How to Build a Hospital Float Pool in 90 Days: A Step-by-Step Operator

Most healthcare organizations that try to build an internal resource pool spend six months on planning and never launch, or launch with a bench too shallow to make a meaningful dent in agency spend.

The ones that get operational in 90 days do three things differently: they build the bench before the platform goes live, they set compensation before recruitment, and they treat the first 30 days as infrastructure, not pilot.

Here is how to do it.

Before Day One: The Decisions That Determine Everything

Two decisions need to be locked before any recruitment or platform setup starts. Getting them wrong costs months.

Decision 1: What shifts are you trying to fill?

Define the target scope before you build. Which units, which roles, which shift types are you trying to cover with the IRP? Start narrower than you think you need to. A focused IRP covering med-surg RN gaps across three units will be functional in 90 days. A sprawling IRP covering every role across every department will still be in setup at month six.

Most organizations should start with their highest-volume agency spend category, usually med-surg or step-down RNs, and expand from there.

Decision 2: What will you pay IRP staff?

Compensation structure must be set before you recruit a single nurse into the pool. IRP staff are choosing between your pool and per-diem agency work. If your rate is not competitive, you will not build a bench.

The benchmark: base hourly rate plus $3–$8/hour IRP premium, or a flat bonus of $50–$100 per shift accepted. Some organizations also build in an urgent-fill bonus for shifts accepted within 24–48 hours of start time. Get your finance and nursing leadership aligned on this number before Day 1.

Days 1–30: Infrastructure

This phase is all setup. Nothing goes live. The temptation is to rush recruitment, resist it.

Week 1–2: Platform configuration

If you are using DirectShifts, this is platform onboarding: building your facility profile, unit configuration, shift templates, and credential requirements by role. Get your scheduler or workforce management team involved early, they are the daily operators once the IRP launches.

Set up your shift posting workflow. Decide who posts shifts, what notice window is standard, and what the escalation path is when an IRP shift goes unfilled (the fallback to agency should be defined, not improvised).

Week 3–4: Credential matrix

Build the competency and orientation matrix for each unit before you credential anyone. This is the document that defines which nurses are cleared to float where. It should include: required unit-specific orientation hours, documented competency checkoffs, and any specialty certifications required (ACLS, BLS, NIH Stroke Scale, etc.).

Most organizations have something like this already. It usually needs to be formalized and digitized before it can be operationalized in an IRP platform.

Days 30–60: Bench-Building

This is the highest-priority phase. Your IRP is only as functional as the number of credentialed staff in it.

Week 5–6: Internal recruitment

Start with your existing workforce. Current per-diem staff, part-time nurses looking for additional hours, and recently retired staff who want occasional shifts are the fastest path to bench depth. Internal recruitment requires leadership buy-in, unit managers who feel threatened by a float pool that pulls their staff will slow this down. Get ahead of that conversation early.

Your messaging to potential IRP staff: more control over schedule, premium pay for IRP shifts, no minimum hour commitment, work across different units if they want the variety.

Week 7–8: External IRP enrollment

Open the pool to external per-diem and part-time candidates. This is where a staffing platform with an existing clinician network accelerates the timeline significantly. DirectShifts can seed your IRP bench with pre-screened candidates from our existing network, nurses and allied health staff who are already in the system, credentialed, and looking for per-diem work.

Target bench size before launch: 2–3x the number of shifts you plan to fill per week through the IRP. If you want to fill 20 IRP shifts per week at launch, you want 40–60 enrolled IRP staff minimum. More is better. A thin bench means the same nurses get called for every shift and burn out fast.

Week 7–8 (parallel): Unit orientation

Begin running orientation sessions for enrolled IRP nurses. For nurses cleared to float to multiple units, each unit requires separate orientation. This takes time and requires unit manager coordination. Start it before you need to, do not try to credential nurses into new units at the same time you are trying to fill live shifts.

Days 60–90: Soft Launch and Calibration

Week 9–10: Soft launch

Post IRP shifts for a limited set of units, your pilot units from the initial scope definition. Fill what you can through the IRP. Track what you cannot. Use agency for the overflow.

The goal of the soft launch is not to eliminate agency spend immediately. It is to test the shift-posting workflow, the notification and acceptance process, the documentation, and the payment flow. You will find friction points here. Fix them before full launch.

