Applicant response rate to text messages: 4 in 5 reply, most within a minute

What happens when a person applies for an hourly job and gets a text within minutes?

They answer it. Four in five replied to a single text. The median applicant wrote back in 42 seconds. Half the screens that reached a verdict, nine questions covering the shift, the site, and a background check, were done inside five minutes.

In Indeed’s 2023 ghosting survey, 61% of U.S. job seekers who admitted to ghosting had done it to two to four employers in the past 12 months. In SHRM’s 2025 Talent Trends, 41% of HR professionals at organizations having trouble recruiting said candidates ghosted them during the interview process.

Ghosting happened here too. 395 applicants never answered, and 349 threads went silent partway through the questions. There is no comparison group on this page, so it cannot say whether texting prevented anything. What it can say is when the ghosting happened: inside the first 15 minutes, before anyone had scheduled an interview or held a shift open. The surveys above measure ghosting after days of waiting, at a stage where it costs the employer something.

The employer is a commercial cleaning company based in Bedford, Ohio, hiring part-time evening cleaners for buildings around Cleveland, Columbus, and Cincinnati. Each new application in SmartRecruiters triggers one text, sent in the company’s name within a minute, asking whether the applicant is still interested. Between July 1 and September 25, 2026, 2,185 people got that text.

Four in five applicants answer the first text

1,790 of 2,185 applicants replied. That is an SMS recruiting response rate of 81.9%, for a message that arrived minutes after they applied and asked one question.

Most answered right away. Two in three replied within 15 minutes. The rest trickled in over the day, the person who applied at 1am and answered at breakfast. Nine in ten of those late replies came the same day. Very few came after that.

18.1% never answered. That is the ghosting rate for this moment: one text, minutes after the application, no follow-up.

Nobody in this data got a second message. Once a second and a third text have run for a full month, we’ll report the reply rate.

How fast job applicants respond to a text

Among applicants who replied while the thread was open, the median first reply came 42 seconds after the text landed. Three in five answered inside a minute. Almost everyone who was going to answer that sitting had done so within ten.

This is faster than the “3 minutes” figure that circulates in SMS marketing decks, and the reason is the time to first contact. The text arrives while the applicant is still holding the phone they applied on. They have not moved on to anything else yet.

How long a text message screening takes

Among the 1,066 threads that reached a pass or fail in the same sitting, the median screen took 4.7 minutes, from the first text to the agent’s last message.

A quarter ended in under 3 minutes. 80% ended inside 10. The applicant sent a median of 9 messages.

Those 9 messages are the screen. The applicant answered each question in a median of 18 seconds, and the next one went out the moment the answer landed. Across 8,614 answers, nobody on either side was waiting for the other.

Recruiters who run the same screen by hand put it at about 20 minutes. That is their estimate, not a measurement; if you have measured it, I would like to see the number. The difference is not the applicant. It is the gap between the answer and the next question, and how many applicants are still there when it arrives.

Six in ten screens ran when no recruiter was at a desk

635 of the 1,066 completed screens started outside Monday-to-Friday, 9-to-5 Eastern. That is 59.6%.

Weekday evenings alone were 237 screens, more than half the daytime total. Weekends were 209. The 189 that started between 10pm and 9am on a weekday are the applicants who applied after a shift, or before one, and were screened before anyone in the office had seen the application.

In total the 1,066 screens were 7,304 minutes of conversation, about 122 hours, across 87 days. 67 of those hours fell outside 9-to-5.

Why applicants fail a screening

The screens recorded 1,956 verdicts in the window, 1,600 passes and 356 fails. That is a pass rate of 81.8%.

It is a wider count than the 1,066 same-thread verdicts above, which were 53.8% of the 1,983 standard first texts. It adds 377 verdicts from first texts that used different opening wording and 513 from threads the applicant started, mostly late replies. Another 349 threads got a reply but no verdict because the applicant stopped answering. Those are excluded from the pass rate.

The pass rate is high because the applicants already chose this job. The fails show what the listing did not screen for.

Schedule is the biggest recorded reason by a wide margin. “Declined after hearing the details” is a separate thing, and the agent’s notes say what the detail was: pay in 27 of the 66, the commute or the site’s location in 19, the shift or the hours in 8, and no single reason in 12. “Background check” means the applicant disclosed a record or would not confirm they could clear the required check.

