Early Attrition Isn't an Onboarding Problem. It's a Hiring Problem.

When an engineer leaves within 6 months of joining, companies almost always blame onboarding. Sometimes, onboarding is the problem. But in my experience, most early attrition traces back to something that happened - or didn't happen - long before day one. The hiring process set expectations that reality couldn't meet.

The expectation gap

This is the number one cause. The candidate was told one thing during the interviews and experienced something different after joining.

The work was different

"You'll be building our next-generation recommendation engine" became "you'll be maintaining the current one and fixing bugs while we figure out the v2 roadmap." The future tense in job descriptions is the biggest source of expectation gaps. If the work doesn't exist yet, say so explicitly.

The team was different

"You'll be working with a strong team of 8 engineers", - but 3 were about to leave, and the remaining 5 were overloaded and unable to help with onboarding. The team the candidate evaluated during the interview isn't always the team they join.

The manager was different

The candidate met the VP of Engineering during the interview, who was charismatic and visionary. Their actual manager - the one they interact with daily - is overwhelmed and not particularly good at people management. The candidate felt sold by the VP and abandoned by the EM.

The culture shock

Process chaos

Engineers coming from structured environments are unprepared for the lack of process at early-stage companies. No sprint planning, no code review standards, deployments whenever someone feels like it. If you hired someone from a structured environment, acknowledge the cultural difference during onboarding.

Decision-making dysfunction

Decisions get made, unmade, and remade weekly. Engineers build features that never ship. This is particularly common at founder-led startups where the founder is still learning to delegate. Engineers who need clarity and direction will burn out and leave.

What the hiring process gets wrong

Selling over evaluating

When every interaction is about convincing the candidate to join, there's no room for unfiltered conversation about challenges. The interviewer who says "the work is amazing, the team is incredible" isn't helping the candidate make a good decision. They're setting up a disappointment.

Not involving the direct manager

If the candidate doesn't meet and have a real conversation with their future day-to-day manager during the process, they're making a career decision with incomplete information. The manager should be part of the interview as a genuine interaction, not a formality.

Not discussing what the first 90 days look like

If you can describe, concretely, what the engineer's first month looks like - what they'll work on, who they'll work with, what they'll learn - the candidate can evaluate whether it matches their expectations. If you can't describe it, you're not ready to hire for the role.

The financial trap

There's a pattern I've seen enough to name: the "golden handcuffs quit." A candidate joins primarily for the salary hike. The work doesn't excite them. Within 2–3 months, the money doesn't compensate for daily dissatisfaction. They start interviewing again from a higher base and leave by month 5. The company offered money without ensuring alignment. The candidate accepted money without ensuring interest. Both are optimised for the wrong variable.

What to do about it

Be transparent during hiring

"The team is going through a transition right now. Two people left recently, and we're rebuilding. Your first 3 months will involve some catch-up before you get to build new things." A candidate who accepts this offer knows what they're signing up for.

Have a structured 30-60-90 day plan

Not a checklist - a genuine plan. By day 30, you'll have shipped your first change. By day 60, you'll own a specific area. By day 90, you'll be fully ramped. When something concrete exists, both sides have milestones, and the manager has triggers for early intervention if things aren't going well.

Conduct stay interviews

At the 30 and 60 day marks, have a simple conversation: "How's it going? Is the reality matching what we described? What's surprising you?" These conversations cost 20 minutes and surface problems while they're still fixable. By the time you do a formal 6-month review, the unhappy engineer has already started interviewing.

Early attrition is expensive. But it's also a signal. If you're losing engineers in the first 6 months, the problem almost certainly starts before day one. Look at your interview process. Look at what you're promising. Look at what you're delivering. The gap between those two things is where the attrition lives.