What to Outsource First in an eCommerce Business, and What to Keep

Most eCommerce businesses reach a point where the founder or the small team is doing work that keeps the lights on instead of work that grows the business. Answering the same customer questions. Fixing a listing that was suppressed overnight. Reconciling a shipment. None of it is optional, and none of it needs the person who decides what the business sells.

The question is rarely whether to outsource. It is what to hand over first, and what never to hand over. This is a test for telling the two apart, the order that usually makes sense, and how to hand work over without losing control of the accounts it touches.

Ask three questions of each task. Is it done the same way most of the time? Can you tell within a week whether it was done well? Does a platform put a clock on it? A task that answers yes to all three is the best first candidate, because it can be written down, checked, and it costs the business something every time it is late.

A task that answers no to the first two is a poor first candidate however much time it takes. Deciding which product to launch, how to price against a new competitor or what the brand should say about itself depends on judgement and context that live in the business. That work can be supported, researched and prepared by somebody else, but the decision stays where the consequences land.

Customer service passes the test more clearly than anything else. Most messages are the same dozen questions, the answers can be written once and kept current, and the platforms measure it. On Amazon, sellers describe a 24 hour expectation for answering buyer messages, weekends and holidays included; Amazon's own help page on responding to messages sits behind the Seller Central sign in, so check the current rule there. A message left over a long weekend is how a question becomes a claim.

The other marketplaces measure the same service through different numbers. Walmart ties its seller tiers to delivery and cancellations, set out in Walmart's Pro Seller tiers, and eBay's are in the Top Rated Seller requirements. A team that covers messages across time zones, working from written answers, is the most direct way most businesses protect those numbers.

Creating listings from a template, keeping attributes complete, fixing suppressed listings, updating images and prices across marketplaces: this is repeatable work with a checkable result. A listing is either live and complete or it is not, and the suppression count tells you in a day whether the work is being done.

Keep the standard in house and hand over the upkeep. The style guide, the words that describe the brand, the claims the products can and cannot make, are decisions. Applying them to the two hundredth listing is not. The work that builds a listing in the first place, the research and the writing, is a different job again, which is what product listing optimisation covers.

Shipment creation, receiving discrepancies, removal orders, returns processing and the checks that find money a marketplace owes. It is detailed, it is repeatable and it has deadlines, some of which cost money when they pass. FBA reimbursements shows how much of it runs on dates: a claim that is not made in its window is lost.

This is also where a written process pays for itself fastest, because the work is spread across several screens and a few people, and a step that exists only in one person's head is the one missed when they are away.

Keep the decisions: what to sell, what to charge, what to spend on advertising and what the brand says. Keep the relationships that the business depends on, with its suppliers and its biggest customers. And keep ownership of the accounts. The primary login to a marketplace account, the payment details and the ability to add and remove users belong to the business, never to whoever does the work.

Give every person who works on an account their own login with the permissions the task needs and no more, using the marketplace's own user permissions. A shared password cannot be taken back from one person without being changed for everyone, and a single set of credentials in several hands is also a single point of failure.

Write the playbook before the handover, not during it: the tasks, the answers, the standards, the escalation route for anything unusual. The first version will be wrong in places, and the handover is how it gets corrected.

Agree the measures first. For messages, the response time and the share resolved without escalation; for listings, the errors and the suppressions; for administration, the backlog and the deadlines met. Read them weekly for the first month, then monthly. Run the old way and the new side by side for the first week or two, so a gap shows up as a comparison rather than as a customer complaint.

Start narrow. One task on one marketplace, handed over completely, teaches more than five tasks handed over halfway. When the measures hold for a month, add the next task, and let the playbook grow with it. A handover that stalls is almost always one where the business kept doing part of the work itself and nobody could tell which part was whose.

The model matters as much as the list. A ticket queue answers what it is sent; a team that knows the account notices what nobody sent. Our BPO services are built on the second: "Marketplace specialists, advertising specialists, catalogue and customer support staff who work as part of your team rather than as a ticket queue. The same people, every day, on your systems." Which tasks go first is a decision we make with the business, on the test above.