A warehouse candidate who applies on Monday morning has applied to three agencies by lunch. All three have her number. The one that calls first almost always gets her — not because its assignments pay better, but because she wants the matter settled, and the first recruiter to reach her settles it.
The race runs in both directions
Staffing loses the same race twice. On the candidate side, the first agency to make contact places the person; the resume sitting in the second agency's inbox is a record of work that will never be paid for. On the client side, an order rarely goes to one firm — it goes to two or three at once, and the first to present a qualified, credentialed, confirmed person wins the fill. The runner-up did all the same sourcing and screening for nothing. Both contests reward a single capability: how quickly a human at your firm makes contact with another human. That is a measurable number, and most firms have never measured their own.
Speed-to-contact, measured
The metric is simple: elapsed time from application received to first contact attempt. The mechanism for collapsing it: the application triggers a text within a minute, acknowledging receipt and asking the two questions that matter first — shift availability and earliest start date. The replies sort the recruiter's morning queue, so the day begins with candidates who are responsive right now instead of with the top of an alphabetical list.
The best-known research on response speed — an audit of 2,241 companies published in Harvard Business Review in 2011 — found the average firm took 42 hours to respond to a new web lead, and that responding within an hour multiplied the odds of qualifying it roughly sevenfold. That research is about sales leads and it is fifteen years old, but a candidate with three applications out behaves exactly like a buyer with three tabs open. The same logic covers the inbound line: candidates call from break rooms on ten-minute breaks, and an unanswered call that triggers a text inside a minute keeps the thread alive until the next break. The candidate who cannot reach you at 12:20 has been placed by someone else by 2:00.
The paperwork wall between yes and day one
A placement is not done when the candidate accepts. Between yes and the start date stands a stack: the I-9 — the employee's section completed by the first day of work and the employer's verification within three business days, which is a federal requirement rather than a preference — the W-4, direct deposit details, background-check consent, drug-screen scheduling, and whatever the role demands on top. A forklift operator card for the warehouse. A current TB test and BLS card for a clinical placement. A DOT medical card for a driver.
The manual version of collecting all that is a recruiter texting "did you ever send the back of your license" at 4:50 on the Friday before a Monday start. The automated version is a per-placement checklist that requests one item at a time by text, receives photos back, reads the document type and expiration date off each one, and files it. Run across the whole bench instead of one placement, the same extraction flags the CNA whose BLS card lapses in three weeks before she is submitted to a client rather than after. Reminders escalate to a recruiter only when a candidate stalls. The judging stays human: a person examines the original documents the I-9 requires, and a person decides whether a credential is genuine. The system does the chasing.
The 9pm order
Light-industrial and clinical desks both know this call: a client at 9pm, two no-shows on third shift, four people needed by 10:30. The manual response is a recruiter dialing down a list of people who are asleep or already working. The wired response filters first — right credential, right shift pattern, marked available, not currently assigned — and sends that pool a claim-by-reply text. First confirmations take the slots; the recruiter confirms the roster with the client and only starts dialing if the pool comes up short. Two things make this work. The pool is built at registration, where availability and written consent to receive these texts are captured properly. And the client-facing call stays human, because a client at 9pm is buying reassurance as much as labor.
The 6am problem
First-day no-shows are the trade's chronic disease, and the damage lands twice — the shift goes unfilled, and the client's confidence goes with it. The counter is a confirmation sequence: a day-before text with report time, address, the supervisor's name, and dress or PPE requirements; a morning-of check-in that asks for a one-character reply when the candidate is on the way; and an alert to the recruiter when that reply does not come by a set cutoff. The difference between calling the client at 5:40 with a warning and a replacement already moving, and the client's floor supervisor calling you at 7:05, is frequently the whole relationship. Logged over months, no-show rates by shift and by source also tell you which job boards send people who actually appear — a number worth reallocating ad spend over.
What this will not fix
No contact speed compensates for a pay rate two dollars under the shop across town, and the data will say so plainly: an order broadcast to forty qualified people that draws two replies is a pricing signal, not a messaging failure. Speed also cannot make a bad match stick — churn from oversold assignments is a screening problem that lives upstream of any automation. Texting a pool obligates you to collect consent at registration and to honor an opt-out immediately and permanently; build that in from the first message, not after a complaint. And extraction reads cards — it does not decide that a card is genuine. That examination, like the I-9's, is a person's responsibility and stays one.
Where to start
For one week, timestamp two things: when each application arrives, and when a human first attempts contact. Add the share of inbound calls answered live. Those numbers decide whether speed is actually your problem before anyone spends a dollar on fixing it.