Home › Resources › Insights › Operations › Outsourcing eCommerce management
We are an agency, so treat the following with appropriate suspicion. It is still true: for a good number of businesses, hiring is the better answer, and for a smaller number neither hiring nor outsourcing will fix what is actually wrong.
The Dazzle Commerce Team
eCommerce Management & Marketing
Updated
5 September 2026
Reading time
8 min read
The question is rarely in-house versus agency. It is which parts of the job need which.
The outsourcing decision is usually framed as a cost comparison: a salary against a retainer. That framing produces bad decisions in both directions, because it compares two prices without comparing what you actually get for them.
A better framing is about the shape of the work. Some parts of running an eCommerce business reward deep, permanent knowledge of your customers and your catalogue. Other parts reward having seen forty accounts break in the same way. The first belongs inside. The second usually does not.
KEY TAKEAWAY
Keep the work that compounds with knowledge of your business. Buy the work that compounds with exposure to many businesses.
1
Is this a permanent function or a project? Merchandising and customer service are permanent. A replatform, a market entry or a marketplace launch has a beginning and an end, and hiring for it leaves you with a team you do not need in eighteen months.
2
Can you interview well for it? Most founders can assess a merchandiser. Very few can tell a good paid media hire from a confident one, which is precisely how expensive mis-hires happen in that seat.
3
What happens when this person is away? A single in-house specialist is a single point of failure. If a fortnight of annual leave stops the work, you have bought a person rather than a capability.
The decision that works is usually a split, not a swap: keep the knowledge that is specific to you, buy the knowledge that is specific to the discipline.
Three situations where changing who does the work is treating a symptom. In each case the same problem will reappear with a different team attached to it.
1
The unit economics do not work. If contribution margin after returns, carriage and fees is thin or negative, no amount of marketing skill fixes it. Pricing, product mix or cost base has to change first.
2
Nobody can decide. If every change needs three approvals across two departments, an agency simply joins the queue. The constraint is the decision process, and it is not one an external team can remove.
3
The last three attempts failed the same way. Two agencies and a hire, all ending at the same point, is a signal about the brief, the data or the internal ownership, not about the suppliers.
to start a hire
And roughly six before they are fully effective
most brands need part-time
The gap that makes a blended team economic
is about the brief
Not about the third supplier
Rough planning figures we use in these conversations, not research findings.
Almost every arrangement we see working well is a split, and the split falls in roughly the same place each time.
Function
Usually belongs
Why
Merchandising and range
In-house
Compounds with knowledge of your customers and margins
Customer service
In-house
It is your brand voice and your best source of product insight
Paid media
External
Pattern recognition across accounts beats time in one
SEO and content
Mixed
External strategy, in-house subject knowledge
Lifecycle and retention
Mixed
External build, in-house calendar ownership
Marketplace operations
External
Policy churn is a full-time reading job on its own
Development
External
Peaky demand that almost never fills a permanent role
The exception is scale. Past a certain size every one of these moves in-house, and the useful question becomes which ones move first.
EXPERT VIEW
“The engagements that fail are almost always the ones where the client wanted to hand over a problem rather than a function. Nobody can own an outcome they do not control the inputs to.”
Dazzle Commerce · Client services
If it turns out the work is genuinely a permanent, full-time function, we will say so. Our management engagements are built around the split above rather than around taking everything.
Whichever way the decision goes, the same five things determine whether it succeeds:
1
Name the constraint, not the deliverable. “Our repeat rate is half the category norm” is a brief. “We need more emails” is an order.
2
Give access on day one. An engagement that spends three weeks waiting for analytics access has spent a fifth of its first quarter.
3
Agree one number. Not a dashboard — one number that decides whether this worked, and who owns it.
4
Decide who says no. External teams need one internal person who can settle a disagreement inside a day.
5
Set the review date before you start, and hold it whether the news is good or bad.
BEFORE YOU DO EITHER
If you cannot write the constraint down in one sentence, do not hire and do not appoint. Run the diagnosis first — whoever does the work next will need that sentence anyway.
What an engagement with us actually looks like, priced and described plainly.
The engagement model, what we own, what stays with you, and how the split is agreed.
Engagement models and indicative ranges, so the salary comparison can be made honestly.
Who does the work, how the team is structured, and the markets we operate across.
WRITTEN BY
eCommerce Management & Marketing
Dazzle Commerce runs eCommerce management for brands that have decided the work sits better outside the building. We turn down engagements where it clearly does not, which is where most of the judgements below come from.
More from the operations side of the practice.
The diagnosis that produces the one-sentence constraint this decision needs.
How to tell whether the reporting you are being shown supports a decision.
A worked example of the kind of project that is almost never worth hiring for.
A short conversation usually settles it. If the answer is that you should hire rather than outsource, we will tell you that.