“Other” is mostly screens the agent judged too thin to pass. In 23 of the 40, the applicant said yes to everything but never actually answered the transportation or background questions. The remaining 17 are one-off requirements: a flu shot at a medical site, lifting 50 pounds, being under 18, no computer for the onboarding quizzes.

Experience was the reason for only 26 of 356 fails. For an evening cleaning role, whether someone can be at the building at 9:30pm on weeknights matters more than whether they have cleaned before.

Reply rate by hour of application

The text goes out when the application arrives, so there is no choosing the hour. But the hour of the application predicts the reply. Applications that came in between 10pm and 6am Eastern were answered within 15 minutes 52.1% of the time. The three daytime blocks ran 64.7% to 67.2%.

Weekday and weekend applications replied at nearly the same rate: 64.0% of 1,586 on weekdays, 62.5% of 397 on weekends.

People apply at 1am. They just do not always want a conversation at 1am. Overnight applicants were the most likely to come back later, usually the next morning: 26% of overnight first texts were answered in a later thread, against 16% to 18% for the daytime blocks. They were also the most likely to never answer, 21% against 17% to 18%. That later-reply group is who a next-morning text would reach. The reply rate after that text is the number this page does not have yet.

Method

Every number on this page comes from one employer’s SMS threads, sent and logged by TextWork.

• First text. The automated message sent when SmartRecruiters reports a new application. 2,537 of them went to 2,185 applicants, July 1 to September 25, 2026. 255 applicants got more than one, usually for a second job.
• Reply. Any message from the applicant. System events, such as a signed document arriving, do not count. Reply rates count applicants: a person counts as replied if they answered any of their first texts. Reply speed, screen length, and the hour-of-day tables count threads.
• Within 15 minutes. A thread closes after 15 minutes with no reply. “Replied within 15 minutes” and “replied later” are defined by that timeout.
• Late-reply timing. Measured from thread timestamps. Accurate to within about an hour, not to the minute. Late-reply status could be matched for 1,925 of the 1,983 standard first texts; the by-hour later-reply shares use that set.
• Standard wording. 1,983 of the 2,537 first texts used the same opening message. Reply speed, screen length, and the hour-of-day tables use only those, so they compare like with like. The other 554 used earlier or job-specific wording and are counted in reply rates and verdicts but not in timing.
• Screen length. Runs from the first text to the agent’s last message in the thread, not to the report, which is filed after the thread closes. Computed on the 1,066 standard threads that ended in a pass or fail in that sitting.
• Answer time. The gap between the agent sending a question and the applicant’s next message, within those 1,066 threads. 8,614 gaps. “Outside 9-to-5” means the first text went out on a weekend or on a weekday before 9am or after 5pm, Eastern Time.
• The 20-minute figure. An estimate from hiring managers and recruiters who have run text screens manually. It is not measured and not from this dataset. It is on the page as the comparison a reader will want, labeled as what it is.
• Hours. Eastern Time, the applicants’ zone. Morning is 6am to noon, afternoon noon to 5pm, evening 5pm to 10pm, overnight 10pm to 6am. The hour is when the application arrived and the text went out, which are the same minute.
• Verdicts. The pass or fail the screening agent recorded at the end of a thread, July 1 to September 25, 2026. A pass means the applicant confirmed the shift, the site, transportation, and that they can clear a background check. A fail carries one reason from a fixed list. The sub-reasons given for “declined” and “other” are my reading of the agent’s free-text notes, not a recorded field.
• No-verdict threads. 349 threads got a reply and then went silent before a verdict. 210 are filed as “no response” (the applicant timed out mid-screen) and 139 as “incomplete screen” (stalled at a specific question). On September 25, 110 records were re-labeled from the first bucket to the second so the stall point is recorded. Neither bucket is a verdict, and neither is in the pass rate.

This is one company, one industry, one state, and one channel. It is not a benchmark for cold outreach, for salaried roles, or for texts sent days after an application. It is a measurement of a single, common moment in hourly hiring.

If you are writing about this and need a cut the page does not show, by hour, by job, by city, or the underlying counts, email [email protected]. Replies go out within one business day. The employer’s name and the pay rate for these roles are not published here; both are available to reporters on request.