"We Rejected 500 Candidates in 30 Days": A Hiring Manager's Honest Account

This account is based on a real conversation with a hiring manager at a B2B SaaS company. Details have been lightly changed to preserve anonymity.

---

In Q1 of last year, we had three roles open simultaneously: two senior engineers and a product manager. We received just over 500 applications in the first 30 days. We made three hires. We rejected, by my count, 497 people — most of whom deserved better than what we gave them.

I'm writing this because I think the honest account is more useful than the version where the hiring team is the hero. We made a lot of mistakes. Some of them cost us good candidates. Some of them damaged the company's reputation in ways I'm still not sure we've recovered from.

What We Got Wrong: The Acknowledgment Problem

We had an ATS. It was configured to send an automatic acknowledgment when an application was received. What we didn't realize was that the email was going to spam for roughly 40% of applicants due to an authentication issue none of us had checked. Hundreds of people applied and heard nothing — not even a confirmation that their application had been received.

When candidates have no confirmation, they follow up. We received over 60 follow-up emails in the first two weeks, all politely asking if their application had gone through. We answered about half of them. The other half decided we weren't interested and moved on.

Lesson one: audit your ATO's transactional emails from a real external address before you open a role. Send yourself the application flow. Make sure it works.

What We Got Wrong: The Timeline Drift

We told candidates in the job description that we'd provide a response within two weeks. Our actual time to first response was closer to five. Three candidates who progressed to the technical screen stage accepted other offers while they were waiting for our scheduling email.

None of this was intentional. We just didn't have buffer built in for what the volume actually looked like. We assumed it would be manageable and didn't create explicit triggers to escalate when the timeline was slipping.

Lesson two: if you publish a timeline, staff the process to meet it. If you can't meet it, update the JD and communicate proactively to candidates in process.

What We Got Wrong: The Rejection Message

Our automated rejection email was three sentences. It said we appreciated the candidate's interest, we'd found a profile more closely aligned with our needs, and we wished them well. It was the same for everyone — from candidates who had completed two rounds of interviews to candidates who had applied and not been screened.

I know this is standard practice. It still feels wrong when I think about the senior engineer who did a four-hour technical assessment for us and received the same form rejection as someone who'd been screened out in 30 seconds.

We changed this. Candidates who completed interviews now receive a message that references something specific about their work and gives at least a high-level reason for the decision. It takes an extra few minutes per rejection. It's worth it.

What We Got Right

We ran structured interviews. Every candidate at the interview stage was evaluated against the same criteria by multiple people before anyone shared their opinion. We didn't love every hire decision — sometimes the structured scores pointed somewhere different than our instincts — but our instincts have been wrong before in ways we've paid for.

We also ran candidate surveys after the process. The feedback was how we found out about the email authentication issue and about the timeline drift. Building a feedback loop in before you need the data is the thing I'd tell any hiring manager to do first.

The Number That Still Bothers Me

497 rejections in 30 days. Even at two minutes of quality per rejection — which is less than most candidates deserve — that's 16 hours of time we didn't invest. We invested zero, because the volume felt unmanageable.

What we've learned is that "unmanageable" is usually a process problem, not a volume problem. The right templates, the right triggers, the right automation in the right places — these don't eliminate the human work of treating candidates like people. They make it possible at scale.

---