What is on-demand staffing, and how do you run it?

On-demand staffing means maintaining a pool of workers you can call on when demand spikes, instead of carrying enough permanent staff to cover the highest peaks all the time. It is most common in operational teams with variable work: events, catering, hospitality surge cover, festival crews, volunteer coordination, retail seasonal peaks, and staffing agencies that supply workers to client businesses.

The mechanic is straightforward. Workers signal availability, jobs get offered to the pool, available workers respond, the business picks. The operational reality is harder. Building the pool, keeping it engaged, dispatching cleanly, and handling pay and classification properly are where on-demand staffing succeeds or quietly fails.

This guide covers when on-demand staffing earns its place over permanent hiring, how to build and maintain a reliable pool, the mechanics of dispatch and pay, and the failure modes to watch for.

When on-demand staffing earns its place

On-demand staffing has real overhead: recruiting and onboarding the pool, maintaining engagement, handling dispatch, managing pay and compliance. It pays back when three conditions hold:

Demand is genuinely variable. If your busiest week needs three times the headcount of your quietest week, an on-demand pool absorbs the spike without carrying that cost in the off-weeks. If your demand is reasonably steady, permanent hiring is simpler and produces better quality per hour worked.

You have access to a real labour pool. Some markets have plenty of casual workers (urban hospitality, events in major cities, university towns with students looking for shift work). Others do not (rural areas, specialist skills, small markets where everyone qualified is already employed). If you cannot actually recruit the people you need, on-demand is a wish, not a strategy.

The work does not require deep institutional knowledge. On-demand workers will not know your specific clients, processes, or unwritten rules at the level a permanent team would. If the work depends on that depth (client-relationship work, brand-sensitive customer interactions, complex multi-step setups), keep the core permanent and use on-demand only for the surge.

It does not work when:

You need consistent quality across every shift. On-demand workers vary more than permanent staff because you cannot develop them the same way. For work where quality varies materially with familiarity, the permanent team carries the standard and the pool fills the gaps.

Demand variability is artificially created. If you are running an on-demand pool because you cannot or will not commit to permanent headcount that the work actually justifies, the pool will be unstable and the cost will be hidden in turnover and quality drift. Permanent work pretending to be on-demand burns through workers.

Classification risk is high in your jurisdiction. Worker classification rules vary widely. In some places on-demand workers can legitimately be independent contractors. In others (the UK after the Uber and Pimlico Plumbers rulings, several US states with ABC tests, EU member states with strict tests), workers who look like employees are employees for legal purposes regardless of contract wording. Misclassification is expensive (back pay, taxes, penalties) and rarely worth the upfront savings. Worth checking with employment counsel before scaling the model.

Building the pool

A pool you cannot dispatch is not a pool. Three sources, usually used in combination:

Direct recruitment. Job boards, social media, referrals from existing workers. The pitch differs from permanent recruitment: you are selling flexibility, variety, and income on the worker’s schedule, not career path. Make the flexibility specific in the listing (no minimum hours, work as much or as little as you want, pay structure clearly stated). Vague flexibility promises in a job ad attract people who will be disappointed when the reality is more constrained.

Conversion from other relationships. Former permanent staff who left for good reasons, returning seasonal workers from previous peaks, volunteers who want occasional paid work. These people already know your operation, which makes onboarding lighter and reliability higher than cold-recruited workers. Many on-demand pools are built primarily from this source.

Agency supply. If you need a large pool quickly and do not want to manage recruitment yourself, a staffing agency runs the pool and dispatches workers to you. The trade-off is cost (agency margin on top of worker pay) and that the workers are not exclusively yours, so reliability on your specific needs is mediated through the agency relationship.

For all three sources, the new pool member needs onboarding before they are dispatched. A worker who signs up but does not get briefed properly on your operation will accept their first shift, have a confused experience, and not respond to future calls. Match the depth of onboarding to the assignment length: a worker doing one event shift needs a 15-minute briefing plus a pre-arrival package; a worker doing a month-long temp placement needs two or three days. The principle is not to skip onboarding because the worker is “only” on-demand.

One operational practice worth adopting: a low-stakes trial shift before counting on a new pool member for high-stakes work. The trial tells you whether the person actually shows up, on time, ready to work, with a basic attitude that fits the operation. People who pass the trial join the active pool; people who do not get a polite note and stay on the nominal list. This is how staffing agencies and event companies sort their pools without taking reliability risks on important bookings.

