Online retail has an unusually clean line between outsourceable and non-negotiable work. Most of it is not strategic at all: it is the repetitive operational load that grows faster than the revenue behind it.

The exception is brand. Everything that shapes how your store looks, sounds and prices is worth holding onto — not because it cannot be delegated, but because it is what a customer is actually choosing between you and the next tab.

Strong candidates

TaskWhy it outsources wellWhat you must still own
Bulk product uploadsRule-based, template-driven, volume-elasticBrand voice in descriptions; final copy approval
Catalogue maintenanceRepetitive; mistakes are visible and correctableCategory structure, naming conventions
Marketplace listingsFollows platform rules more than tastePrice floors, positioning, what you claim
Inventory updatesMechanical, and errors are measurableReorder thresholds, stock accuracy standards
Review monitoring and triageVolume and pattern-matching, not judgementHow you reply, what you concede
Order exception handlingDocumented rules cover most casesRefund limits; anything customer-facing and unusual
Report assemblySame inputs, same outputs, every weekWhich numbers you look at and why

Keep these close

  • Pricing and margin decisions — they encode strategy and change with context a supplier will not have.
  • Product selection and ranges — the commercial judgement about what you sell.
  • Brand copy and imagery direction — reviewable by someone else, not substitutable.
  • Customer escalation and complaints — the moments that decide whether someone orders again.
  • Supplier and platform relationships — negotiations and disputes.

The two things that make it work

A listing standard you can test against

Before handing over catalogue work, write the standard: title format, image requirements, attribute completion rules, description structure, prohibited claims, and how out-of-stock or discontinued items are treated. Include a good example and a rejected example with the reason. Two things happen next. Your supplier can meet it, and you usually find the inconsistencies you had been working around for months.

A sampling routine instead of full checks

Checking every deliverable defeats the purpose of delegating it. Check a fixed sample each cycle — ten listings, twenty data rows, five responses — against the written standard, and record what you find. Sustained errors mean the standard is wrong or the training is missing; isolated errors mean a review step needs adding. Either way, the finding is about the process, and that is the point of checking.

Where it usually goes wrong

  • Outsourcing descriptions before writing the voice — you get thousands of interchangeable listings and a worse brand than you started with.
  • Handing over the whole catalogue in one go — start with one category, get the standard right, then scale.
  • Not defining who updates what — when both sides assume the other owns a change, listings drift silently.
  • Losing the seasonal flexibility — peaks are when the arrangement is actually tested. Agree how volume changes are handled before your first one.

A sensible first piece of work

Take one category, one platform, and a few dozen products. Supply the template, the standard and the examples. That is enough to test whether your documentation is clear and whether the review rhythm fits your week, without risking the catalogue. If you sell on Shopify, Amazon or WooCommerce, our e-commerce management service covers exactly this work, and there is a parallel page on e-commerce outsourcing for broader operational support. The case studies describe the kinds of projects we take on, without naming clients.