Vendor Onboarding KPIs
Which measures actually tell you something
Most onboarding dashboards measure what is easy to count rather than what would change a decision. A handful of measures are genuinely useful, and two of the most commonly reported are actively misleading.
The most useful vendor onboarding metrics are time to approved supplier measured from invitation rather than from internal request, completion rate of invitations sent, the stage at which vendors abandon, the share of records with expired documents, and the proportion of vendors re-screened within the last cycle. Average cycle time and total vendors onboarded are the two most commonly reported and the least informative.
Five that change decisions
Each of these, when it moves, tells you something specific to do.
| Metric | How to define it | What it tells you |
|---|---|---|
| Time to approved supplier | From invitation sent to approval recorded, in calendar days | The end to end experience, including vendor response time, which is usually the largest component |
| Invitation completion rate | Submissions completed divided by invitations sent, over a fixed window | Whether your questionnaire is finishable. A falling rate is almost always a form problem |
| Abandonment stage | The page or section at which incomplete submissions stop | Precisely which question is costing you vendors, which no aggregate number will show |
| Records with expired documents | Active vendors holding at least one expired document, as a share of the population | Live exposure right now, rather than how well onboarding ran historically |
| Re-screened within the cycle | Share of active vendors screened within the last screening period | Whether monitoring is actually running or has quietly become a one time check |
Four decisions to make once
Pick the start point and keep it
Measuring from internal request flatters nothing and measuring from invitation is the honest version. Whichever you choose, changing it later makes your history meaningless.
Use calendar days, not working days
Working day calculations hide the effect of a vendor who goes quiet over a holiday period, which is exactly the delay you want visible.
Exclude nothing without recording it
Removing outliers produces a better looking number and a worse understanding. If you exclude, record the rule and apply it consistently.
Measure the population, not the sample
Share of vendors with expired documents is only meaningful against all active vendors, not against those somebody happened to review.
Three worth dropping
These appear on most dashboards and each of them can improve while the underlying situation gets worse.
| Metric | Why it misleads |
|---|---|
| Average onboarding cycle time | Averages hide the tail, and the tail is the problem. A median plus the share taking over a threshold tells you far more than a mean that a few very slow cases distort |
| Total vendors onboarded | Counts activity rather than outcome. A rising number can mean growth or it can mean duplicates and vendors onboarded who were never needed |
| Documents collected | Counts arrival rather than validity. A collected certificate that expired last quarter counts identically to a current one |
Be careful with published numbers
| Claim | Why to treat it carefully |
|---|---|
| Industry average onboarding times | Rarely define their start point, so a figure measured from invitation is being compared against one measured from approval readiness |
| Percentage improvement claims | Usually vendor-supplied, frequently from a single customer, and almost never state the baseline |
| Your own first measurement | Treat it as a baseline rather than a verdict. The useful comparison is against yourself over time, not against a number in a brochure |