Running the pool

Two dimensions need clear rules: how jobs get to workers, and how workers get paid.

Dispatch. Post the job to the pool with the information a worker needs to decide whether they want it: outcome (what they will be doing), location, shift times, pay rate, supervisor on site, anything role-specific. Available workers signal interest, you pick from the responders, the unassigned people get the next post when it arrives. This is the on-demand bidding pattern, covered in more detail in the shift bidding guide.

The pace matters. Workers in active pools respond to posts within minutes. If you take hours or days to pick, the responders cool off, take other work, or stop trusting that responses lead to assignments. Pick fast, or set an explicit expectation about when picks happen. A “we will pick by 2pm” note on the post is enough. Silence after responses is the thing that erodes the pool.

Pay structure. Several patterns work, usually in combination:

Hourly or per-shift base pay. The simplest. Workers paid for the hours they actually work.
Day rate. One number for the day, regardless of exact hours within reason. Common in events and catering where the shift length is approximate.
On-call retainer. A smaller flat payment for being available during a specified window, whether or not called in. The retainer is an honest acknowledgment that “be ready to come in” has value even if the call does not come. Used in emergency services, some healthcare, occasionally for premium events.
Premium pay for short-notice. Higher rate for shifts called with less than 24 or 48 hours notice. Reflects the worker’s higher coordination cost and helps fill less appealing same-day shifts.
Cancellation pay. A token payment to workers whose confirmed shifts get cancelled by the business at short notice. Preserves goodwill with the pool when the business causes the disruption.

Document the pay structure in writing and stick to it. Workers compare notes; inconsistency reads as favouritism and erodes the pool faster than low pay does.

Common failure modes

On-demand staffing works on a trust loop. The pool trusts that responses lead to assignments often enough to be worth staying engaged. The business trusts that the pool will show up when called. Either direction breaking causes the system to fail.

Four patterns that recur:

Pool decay. Your nominal pool size is 200, your active pool is 40, and the active pool is shrinking. Caused by some combination of slow manager response, picking the same favourites repeatedly, vague posts that careful workers ignore, and pay or cancellation handling that erodes goodwill. Watch the response rate over time, not just the headline pool size. A 40-person active pool with strong response rates is more useful than a nominal 200 where you cannot tell who actually picks up calls.

Coverage gaps in the off-peak. The pool wants to work the peak-rate shifts (Saturday nights, holiday weeks, special events) and ignores the baseline calls. You end up over-supplied during peaks and short-staffed for everything else. The fix is either differential pay (more pay for the unpopular slots) or a soft commitment structure (people who pick up baseline shifts get priority for peak calls).

Same-favourites trap. Five people always say yes, so they get picked for everything. The other 195 stop responding. You think you have a 200-person pool but really you have a five-person dependency, which fails the moment one of those people goes on holiday or takes another job. Rotate intentionally, even if it costs a little quality on individual shifts, because the alternative is fragility.

Quality drift. On-demand workers see your operation occasionally, not continuously. They forget changes, miss training updates, and slowly drift away from current standards. Build a way to push updates to the pool (a short pre-shift brief on what has changed, a periodic refresher, written reference material), or accept that the pool will be slightly behind your permanent team and plan accordingly.

The fix across all four is the same: honest two-way feedback after shifts. The business records reliability for the worker (showed up, on time, quality of work). The worker has some way to feed back on the post (was the information accurate, was the experience as described, was the pay processed properly). Over time the data shapes who gets called for which work, and the pool self-sorts toward people who genuinely want to participate on terms that work for both sides.

Where this fits

On-demand staffing earns its place when demand is genuinely variable, you have access to a labour pool that wants the work, and the work does not require depth of institutional knowledge for every shift. Events, catering, hospitality surge, festival crews, volunteer coordination, and staffing agencies are the natural fits. So is the operational core of any business that mixes a permanent backbone with surge capacity from a pool.

Zelos works for teams running on-demand staffing at the scale where building and maintaining the pool is worth the effort. Open jobs get posted to your pool, available workers signal interest from a mobile app, you pick from the responders, and the dispatch data feeds the trust loop on both sides. The simple staffing software page explains how it works.