Starting Without an Order
Sometimes you start before the paperwork exists. It is a decision, not an accident, and there is a way to make it survivable.
Sharedwho controls this stretch
Sometimes the work is urgent, the order will take three weeks, and the customer asks you to begin anyway. Refusing costs the job; agreeing costs the protection the order would have given you.
The commercial boundary in “Starting Without an Order” is clearer when estimated work, recorded time and authorised scope can be compared. For teams researching work weeks in a year, open this official product note can add that operational context, but purchase terms and payment obligations should remain in the signed agreement.
It is a commercial decision, it is frequently the right one, and the damage comes from making it without saying so.
As a separate reference point for “Starting Without an Order”, see Fiverr business resources. Confirm current legal, tax and accounting requirements with a qualified adviser where necessary.
What you are actually risking
Not usually non-payment. More often a long delay while a retrospective order is raised, and occasionally a refusal because nobody will authorise spend after the fact.
Two situations look identical from outside and are not. In the first, the spend is authorised and the order is administration that has not caught up. In the second, nothing internal has happened at all. The question that separates them is whether the budget has been approved — and it is a reasonable thing to ask before beginning.
In organisations with strict controls, the person who asked you to start may not have had authority to ask.
The email that makes it survivable
Short, same day, to the person who asked. What you are starting, on whose instruction, at what price, and that an order is to follow.
Four facts. It converts an informal request into a record.
Why that email matters later
Because the question that arises is always "who authorised this". With the email, there is an answer with a name and a date.
Without it, there is a supplier asserting that somebody said something.
Asking for the requisition number
Often an order is preceded by an internal requisition, which exists days before the order does.
Asking for that number gives you something to quote and shows the request is genuinely in the system.
Limiting the exposure
Start the part that is cheap, or the part that is useful to you regardless, and hold the expensive part until the order exists.
Staging is available even when refusing is not.
Invoicing without an order
Usually impossible to get paid on. The invoice is parked and the exception is resolved only when an order appears.
Sending it anyway is harmless and should not be confused with progress.
The retrospective order
Common, slow, and frequently requires somebody senior to approve something they would rather not have to.
Asking early makes it easier. Asking after three months makes it a problem somebody must own.
When it goes wrong
The person who instructed you has left, or denies the scope, or the budget no longer exists.
This is where the email is the whole of your position, and where its absence is usually decisive.
Doing it repeatedly for one customer
A customer who always asks you to start early and always produces the order eventually is a known quantity, and the risk is manageable.
A customer where it goes wrong once has told you something, and the second time is a choice.
The policy worth having
Start without an order only on a written request, only with an email recording it, and only for work you could absorb losing.
Three conditions. They permit flexibility without pretending it is free.
The two situations that look identical
Starting without an order covers two quite different positions, and suppliers routinely treat them as one.
In the first, the buyer has decided, the budget exists, somebody has instructed you in writing, and the order is a piece of internal administration that has not caught up. The risk here is delay and paperwork.
In the second, nothing internal has happened at all. Your contact wants the work, is confident it will be approved, and is asking you to begin on that confidence. The risk here is that the approval never comes and the work is never payable by anybody.
The two feel the same from outside — both are "we're waiting on the PO" — and the question that separates them is whether the spend has been authorised yet. Asking it directly is reasonable, and the answer decides whether you are accepting a delay or lending the customer a job.
Making the exposure finite
If you are going to start, the useful discipline is to decide in advance how much of the job you are prepared to have unauthorised, and to stop at that point rather than at the point where you notice.
A week of work, a defined first stage, or whatever sum you could write off without it mattering.
This converts an open-ended risk into a known one. It also creates a natural moment to raise the order again, with a reason attached: the first stage is complete, here is what was done, the order is needed before the next part begins. That is a much easier conversation than an abstract request for paperwork, because it attaches the administration to something the buyer wants.
Suppliers who do not set the limit in advance almost always discover it retrospectively, at whatever figure the job happened to reach before something went wrong.
The moves that make it worse
Start quietly, say nothing, and raise it when the invoice is rejected.
That sequence turns your own decision into their problem, and it is usually resolved at your expense.