Program managers inherit dashboards built by VMS vendors for VMS vendors. The metrics that populate those dashboards — submit volume, requisition count, days-to-submit — were designed to show that the VMS is processing activity. They were not designed to predict whether your program is actually working.
The result is a common misalignment: program managers who run tightly controlled scorecards on the wrong numbers. Vendors game the metrics that are measured. The metrics that would actually reveal program health either go uncounted or get buried under vanity indicators that look good in quarterly reviews but say nothing about quality.
After 14 years managing MSP programs — including a top-3 ranking among 30+ vendors for Cox Communications — here is what we have learned about which numbers actually matter.
The two metrics to deprioritize
Before the five that matter, two that do not — at least not the way most programs use them.
Speed-to-submit. The faster a vendor submits candidates, the higher they score on most dashboards. The logic seems reasonable: faster submits mean the vendor is engaged and active. In practice, speed-to-submit correlates strongly with spray-and-pray behavior, not quality. A recruiter who submits eight candidates within four hours of receiving a requisition has not screened eight candidates — they have copy-pasted eight résumés and called none of them. The real signal is not how fast the first submit came in. It is what percentage of those submits resulted in interviews.
Speed matters, but it matters at the right stage. Our average time to first submit is 24 hours across active programs. That is fast. The difference is that every candidate we submit in that window has been spoken with, screened for compliance eligibility, and confirmed on availability and salary alignment before they hit the VMS. Fast submits with no pre-screen are not a service — they are administrative noise that costs your hiring managers time.
Vendor panel size. More vendors on the panel means more competition, which should mean better coverage — that is the conventional wisdom. The data does not support it. Beyond a certain threshold — usually six to eight active vendors on a given requisition category — adding more vendors produces diminishing returns on fill quality and increasing overhead on program administration. Larger panels also tend to depress individual vendor fill ratios, because each vendor is working fewer requisitions at full effort and more at a low-probability attempt. The programs with the best fill quality we have worked in tend to have tighter panels with deeper accountability per vendor, not broader panels with diluted accountability.
The five metrics that matter
These five are the indicators we watch on every program we support, and the ones we recommend program managers weight most heavily in vendor scorecards.
1. Fill ratio per vendor (not overall fill). Overall program fill rate is a useful headline number, but it masks enormous variance between vendors. A program with a 35% overall fill rate might have two vendors filling at 50%+ and four vendors filling at 15%. The vendors at 15% are not just underperforming — they are consuming program capacity: intake calls, requisition access, VMS slots, and hiring manager interview time. Vendor-specific fill ratio is the most honest diagnostic of who on your panel is actually contributing. It should be reviewed quarterly, broken down by requisition category, not aggregated across all job types.
2. Submitted-to-interview rate. This is the pre-screen quality indicator. It answers the question: of the candidates a vendor submits, what fraction actually makes it to a hiring manager conversation? A high submitted-to-interview rate — our benchmark is 40% — means the vendor is doing real qualification work before the submit, not after. A low rate, say 12–15%, means the hiring manager is doing the recruiter's job: sorting through unqualified candidates to find the ones worth speaking with. Track this per vendor, per role category, and over time. It should be stable or improving. A vendor whose submitted-to-interview rate is declining has a quality problem that will not self-correct.
3. Interviewed-to-offer rate. Where submitted-to-interview measures pre-screen quality, interviewed-to-offer measures candidate-role calibration quality. A candidate who passes the recruiter screen and reaches the hiring manager interview should have a high probability of receiving an offer, if the recruiter genuinely understood the role requirements. Our benchmark is 48%. A rate below 25% typically indicates a calibration gap — either the recruiter does not fully understand what the hiring manager is actually looking for, or the requisition is mis-scoped and the vendor is not flagging it. Both problems are fixable, but only if you are measuring the right indicator.
4. Time-to-backfill. Every placement eventually ends — contract completions, voluntary departures, performance exits. Time-to-backfill measures how quickly a vendor can replace a departed placement on the same requisition or comparable role. This is a program resilience metric. Vendors with deep, maintained talent pipelines can backfill quickly because they are not starting from scratch on role knowledge and candidate sourcing. Vendors who built their placement on a one-time sourcing sprint cannot. Track it. A vendor who fills roles at 45% but takes 60 days to backfill departures is creating a different kind of program fragility than their fill rate suggests.
5. 90-day retention rate. This is the lagging indicator that validates everything above. A placement that does not survive 90 days is, by most program definitions, a failed placement — the role goes back to open, the sourcing process restarts, and the hiring manager resets their timeline. The 90-day retention rate per vendor reveals whether a vendor's placements are genuinely well-matched or whether they are winning on speed and losing on fit. We track 90-day retention on every placement we make, on every program. It is the metric we optimize the entire process toward — because it is the metric that proves the placement worked.
"Fill ratio is the single most honest number in an MSP program. It does not lie."
How to structure vendor scorecards
Most vendor scorecards weight speed heavily because it is easy to measure and easy to surface in a VMS dashboard. A better weighting model for a quality-focused program looks something like this:
- Fill ratio per vendor: 25% of scorecard weight
- Submitted-to-interview rate: 20%
- Interviewed-to-offer rate: 20%
- 90-day retention rate: 20%
- Time-to-backfill: 10%
- Speed-to-submit: 5% (gating metric — must be within program SLA, not a differentiator above the threshold)
Review cadence matters as much as weighting. Quarterly scorecard reviews with individual vendor debriefs — not just aggregate reporting — are what create accountability. A vendor who sees their submitted-to-interview rate declining quarter over quarter has a specific problem to solve. A vendor who only sees an aggregated program health score has no clear signal.
It also helps to separate volume vendors from specialty vendors in your scorecard analysis. A high-volume generalist vendor and a specialty technical vendor should not be evaluated on identical criteria. The generalist's value is throughput and breadth; the specialist's value is conversion rate on hard-to-fill roles. Applying one scorecard to both obscures the contribution of each.
Our own numbers
We do not ask clients to take our word for any of this. Our performance record is public and specific.
At Cox Communications, we have supported their contingent workforce program for 14 years, delivering more than 600 placements and consistently ranking top-3 among a vendor panel that has included more than 30 staffing firms. Our fill ratio on the Cox program has held a 12% fill ratio relative to total program requisitions — meaning we fill more than we are allocated relative to our panel share. That performance, sustained over 14 years, is what the five metrics above look like in practice.
At SiteOne Landscape Supply, where we have been a primary vendor since 2017, our fill ratio is 48%. We are their top-ranked staffing vendor on the program. That number reflects a combination of Leap Tiger™ candidate matching, compliance pre-screening, and 10-year average recruiter tenure on the account — not a single variable.
If you manage an MSP program and want to walk through how your current scorecard compares to this model, our MSP team is happy to do a 30-minute diagnostic conversation. No sales pitch — just a structured look at what your current metrics are telling you and what they might be missing.