Bombed One Round of Your Onsite Interview? Reset Before the Next One

If you bombed one round of your onsite, the loop is usually still open. Many loops collect each interviewer’s feedback separately. In those, the next interviewer hasn’t read a word about your last round. So the bad hour only spreads if you carry it in with you. Use the break to write down what went wrong in three short lines. Then change one habit for the next round, like saying your approach before you code. Don’t mention the bad round to the next interviewer. They’d have no reason to know about it. Mentioning it tells them. After the loop, write to your recruiter only with a fact they can act on. A dropped call or a crashed editor counts. An apology for a weak answer doesn’t. Your own read of the round is also shaky. Candidates often think they did worse than they did. This post follows one Android engineer through a four-round virtual onsite. It runs from the bad first round to the email afterward.

Each interviewer writes feedback before reading anyone else’s

The fresh start is built into how many companies collect feedback. The Holloway Guide to Technical Recruiting and Hiring is written by Osman (Ozzie) Osman and other contributors. Its section on interviewer feedback was last updated in August 2022. Its rule is firm. The guide says “no interviewer be exposed to other interviewers’ feedback before they have submitted their own.” The reason it gives is bias. Seeing other notes can tilt an interviewer’s judgment.

Not every company follows that rule. Some interviewers do talk in the hallway. But in a loop that follows it, your second interviewer walks in blind. They know the role, the question and their rubric. They don’t know you stalled for 25 minutes an hour ago.

So each round produces its own written record. The decision comes later, when someone reads the records side by side. Our post on why you can be rejected after a good interview covers that later step. The same mechanism can also work for you. Three strong records can carry one weak one.

Your own score for the bad round is probably wrong

Gayle Laakmann McDowell was an interviewer at Google and wrote Cracking the Coding Interview. In a December 2011 post, she argued that one bombed interview doesn’t rule out an offer. Her reason was that rounds aren’t scored as right or wrong. Interviewers weigh qualities like speed and how you handle bugs. She also wrote that candidates often misjudge which round went worst. The post is old. Its point about self-assessment has data behind it, though.

Aline Lerner at interviewing.io studied it. She compared how candidates rated their own mock interviews with how interviewers rated them. Her analysis uses data as of August 2016 and has no visible publish date. It found only a weak link between the two. Candidates underrated themselves roughly twice as often as they overrated themselves. Interviewing.io sells interview practice, so it has an interest in this topic. The finding still matches what McDowell saw.

So treat “I bombed it” as a guess. It might be right. It’s still a weak basis for the rest of your day.

The running example: one stalled coding round in a four-round loop

Take an Android engineer we’ll call Sam. Sam has a virtual onsite for a mid-level role. There are four rounds. Round one is coding. Round two is Android system design. Round three is a second coding round. Round four is behavioral.

Round one goes badly. The problem is a graph search. Sam starts with a greedy approach and codes it for 25 minutes. Then a test case breaks it. The interviewer hints at breadth-first search. Sam rewrites it with five minutes left. The code runs on one example. There’s no time for edge cases or the complexity discussion.

Sam has 15 minutes before round two. The rest of this post is what Sam does with them.

Use the 15-minute break to write three lines and change one habit

Replaying the round in your head won’t fix it. Writing it down ends the replay. So Sam opens a notes file and writes three lines.

I coded greedy for 25 minutes before testing it on a hard case.
That cost me the edge cases and the complexity talk.
Next round, I say the approach and one breaking case before I write code.

The first two lines are for later. They’ll matter when Sam talks to the recruiter or plans practice. The third line is for today. It turns the mistake into a habit Sam can use in the next round.

Then Sam closes the file. With the minutes left, Sam stands up, drinks water and rereads the job posting. That last part points attention at round two. The system design round will ask about Android architecture, not graphs.

The changed habit should fit every remaining round. “Test before you commit” works in system design too. There, it means naming one failure case before you pick a design.

Don’t tell the next interviewer about the bad round

An apology at the start of round two feels honest. It’s usually a mistake. The interviewer had no information about round one. Now they do. And it’s negative. If the loop shares notes later, a debrief reader might also see it in their write-up.

Holloway’s guide makes a related point about uncertainty. It says “there is no such thing as a neutral opinion” in hiring feedback. The guide treats an unsure interviewer’s view as a no hire. A candidate who opens with self-doubt gives them a reason to be unsure.

So Sam opens round two the way they’d open any round.

Thanks for having me. Before I start on the design, can I ask two questions about scale? How many users should I plan for? Does it need to work offline?

Nothing in that opening mentions round one. It moves straight to the scope questions that system design rewards. Our post on running out of time in a system design interview has the checkpoints.

If you bombed one round, how badly still matters

A fresh start in round two doesn’t erase round one. Its record still goes to whoever decides. A mild miss, like Sam’s, is a weak data point among four. A round that shows you don’t know the basics of the role is heavier.

Some loops also give one interviewer extra weight. At Amazon, the bar raiser has that role. Our post on the Amazon bar raiser interview explains why guessing which round that is backfires. You can’t control which round went badly. You can control the next three.

Mixed results aren’t always a final answer either. Holloway notes that comparing feedback can surface open questions. Those can lead to another interview or a call with the hiring manager. The company decides whether to offer that. It’s still one more reason not to give up on round two.

After the loop, write to the recruiter only with a fact they can use

Sam finishes the day and wants to send an email explaining round one. That email rarely helps. The interviewer’s feedback is already written. A note saying “I know BFS, I just froze” can’t change it. It adds nothing the recruiter can act on.

The case changes when something outside your control broke the round. Say Sam’s audio had dropped for ten minutes in round one. That’s a fact the recruiter can check and act on. Send it the same day.

Hi Maya, thanks for organizing today’s loop. One thing to flag for the first coding round. My audio dropped for about ten minutes around 10:20. We lost time reconnecting. I wanted you to have that context. If a shorter follow-up round would help, I’m glad to do one.

In Sam’s real case, nothing broke. So Sam sends a short thank-you and asks about the timeline. The three-line note goes into practice instead. Two weeks of drills on testing an approach before coding it are a better use of it. After a no, our post on how to reapply after a failed interview is the next step.

Why each part of Sam’s day works

The three-line note stops the replay. It also turns a feeling into one habit Sam can change in the next 15 minutes.

The new habit is general. “State the approach and a breaking case first” helps in coding and design alike. A fix that only applies to graph problems would help no one that afternoon.

The round two opening protects the fresh start. The interviewer forms their own view from what Sam does in that room.

The recruiter email follows the same rule. It carries a checkable fact and an offer to help. A defense of a weak answer would give the recruiter nothing to act on. If there’s no fact, the email stays a thank-you.

1 thought on “Bombed One Round of Your Onsite Interview? Reset Before the Next One”

Leave a Comment