Week 10–11: Measure and adjust

Key metrics to track in the soft launch:

• IRP fill rate: what percentage of posted IRP shifts are being accepted
• Time to fill: how long between shift posting and acceptance
• No-show rate: what percentage of accepted IRP shifts result in a no-show
• Agency fallback rate: how often you are falling back to agency because the IRP could not fill

If your fill rate is below 60%, your bench is too thin or your compensation is not competitive, identify which and fix it before full launch.

Week 12: Full launch

Expand to all target units. Post agency fallback as a defined last resort rather than a first call. Track savings monthly against your pre-IRP agency spend baseline.

What to Expect in Months 4–12

The 90-day target gets you operational. Full optimization takes longer.

Month 4–6: Fill rates improve as IRP staff get familiar with the platform and units. Agency spend starts declining but does not fall off a cliff. Bench depth is the primary variable, keep recruiting.

Month 6–9: The IRP is handling the majority of predictable gaps. Agency is being reserved for true spikes and specialty coverage outside IRP scope.

Month 9–12: Savings become material and visible in the budget. Some organizations reach 50% reduction in agency spend by month 12. Others take 18 months. The difference is almost always how much bench-building effort went into months 1–4.

The Three Things That Kill IRPs

Know these before you build.

Launching without enough bench depth. An IRP that posts shifts and does not fill them trains your schedulers to distrust it and fall back to agency immediately. You lose the behavioral change before the tool ever has a chance to work.

Paying below-market IRP rates. If your per-diem premium is not competitive with what local agencies are paying for per-diem shifts, your nurses will choose the agency. Survey your market before setting rates.

Unit manager resistance. Managers who feel the IRP pulls their staff away from their unit will quietly discourage participation. The solution is not to fight it, it is to show managers that the IRP fills their gaps, not just pulls their people. Framing matters and executive leadership sponsorship for the IRP rollout is essential.

See how DirectShifts builds your IRP from bench to platform. Book a demo and get a 90-day implementation timeline specific to your facility.

Estimate your annual savings with the DirectShifts IRP cost calculator.

Frequently Asked Questions

How long does it take to build a hospital float pool?A focused internal resource pool covering 2-3 units can be operational in 90 days if you prioritize bench-building in the first 30 days. Full operational maturity, meaning the IRP handles the majority of flex staffing needs without agency fallback, typically takes 12-18 months. The biggest variable is how aggressively you recruit and credentialize IRP staff in the early phase. Organizations that launch with shallow benches take much longer to see real impact.

What is the first step to building an internal resource pool?Before anything else, decide which units and roles you are targeting and set your per-diem compensation rate. Those two decisions determine everything downstream. Organizations that start platform setup or recruitment before locking compensation end up revising it mid-build, which disrupts early recruitment. Set the rate first, then recruit to it.

How do you get nurses to join an internal resource pool instead of working for an agency?Competitive per-diem premium pay is the primary lever. IRP nurses need a financial reason to choose your shifts over agency work. Most organizations pay base hourly rate plus $3-8/hour for IRP shifts, or a flat shift bonus of $50-100. The secondary lever is convenience: if accepting an IRP shift takes more steps than calling an agency recruiter, staff will choose the agency. A mobile-first platform where nurses see available shifts and accept in one tap is the baseline.

How many nurses do you need in an IRP before it is worth launching?At least 2-3 credentialed nurses per shift you want to fill per week, before you go live. If you want to fill 20 shifts per week through the IRP, you need 40-60 enrolled nurses at launch. Less than that and your fill rate will be too low to change scheduler behavior. Schedulers who try the IRP twice and cannot fill the shift both times go back to calling the agency and do not try again.

What is the biggest reason internal resource pools fail?Launching with a bench that is too small. Organizations that get excited about the platform and go live before they have enough credentialed staff in the pool end up with a tool that does not work, schedulers who give up on it immediately, and no behavior change. The platform is not the hard part. Building the bench is. Everything else follows from that.

Can you build an IRP while still using agency staff?Yes, and this is how most organizations do it. The IRP fills what it can and agency covers the overflow. Over time the IRP proportion grows and agency usage declines. You do not need to cut agency contracts to start building, and trying to do both at once usually creates coverage gaps that make leadership nervous and slow the transition.