RPO

What has to be true internally before an RPO will work?

A provider inherits whatever process exists. If requisitions are approved inconsistently, if hiring managers disagree about what the role is, or if offers take weeks to sign off, an outside team will surface those problems rather than absorb them. Buyers who treat the engagement as a way to avoid fixing an internal bottleneck are usually disappointed, and the honest sequence is to decide which parts of the process the client will own and hold to before handing the rest over.

How is scope creep handled in practice?

Demand rarely arrives as forecast. A workable agreement states what volume the base commercial terms cover, what happens above and below that band, and how a genuinely new category of hiring is priced rather than absorbed. Without that, either the provider quietly under-serves the excess or the buyer receives an unexpected invoice, and both outcomes damage the relationship more than the original conversation would have.

How does an RPO relationship differ from buying agency placements?

An agency is paid per successful introduction and represents candidates to the employer from outside the process. An RPO team runs the employer's process itself, under the employer's brand, and is paid for running it. That changes who the candidate believes they are dealing with, who holds the candidate record, and what the provider is accountable for when a requisition goes unfilled for reasons that have nothing to do with sourcing.

How is an RPO engagement priced, and what does each model reward?

Three commercial shapes are common. A management fee charges a fixed recurring amount for an agreed capacity, which makes budgeting simple and puts the risk of demand fluctuation on the buyer. A per-hire charge ties payment to output, which shifts risk to the provider but rewards volume of hires rather than quality of them. A resource-based charge prices the recruiters and coordinators assigned, which is transparent about what is being bought but does not by itself guarantee anything about results.

Most live contracts blend these, typically a base fee covering the team plus a variable element tied to delivery. The discipline is to read each model for what it quietly incentivises. Per-hire pricing rewards filling requisitions and is indifferent to whether the person is still there a year later, so if quality matters it has to be brought in through the measures rather than assumed. No pricing model removes the buyer's obligation to define what a good hire is.

Which service levels are worth writing into the agreement?

Useful service levels measure things the provider genuinely controls. Time from requisition release to first shortlist, shortlist quality expressed as the proportion of submitted candidates the hiring manager agrees to interview, and responsiveness to candidates all sit largely within the provider's hands. Measures that depend on the client, such as end-to-end time to fill or offer acceptance, are shaped heavily by hiring manager availability and pay decisions, and holding a provider to them without acknowledging that produces argument rather than improvement.

The other half of a workable service level is the client's own commitments: how quickly feedback is returned on a submitted profile, how fast interview slots are made available, and who approves an offer. Agreements that impose obligations in only one direction fail predictably, because the provider is then accountable for a cycle time it cannot influence at several of its steps.

What happens during transition, and why does it decide the outcome?

Transition is the period between signature and the provider running live requisitions, and it is where most of the eventual friction is either resolved or baked in. It covers access to systems, mapping the client's stages and approval flows, agreeing templates and branding for candidate communication, documenting each hiring manager's expectations, and deciding what happens to requisitions already in flight.

Engagements that go badly usually skipped this. The provider starts taking requisitions before anyone agreed what a submitted candidate must include, hiring managers have not been told the process changed, and in-flight roles fall between two teams. A transition plan with named owners and a date at which each element is considered live is unglamorous and is the single highest-return part of the setup.

What does governance look like once the engagement is running?

Two rhythms are needed. An operational one, usually weekly, works through live requisitions: what is stuck, which roles need a specification revisited, where feedback is outstanding. A strategic one, usually monthly or quarterly, looks at trends, service level performance, forecast demand for the next period, and anything that needs a commercial change.

Both need a single named owner on each side with authority to decide. The pattern that fails is a governance meeting attended by people who report findings to absent decision makers, so nothing changes between meetings and the same issues recur. It also helps to agree in advance how a disagreement about whether a service level was met is settled, because that argument is far easier to design a route through before it happens.

Who owns the data, and what happens at exit?

The talent the provider builds over the life of the engagement is the asset most often left unaddressed. The agreement should state who owns candidate records, pipelines, notes, and the reporting history, in what format they are returned, and within what period. Where the provider works inside the client's own applicant tracking system this is largely settled by default; where the provider uses its own tooling it very much is not.

Exit terms should also cover the practical handover: notice periods, whether provider recruiters can be offered roles by the client, continuity for requisitions open on the last day, and how candidates mid-process are told what is happening. Handling personal data at exit also carries obligations that differ by jurisdiction and by what candidates were told when their data was collected, so this is a point to confirm with counsel rather than to resolve commercially.