Most outsourcing relationships start in the wrong place. The founder hands over the work they find most painful — the finance reconciliation they have been avoiding, or the recruitment pipeline they never built — and six weeks later both sides are frustrated, because the hardest work is also the work with the least defined process.
The fix is not a better vendor. It is a better first choice.
Bucket one: the work you already know how to do
Start with a workflow that already runs inside your company, that someone on your team used to do, and that has a measurable output. Onboarding checklists. Weekly bookkeeping. Lead list building. Support ticket triage. Appointment setting.
You can write the standard for it today, because the standard already exists — it lives in somebody's head. When a partner runs it, the gap between "how we do it" and "how it is done" is small enough to close in one review cycle.
This is the bucket that pays for the relationship quickly, because you can compare the before and after numbers without a spreadsheet full of assumptions.
Bucket two: the work that is blocking your team
Second comes the work that talented people are doing because nobody else is available for it. A senior engineer maintaining the CRM. A sales lead formatting reports every Friday. A founder answering the same six customer questions.
Here the return is not measured in salary saved. It is measured in hours returned to the people whose time is scarce. Watch what happens in the first month: the person who handed the work over should be visibly lighter, not quietly doing the same work plus supervision.
Bucket three: the work nobody owns yet
Market research. Data hygiene across three systems. Building the reporting layer. Video editing backlog.
Do not start here. It is tempting, because it looks like "new capability" — but with no internal owner there is no one to answer the partner's questions, so quality drifts before anyone notices. Get a relationship running on buckets one and two first, then widen the scope.
The two questions that decide the first workflow
Before you commit, ask:
- Can I describe done? If the output cannot be described in a sentence with a number in it — "a list of 400 qualified accounts, enriched, with a report on what landed" — the scope is not ready.
- Who signs off? Every output needs a named reviewer on your side. Not a committee. One person who catches the miss before a customer does.
If both answers are clear, you have your first workflow. If either is vague, the first job is not delivery — it is writing the standard, and that work is worth doing yourself.
What good looks like in week one
A scope in writing with owners, turnaround and price. One workflow running. A weekly report that shows what moved, not what was attempted. And a short list of things the partner deliberately did not do because they were outside the agreed scope — the clearest sign that boundaries are being respected rather than guessed.