The First 90 Days: No-Shows, Week-One Quits and 90-Day Exits, and Why Most of Them Were Decided Before the Ad Was Written

Last updated August 27, 2026 · Closes the loop after the hire · Pairs with caregiver turnover and truck driver turnover

First-90-days turnover is the hire that never shows, the one who quits in week one, and the one who's gone by day 90 — and in frontline work it's most of the turnover there is. The numbers are third parties': CNA turnover ran 42.34% in 2025 (AHCA's salary and benefits report), home care caregiver turnover 75% (Activated Insights, 2025), and truck drivers file a new application after an average of 3.8 months in a role (Tenstreet, June 2026). We publish no retention data of our own; what we can say from ours is that the expectation an applicant arrives with was set by the ad, and an ad that names the wage, the shift and the site in the first line is the first retention tool. Below: the third-party numbers, where the loop leaks, the 90-day cadence, and two planners.

Book a Demo

A note on who's writing this

Boostpoint sells recruitment advertising, which ends at the application — we see applicants, not starts, and not who's still there at day 90. So every turnover figure on this page is a third party's, named and dated, and every cadence step is our practice or our clients' rather than a measured result; where we say "same day," that's how we run follow-up, not a statistic. The one claim we make from our own data is narrower: in our benchmark the click-to-application conversion rate explained 70% of the variation in what an applicant costs, and the same first line that converts — the wage, the shift, the site — is the one that sets the expectation the hire shows up with. Cost per applicant is not cost per hire, and nothing here is a cost per hire or a retention rate of ours.

The case that early turnover is a hiring problem, not a management one

The manager's version of early turnover is that people don't want to work anymore. The recruiter's version is more useful: a new hire who leaves in the first 90 days is usually leaving a job that turned out to be different from the one they applied for. The shift was nights when the ad said "flexible"; the pay was the base when the ad said "up to"; the site was forty minutes farther than the ad implied; the manager they met on day one wasn't the one who interviewed them. None of that is a training problem, and no onboarding program fixes a mismatch that was written into the ad. That's the case this page makes first, because it's the part a recruiting company can honestly speak to: the first retention tool is an accurate ad, the second is a fast, specific reply, and the third is a start date that arrives with the paperwork already done. The management half — the buddy, the check-ins, the 90-day review — matters too, and the cadence below has it, but it comes second.

One thing moves quietly across all five of those exits: what a candidate already believed about working for you before they applied. Employer branding on a budget treats that as a tiebreaker rather than a fix, which is the honest framing for frontline roles.

The third-party numbers

Frontline turnover is measured by the industries themselves, and the measured figures are worth reading exactly. In skilled nursing, the 2025–2026 Nursing Home Salary and Benefits Report — from the Hospital and Healthcare Compensation Service, supported by AHCA, covering 917 nursing homes and more than 111,600 employees — put CNA turnover at 42.34% in 2025, down from 44.16% in 2024 and still the highest of any position tracked (AHCA/NCAL, September 9, 2025). In home care, Activated Insights' 16th annual benchmarking report, released July 1, 2025, reported caregiver turnover "dropping to 75%, the lowest level reported in the past five years," and describes "the first 100 days when the risk is at its peak." In trucking, Tenstreet's analysis of roughly 5,000 carriers (June 4, 2026) found drivers submitting new applications after an average of 3.8 months in their current roles, with average tenure rising since late 2023 — and that drivers referred by another driver were 12 times more likely to be hired than the platform average. Across the economy, the BLS quits rate for accommodation and food services was 4.5% a month in June 2026 against 2.0% for all nonfarm employment (JOLTS, released August 4, 2026), and Work Institute, which runs exit interviews, puts it plainly: "Over one-third of newly hired employees quit within their first year." Every one of those figures is somebody else's, and each industry's own page has the detail: caregivers, drivers.

Where the loop leaks: five exits between the application and day 90

Treat the 90 days after the application as a funnel with five exits, because each one has a different fix. The interview no-show: the applicant applied to three employers in one sitting and went to the one that texted first; the fix is the reply, within minutes by text and the same day by phone, with a specific slot rather than "we'll be in touch" — the on-the-spot version collapses the interview into the walk-in (how that works). The offer that goes cold: the gap between "you're hired" and the start date is where a competing offer lands; the fix is a start date within days, paperwork sent the same day the offer is made, and a text the day before day one (the job offer). The day-one no-show: an offer accepted and a start date missed, usually because nothing happened in the days between; the fix is the cadence below. The week-one quit: the job wasn't the ad — the shift, the pay, the site, the manager — and this one is decided upstream, in the first line of the ad and the honesty of the interview. The 90-day exit: the hire found out what the job pays after the differential, who the crew is, and whether anyone noticed they were there; the fix is a buddy, a day-30 and day-60 conversation, and a 90-day review that's a raise conversation when it should be. Which exit is leaking is a number you already have: count the applicants, the interviews held, the offers accepted, the day-one starts and the day-90 survivors for the last quarter, and the biggest drop is where the money goes. Tag the source on each so you can see whether a channel's hires leave faster — the tagging is on the analytics page.

