When does a contingent programme need a VMS?

The instinct is to size the decision by how many contract workers you engage, and that is the wrong axis. A hundred workers from one supplier on one agreed card is administratively simple: one contract, one invoice, one relationship to hold in your head. Fifteen workers from five suppliers is not, because every commercial term now exists in five places and the person who agreed each one has moved on. What a vendor management system centralises is the supply side — who is approved for which category, on what terms, with which documents current. Count your suppliers and the number of separate rate conversations happening each month, and you have a better read on the need than any worker headcount gives you.

Four recur, and any two together are usually enough. Somebody asks what you spend on contract labour in a year and the honest answer takes a week to assemble. Two suppliers put forward the same person and the argument is settled by whoever shouts, because no record shows who got there first. A rate agreed for one role quietly becomes the reference for a different one, and nobody notices the card has moved. An assignment that was meant to run for a quarter is still running, because no date was diarised anywhere that anybody looks. Each is a symptom of the same underlying thing: the commercial facts live in mailboxes rather than in a record.

More than buyers expect, and this is the honest counterweight. Before a programme runs, somebody has to define supplier categories, load the rate card with its levels and inclusions, decide how requisitions reach the panel, set approval routes for timesheets and exceptions, and agree what compliance documents are mandatory. Then every supplier has to be onboarded into it, which means asking organisations that were happily emailing you to change how they work. That effort is fixed regardless of programme size, which is exactly why it repays itself on a large panel and does not on a small one. A system bought and half-configured is worse than the spreadsheet it replaced, because now the record looks authoritative and is not.

For a while, and for genuinely light volumes, yes. An applicant tracking system can hold contract candidates, move them through selection and record who was engaged. Where it stops is the part after selection: an assignment carries a rate, a start and an end date, extensions, timesheets, approvals and an invoice trail, and the counterparty is a supplier that keeps supplying rather than a candidate who is hired once. The two record types diverge at exactly the point where the money starts. If nobody is enforcing rates and no timesheet needs approving, the divergence does not matter yet. Once it does, bolting those fields onto a hiring record tends to produce a system that answers neither question well.

Three things, all free, and doing them first tells you whether you have a software problem or a process problem. Write down every supplier you currently use and what commercial terms you believe apply to each, which is usually the moment somebody discovers a supplier nobody remembers approving. Agree one ownership rule for candidates put forward by more than one party, and tell every supplier what it is. Put every live assignment on one list with a start date, an intended end date and an owner. If those three exercises restore control, you have bought yourself time. If they collapse again within a quarter because the maintenance has no owner, that is the honest evidence a system is warranted.

Ask what would change on the day the platform went live. If the answer is that the same person would keep the same list in a better-shaped container, nothing has been solved and the licence is buying tidiness. If the answer is that rates would be validated before a profile is reviewed rather than argued about at invoice, that duplicate claims would be settled by a timestamp rather than a phone call, and that a spend total would exist without anybody assembling it, the case is real. Programmes that grow into the need usually revisit this every year or so rather than deciding once, because supplier count moves and the threshold moves with it.