The task looks simple until somebody counts. Every supplier who invoices you has to be able to reach you through whatever channel the mandate specifies, and you need to know how to identify each of them.
With forty suppliers that is a morning. With two thousand it is a programme, and the naive approach — a mail merge asking everyone for their details — produces a response rate that makes the remaining work larger rather than smaller, because now you have partial data and no way to tell whose is missing on purpose.
Segment by capability, not by spend
The instinct is to sort suppliers by value and start at the top. That optimises the wrong variable. A large supplier is usually the easiest, because they have software and a team that has done this before. The effort is concentrated in the tail, and the tail is where the obligation will fail.
Sort by what each group actually needs from you.
| Segment | How to identify them | What they need from you | Realistic outcome |
|---|---|---|---|
| Already reachable on the network | A directory lookup against their tax identifier, done in bulk | Your address, and confirmation you are live | Complete without a conversation; this is usually a larger group than expected |
| Capable software, not yet connected | Mid-sized suppliers with a recognisable ERP | A specification, a test contact, and a date | High completion, on a timetable of weeks |
| Willing but without means | Small suppliers, sole traders, local services | A route that costs them nothing, explained in a page | Completes late, and only if the free route is genuinely free |
| Occasional and low value | The long tail, a handful of invoices a year | Nothing elaborate; a fallback channel | Never fully completes, and should not be planned as if it will |
| Will not or cannot move | Foreign suppliers outside scope, or the actively unwilling | A decision from you, not a further email | Handled as a permanent exception, deliberately |
Doing the first row properly is the highest-value hour in the exercise. A bulk lookup against a network directory tells you which suppliers are already reachable, and it removes them from the campaign entirely. Running the campaign first and discovering afterwards that a third of the recipients were already connected is a way to spend goodwill you will need later.
Collect the address, not half of it
One field causes most of the failures, and it fails the same way every time.
An electronic address is two things: a scheme identifier saying which identification system the value belongs to, and the value itself. A tax registration, a company registration and a network-specific identifier are different schemes, and the same digits can be valid in one and meaningless in another.
Ask a supplier for "your Peppol ID" and a proportion will send a number without saying what kind of number it is. It will be stored. It will look complete. Nothing will deliver, and the failure will be silent rather than an error — the mechanism is explained in how a participant identifier becomes a delivery.
So the form asks for both, separately, with the scheme as a constrained choice rather than a free-text box. That single design decision removes most of the rework, and it is the same discipline that master data quality requires everywhere else.
An address collected and stored without being tested is an assumption. Verifying it — a directory lookup at the moment of entry, or a test document before the supplier is marked complete — turns a campaign that reports ninety per cent complete into one that is. The difference surfaces on the first day of live running, when the untested ten per cent all fail at once and the exception queue receives a month of work in an afternoon.
Say what you are asking and why
The communication itself decides the response rate, and most of them are written badly for a predictable reason: they are written by the project, for the project.
A supplier reading it wants four things and will stop reading if they are not in the first paragraph. What is changing. When. What they specifically have to do. What happens if they do nothing.
That last one has to be answered honestly. If the answer is "your invoices will not be accepted", say so, because a vague warning produces no action and a specific one does. If the answer is "nothing yet, we will contact you again", say that instead of manufacturing urgency, because a supplier who is told something is critical and then finds it was not will not believe the second message.
And send it from the relationship, not from the project. A message from the buyer a supplier deals with gets read. A message from an unfamiliar address about a compliance programme is filtered by the same instinct that filters everything else.
What to do about the ones who will not move
Every plan needs an answer here and most defer it.
Three things are true at once. Some suppliers genuinely cannot connect on the timetable. Some are outside the mandate's scope entirely and never had to. And a small number simply will not, for reasons that have nothing to do with capability.
The workable arrangement is a free route — a supplier portal or an equivalent entry point that requires them to buy nothing and learn little — plus an explicit list of permanent exceptions with a named owner. The exceptions are not a failure of the programme; they are a category that has to be operated deliberately, because accounts payable will keep receiving from them whether or not the project acknowledged it.
What does not work is treating the tail as a completion percentage to be driven upward. It reaches an asymptote, and effort spent past that point would have been better spent making the fallback channel work properly.
Start before you think you need to
Onboarding has the longest lead time of any activity in the programme, for a reason nothing can compress: it depends on other people replying.
The receiving obligation typically arrives before the issuing one, which means the deadline for this work is earlier than the deadline everyone is planning towards. A programme that sequences onboarding after the issuing build has scheduled its longest-lead activity last.
Counting the supplier base by segment is cheap and belongs in the readiness assessment, months before anything is built. It costs a database query and it is the number that tells you whether this is an administrative task or a small campaign with its own budget.
The measure that means something
Not the percentage of suppliers onboarded. The percentage of invoice volume arriving through the intended channel, split by segment — because a hundred connected suppliers who send four documents a year is a worse position than twenty who send four hundred each, and the first number reports better.
Track the volume figure, publish the segment split alongside it, and resist the temptation to report a single completion percentage. It is the metric most likely to be green while the operation is not.
