loaded cost is meaningfully higher, and you are also buying a ramp period during which output is low. One person, one skill set.
- Agency. Usually a percentage of ad spend, commonly 10% to 20%, sometimes against a monthly minimum. At $30,000 a month in spend, 15% is $4,500 a month, or $54,000 a year — do that arithmetic against your own spend before the first call, because the percentage disguises the annual number.
- Consultant. Typically a flat fee. Mine are published: $1,500 a month for one channel, $4,000 a month for every channel with strategy attached, from $7,000 a month for the whole marketing function including email, SEO and managing other contractors.
The percentage‑of‑spend model is the one to interrogate. It means your fee rises in the months you should be pulling back, and nobody on that model has ever been paid to recommend a smaller budget. It is not dishonest, it is just an incentive, and you should know which way it points before you sign.
The comparison, on the criteria you decide on
Every option wins a row. If a comparison hands one side every row, somebody is selling.
| Criterion | In-house | Agency | Consultant |
|---|---|---|---|
| Cost at $30k/mo ad spend | A loaded salary plus tools, fixed whether the work needs it or not | Scales with spend — at 15% that is $4,500 a month | Flat. $4,000 a month for every channel, unchanged if spend doubles |
| Breadth of skills | One person, one or two specialisms | Widest. Paid, creative, email, design, dev, under one contract | Deep in a few things, honest about the rest, subcontracts the gaps |
| Who does the work | You know exactly who. They sit with you | Sold by a senior, often run by a junior. Ask who touches the account daily | The person you met. No handover, no account manager layer |
| Speed to start | Slowest — hiring plus notice plus ramp, realistically a quarter | Fast, though onboarding is a process | Fastest. Days, and it can end in days |
| Capacity when volume spikes | Fixed. A launch means overtime or nothing gets done | Elastic. This is the real product an agency sells | Fixed, and honestly the weak point. One person has one calendar |
| Institutional knowledge | Stays in the building and compounds | Leaves when the contract ends, and when their staff turn over | Leaves when the engagement ends, though the account and documentation stay yours |
| Commercial context | Knows the margins, the stock, the returns problem | Sees the ad account. Often never sees a cost of goods | Sees it if you show it, which you should |
| Risk if it goes wrong | Highest. Firing an employee is slow, costly and personal | Contractual, often with a notice period or a minimum term | Lowest. A short commitment and the accounts are already yours |
| Accountability | Diffuse. Performance reviews are not a reporting line for ROAS | Formal. Contracts, reporting cadence, someone to escalate to | Total and undiluted, which cuts both ways — one person, one point of failure |
Why is every other page on this question a binary?
Because most of them are written for a marketer choosing a job, not a founder choosing where to put a budget. That is not a guess — I read them.
The page holding the top organic position is a Reddit thread titled around why anyone would stay at an agency, which is a career question. The highest‑ranking article is published by a recruitment firm, and its sections are the pros and cons of working in‑house against working at an agency. Further down the page sit a graduate careers guide and a coding bootcamp’s blog. Half the results for a query that reads like a buying decision are answering a hiring decision from the other side of the desk.
The genuinely buyer‑facing pages share two omissions. Not one of the three top results names a freelancer, consultant or fractional marketer as an option at all — the comparison is presented as strictly two‑sided. And not one of them puts a number on either side. You can read all three end to end and come away unable to say what either choice costs, which is the only thing that makes it a decision rather than an essay.
Ask any comparison two questions: what does each option cost, and at what size does the answer change? A comparison that answers neither is describing a topic rather than helping you choose.
When is an agency clearly the right answer?
When you need capacity rather than judgment, and you need it across more than two disciplines at once.
The clearest case is a rebrand or a replatform: new creative in volume, a new site, a migration, paid and organic both moving, all inside a quarter. That is genuinely four people’s work happening simultaneously, and there is no version of one person doing it. Hiring four specialists for a project that ends in twelve weeks is worse than paying an agency for twelve weeks.
The second case is scale. Above roughly $100,000 a month in ad spend, the number of simultaneous tests, creative variants and market splits stops fitting into one person’s week, and the percentage fee that looked expensive at $20,000 starts buying you real staffing. The third is when you need somebody to be accountable to a board in a formal way, with a reporting structure behind them.
If any of those describes you, hire the agency, and interrogate one thing on the call: who touches the account daily, and what else are they on. The gap between who sells the work and who does it is the single largest source of disappointment in this category.
When is in-house clearly the right answer?
When marketing is the product rather than a channel, and when the knowledge has to compound inside the company.
If your differentiation is a brand and a community rather than a catalog, the person building it cannot be renting their understanding of it. The same goes for anything requiring constant proximity to merchandising, stock and customer service — a consultant who sees the ad account weekly will always be slower to know that a hero product went out of stock than somebody sitting twenty feet from the person who ordered it.
The honest cost is the ramp. A hire is a quarter before real output and a year before compounding, and if the channel is already losing money you probably do not have a year. A sensible sequence is to bring in a consultant to fix and document the account while you hire, then hand over something that works to somebody who will own it. That is a better outcome than an engagement which quietly renews forever, and it is worth saying out loud at the start so both sides are planning for the same ending.
The third option is a middle: hire a capable generalist in‑house and buy the specialisms. That is what the fractional CMO arrangement is for — senior direction and a plan, executed by your own team and whichever contractors the plan needs.
What flips the answer
Monthly ad spend, more than anything else. Under $5,000 a month, do it yourself; no fee structure survives being that large a share of the budget. Between $5,000 and $50,000, a consultant. Above $100,000, or whenever the work stops being one specialism and starts being four at once, an agency or a team.
Two secondary triggers override the spend number. If you already employ somebody capable and the problem is direction rather than hands, you want senior input and not another supplier. And if marketing is your competitive moat rather than a distribution cost, build it inside, whatever it costs and however long it takes.
A last practical point that applies to all three. Own your ad accounts, your analytics property and your tag manager container, in your own billing, before anyone else touches them. The cost of choosing wrong is small when you can end an arrangement and still be advertising the next morning. Most of what I find in a first look at a store’s marketing is the residue of an arrangement that ended badly, and it is always worse when the previous holder owned the account.
Questions people ask
What are the drawbacks of in-house marketing?
Is a marketing agency cheaper than hiring in-house?
What is a fractional CMO and how is it different from an agency?
Can one consultant really replace a marketing agency?
Should I hire an agency or a freelancer for ecommerce marketing?
Want this run properly?
I am Greg Asuncion, an ecommerce paid media consultant. I run Google Ads, Meta, Amazon and Microsoft for direct‑to‑consumer brands on Shopify — one person, published pricing, and a free audit after a call. Here is how the fractional arrangement works.