What early turnover costs — on your numbers, not ours

We don't publish a cost per hire and won't compute one from ad data, so the cost of an early quit here is built from two figures you supply and one third-party reference. Your figures: how many people you hire a year, and what share are gone by day 90 — from your own records, and if you've never counted, the last quarter's starts against today's roster is a ten-minute exercise. The reference: SHRM's average cost per hire of "nearly $4,700" (April 2022), an all-in, cross-industry figure that's a placeholder for what replacing someone costs you, not a measurement of your business. The advertising side is the smaller number and the one we can speak to: replacing an early quit means buying another set of applicants at whatever the role's cost per applicant is — the benchmark's all-industry median is $13.88 with the management fee inside — times your own applicants-per-hire ratio. The calculator does both and keeps them separate, because the advertising cost of a re-hire is not its cost per hire, and confusing the two is how a "cheap" channel with fast-quitting hires looks like a bargain.

The 90-day cadence: eight touches, with dates

The onboarding programs that reduce early exits are not elaborate; they're dated. What follows is the cadence we recommend and run with clients — practice, not a measured result — and the builder turns a start date into the calendar. Offer day: the offer by phone, confirmed by text within the hour, with the start date, the site, the shift and the pay in the text so there's a record the hire can re-read. Day −7: paperwork sent and the first day described — where to park, who to ask for, what to wear, when lunch is. Day −1: a text from the manager, by name: "See you at 6:45 tomorrow, ask for Dana." This is the no-show touch; most day-one no-shows had heard nothing since the offer. Day 1: the manager who interviewed them is the one who meets them; a buddy is assigned before lunch; the first day ends at a normal time. Day 7: the manager asks one question — "Is the job what you expected?" — and writes down the answer, because the gap between the ad and the job is now visible and fixable. Day 30: the first real check-in, with the pay stub in hand, the differential explained, and the schedule confirmed. Day 60: a conversation about the next thing — the certification, the shift they'd rather have, the crew lead track (the induction program). Day 90: the review, and if the ad said "raise at 90 days," the raise. Every one of those can be a text template; texting is the medium for all of them, and the communication glossary covers the tone.

How to keep new hires engaged when the manager is on the floor

The objection to any cadence is that the manager is running a shift and can't run a program. Three answers. Make the touches texts, drafted once, sent from a template with the name and date filled in — a day −1 text takes eleven seconds. Give the buddy the day-7 and day-30 questions, because a peer gets a truer answer than a boss does and the manager gets the written answer either way. And put the day-90 raise in the ad, so the promise is public and the review is a formality; a promise in the ad is the cheapest engagement tool there is, provided it's kept. What doesn't work is the welcome packet — a hire who got a binder and no text on day −1 is a hire who was welcomed and then forgotten, and the data above says the first 100 days is when they decide.

This page prices the early exit specifically — the hire who leaves before day 90. For the whole-year figure across every role you fill, computed from your own headcount, turnover rate and replacement cost, use the turnover cost calculator.

For the wider picture — how your industry's quits rate compares with the rest of the economy, and whether your hiring difficulty is really a retention problem — our labor shortage page works through the BLS openings, hires and quits figures industry by industry.

When the fix isn't onboarding

Three cases where the cadence won't help. If the pay is below the employer two exits down, the hire will leave at day 30 with the stub in hand, and the honest fix is the pay or the ad — advertise the real number and let the people who'll work for it apply. If the shift is nights and the ad said flexible, the week-one quit is a recruiting problem: fix the first line (the job ad page). And if a channel's hires leave measurably faster than another's — which your tagged data will show — the problem is the channel's pool or its honesty, not the onboarding; the driver version of that story, agencies over-promising home time, is on the truck driver turnover page.

Frequently asked questions

Third-party figures: AHCA/NCAL, "2025–2026 Nursing Home Report Reveals Lower Turnover," September 9, 2025, summarizing the 2025–2026 Nursing Home Salary & Benefits Report by the Hospital & Healthcare Compensation Service (917 nursing homes, more than 111,600 employees, 102 positions) — CNA turnover 42.34% in 2025, 44.16% in 2024; Activated Insights, 16th Annual Benchmarking Report for Home Care, Home Health and Hospice, July 1, 2025 — caregiver turnover 75%, "the first 100 days when the risk is at its peak"; Tenstreet analysis of approximately 5,000 carriers, June 4, 2026 — new applications after an average of 3.8 months, referred drivers 12 times more likely to be hired; BLS JOLTS, June 2026, released August 4, 2026 — quits rate 4.5% accommodation and food services, 2.0% total nonfarm; Work Institute first-year turnover page (undated) — "over one-third of newly hired employees quit within their first year"; SHRM, average cost per hire "nearly $4,700," April 11, 2022. Boostpoint publishes no retention data of its own; every cadence step is practice, not a measured result. Boostpoint figures from the 2026 Social Job Advertising Benchmark — 891 campaigns and 1,334 campaign-months, costs inclusive of campaign management: conversion explaining 70% of cost variation; median $13.88 per applicant, $8.02 volume-weighted, percentiles $2.91 and $66.45. The calculators compute on the reader's inputs. Cost per applicant is not cost per hire; we do not track hires and publish no cost-per-hire figure. A sample of Boostpoint campaigns, not an industry-wide study.