Can't tell why your sales dropped? Sledge reads your store data and ranks the leaks by what fixing them is worth. Free for 14 days →

// Level 05 Making It Last · 5.4

Illustrated temple ruins in violet with an obelisk and carved pylons, linked by a bridge over the cliff under a golden moon.

When to hire or automate

Updated August 2026 6 min read
HOW OFTEN, AND HOW MUCH JUDGEMENT BUY IT IN Rare, needs expertise a one-off build, an accountant, a photographer THIS IS YOUR JOB Frequent, needs judgement what to test next, what to sell, what a number means LEAVE IT Rare, mechanical not worth the setup cost AUTOMATE Frequent, mechanical order routing, tracking emails, recording what shoppers do MORE JUDGEMENT LESS RARE FREQUENT

The quadrant most owners get wrong is the top right: they delegate the judgement work because it is hard, and keep the mechanical work because it feels productive.

The short answer

Two questions sort almost everything. How often does this recur, and how much judgement does it need? Frequent and mechanical is what you automate. Rare and specialised is what you buy in. Rare and mechanical is usually not worth touching. Frequent and judgement-heavy is your actual job, and it is the quadrant owners most often delegate first, because it is the hardest.

The two questions

A revenue threshold is the wrong shape of answer to "when should I hire", because it ignores what the work is.

Ask instead: how often does this recur, and how much judgement does it need. Those two sort every task in the business.

QuadrantAnswerExamples in a dropshipping store
Frequent, mechanicalAutomateOrder routing, tracking emails, recording what shoppers search and abandon
Rare, specialisedBuy it inA one-off theme build, an accountant, product photography
Rare, mechanicalLeave itAnything whose setup costs more than the hours it saves
Frequent, judgementKeep itWhat to test next, what to stock, what a number means

The fourth quadrant produces nothing visible on a good day, so an overloaded owner hands it over first. The first quadrant feels productive, so it gets kept.

Owners automate what they enjoy and delegate what they find hard. It should be exactly the other way round. The Dropshipping Playbook

What you can afford

The threshold is not revenue, and using revenue is how stores hire themselves into a loss.

MONTHLY HEADROOM = CONTRIBUTION PER ORDER × ORDERS
300 orders at $9 contribution is $2,700 to cover everything, not $12,000 of revenue.

Two stores with identical revenue can be nowhere near each other on this, which is the whole point of chapter 5.1. A thin-margin, high-volume store looks larger and has less room than a smaller store with healthy margins.

And one dropshipping-specific caution on top: headroom is not the same as cash. Chapter 1.3 documented reserves holding a share of every transaction for 90 to 120 days. A recurring commitment made against revenue that will not be spendable for four months is how a profitable store misses payroll.

What to automate first

In order of how clearly each one sits in the bottom-right quadrant:

What not to automate early: customer service replies, review requests that read as generated, and anything that makes a promise to a customer without a person having checked it can be kept.

The order people actually hire in

Roughly, and with the caveat that the right first hire depends entirely on which part of your week has become the bottleneck.

First, usually
Customer service. High frequency, moderate judgement, and the first thing whose quality visibly drops when the owner is stretched. It also directly affects the dispute rate that drives payment reserves.
Second
Content or creative production. Frequent, and the constraint on the ad testing in chapter 2.2 is usually creative volume rather than budget.
Third
A media buyer, but only once your monthly spend makes the fee a small share of it. Below that you are paying somebody to run a test that cannot conclude.
Rarely early
An agency for everything. Broad retainers on a small store buy generic work and remove you from your own numbers, which is the one thing chapter 5.3 says not to delegate.

The part that stays yours

You can hand over almost everything eventually. Three things are worth holding longer than feels comfortable.

  1. The weekly read. Ten minutes, four numbers. The moment somebody else does this, you are managing a summary rather than a store.
  2. The choice of what to change. Execution delegates fine. Prioritisation is the judgement quadrant, and it is what the whole of Level 04 is an input to.
  3. The supplier relationship. Your margin, your shipping times and your quality control all live here, and chapter 1.1 is about how little of it you control even when you are paying attention.

Everything else is a question of whether it recurs and whether it needs you. That is the end of the playbook.

Common questions

When can I afford to hire anyone?
Frame it as contribution rather than revenue. A hire has to be paid out of contribution per order times orders, not out of turnover, which is why a store doing good revenue on thin margins can be further from a first hire than a smaller one on healthy margins. Chapter 5.1 covers why the two look identical on a dashboard.
Should I hire a virtual assistant or buy software?
Software for anything that recurs identically and has a defined output: order routing, notifications, recording behaviour. People for anything that varies with context: customer service, supplier relationships, judgement about what to test next. A person doing a software job is expensive and error-prone, and software doing a person's job is worse.
What should I never delegate?
The weekly read and the decision about what to change, per chapter 5.3. You can delegate execution of the change. Handing over the interpretation means the person closest to your numbers is not you, and it is the one dependency that is genuinely hard to reverse.
Is a media buyer worth it early?
Usually not, because at a small testing budget you are paying somebody to spend an amount that cannot produce a conclusive result anyway, per chapter 2.2's 50-conversion rule. A media buyer earns their fee at a spend level where the fee is a small share of it, not at the level where it doubles your monthly cost.
Illustrated violet forest at dusk, with a yellow sun between the trunks.

One number pays for the rest

What a single order is worth decides what every other fix can afford. Chapter 4.1 is where that number moves.

Go to chapter 4.1