How to prepare for an interview comes down to three things done in order: work out which round you are in and what it tests, build a short prep sheet from the job posting itself, and rehearse your answers out loud before the day. Most guides stop at "research the company and be on time". The harder truth is in the data: only about 35% of candidates pass the recruiter screen, according to Ashby's 2026 benchmarks, and at Ashby's customers a technical hire now takes a hiring team 17.6 interviews on average. This guide gives you a timed plan for each round, one week out, the day before and the hour before, and a prep sheet you can copy.
Your prep plan at a glance
Copy this table into your notes when the invitation arrives. Each row is expanded in its own section below.
| Round | One week out | The day before | The hour before |
|---|---|---|---|
| Recruiter screen | Write a 60 to 90 second answer to "tell me about yourself"; decide your pay range and how you answer work authorization | Reread the posting; check the time zone and who is calling whom | Have the posting, your resume and your range on one screen |
| Hiring manager | Build the prep sheet; draft five or six stories in STAR form; read what the team ships | Rehearse each story aloud once, under two minutes | Reread the top three requirements and your three best stories |
| Technical, case or skills | Practice the exact format under a timer, ideally with another person | Light review only; confirm tools, logins and allowed materials | Warm up with one easy problem; test camera, audio and screen share |
| Panel or onsite loop | Map stories to interviewers so none repeats; prepare a different question for each person | Plan travel or tech, meals and breaks | Read the schedule and each interviewer's name and role |
| Recorded or AI-led | Do any practice questions the platform offers | Test the link, lighting and audio on the device you will use | Close other apps; put notes at eye level, not in your lap |
How many rounds to expect, and what the data says
There is no standard number. The best evidence comes from two places: hiring platforms that measure their customers' pipelines, and the few employers that publish their own process.
Start with where people drop out. Ashby, an applicant tracking system, analyzed "over 54M applications and 93K jobs from January 2021 through March 2026" for its recruiting operations benchmarks. Recruiter screens "see lower passthrough rates (~35%)", while "later stages, such as post-onsite and offer, see significantly higher conversion rates at 95% and 81%, respectively." Its chart shows 24% at the onsite stage itself. Read plainly: the first two conversations are where most candidates are cut, so they deserve real preparation, not a quick look at the company's About page.
Then look at how much interviewing a hire takes. Ashby's 2026 recruiter productivity report counts every interview a hiring team runs, across every candidate, for each person it hires:
| Role type (Ashby customers) | Interviews per hire |
|---|---|
| Data | 19.5 |
| Product management | 18 |
| Engineering | 17.9 |
| Technical roles overall | 17.6 (about 11 in 2021) |
| HR | 14.3 |
| Business roles overall | 11.7 (about 8 in 2021) |
| Customer support | 9.5 |
These are not the rounds you personally sit. They are the total effort behind one hire, spread across the candidates who were interviewed and turned down. Gem, another hiring platform, reports the same direction in its 2026 benchmarks, built from over 165 million applicants and 1.2 million hires: "The average number of interviews per hire has increased by 33% since 2021." More interviews per hire.
Few large employers publish a count. Amazon is the exception. Its research scientist interview prep page says "you will have one to two technical phone screenings", each 60 minutes with a senior leader on the team, and then "your loop will include four 55-minute interviews". A separate Amazon page says candidates meet "anywhere from two to seven Amazonians during the interview loop stage". Google and Meta publish no interview count, so any number credited to them on a blog is a guess.
The interviews also take time end to end. Ashby puts the median from a hired candidate's first interview to their last at 17.9 days for technical roles and 14.4 days for business roles. Plan your preparation in that window, not just the night before.
A prep sheet built from the job posting
The posting is the closest thing you have to the interviewer's scoring sheet. Many employers use structured interviews, where every candidate gets the same job-related questions and answers are rated against set criteria. They are also the best-supported interview format: a 2022 review of selection research by Sackett, Zhang, Berry and Lievens found that structured interviews had the highest mean validity of the methods studied, r = .42, as summarized by SIOP. The criteria come from the job, so the job description tells you what they are.
Build the sheet in four columns:
- Requirement. Copy each required and preferred qualification from the posting, word for word. Add any duty that appears twice.
- Likely question. Turn each one into the question you would ask if you were hiring. "Experience with stakeholder reporting" becomes "Tell me about a report that changed a decision."
- Your evidence. One story or fact that proves it, in STAR form: situation, task, action, result.
- The number. A figure from the result: time saved, error rate, revenue, users, size of the team. Amazon tells candidates to "include metrics or data where applicable", and most interviewers will ask for one anyway.
Here is a short sample, written for this post. The posting and the candidate are invented.
| Requirement (from a sample operations analyst posting) | Likely question | Evidence (sample) | Number |
|---|---|---|---|
| "Build and maintain dashboards in SQL and Tableau" | Walk me through a dashboard you built from scratch | Rebuilt the weekly returns report from spreadsheets into one SQL view and a Tableau dashboard | Report prep went from about 4 hours to 20 minutes a week |
| "Partner with warehouse leads to improve throughput" | Tell me about working with a team you did not manage | Ran a two-week pilot with two shift leads on pick-path changes | Picks per hour up 9% on the pilot shift |
| "Comfortable with ambiguity" | Describe a time the goal was unclear | Wrote a one-page problem statement before starting a pricing analysis and got sign-off | Cut the project from three rounds of revisions to one |
| "Clear written communication" | How do you explain data to non-analysts? | Monthly summary email with one chart and three sentences | Read by the regional director, who asked for it to continue |
Then write five or six stories that can each cover more than one row: a success, a failure and what changed after it, a disagreement, leading without authority, a tight deadline, and learning something fast. Amazon's own advice is to prepare examples that show how you have "taken risks, succeeded, failed and grown." Our STAR method guide shows how to keep each one under two minutes.
Last, add a fifth column of your own questions. One per requirement you are unsure about is plenty. Good ones ask how the work is measured, what the first three months look like, and what the last person in the role found hardest. Two specific questions beat ten generic ones. For more by round, see questions to ask the interviewer.
The recruiter screen
This is the round most people under-prepare for and the one with the lowest pass rate. A recruiter is usually checking a few things: that you understand the role, that your background matches the basics in the posting, that the logistics work (location, start date, pay, work authorization), and that you can explain yourself clearly.
One week out. Write and time a 60 to 90 second answer to "tell me about yourself". A sample shape: current role and level, the two strongest matches to the posting, one result, and why this job now. Our guide to why do you want to work here has sample answers for that last part. Decide your pay range from market data before anyone asks. Write down how you will answer the work authorization and sponsorship question, and make it match what you put on the application form.
The day before. Reread the posting and your own application. Confirm the time zone and whether they call you or you join a link. For a call, see the phone interview questions recruiters ask.
The hour before. Put the posting, your resume and your range where you can see them. Have one question ready about the team or the next steps.
A sample answer to the pay question, if you would rather not name a number first: "I'm looking at roles in the range of $78,000 to $88,000 base, depending on the full package. Is that in line with the range for this role?" If the posting lists a range, start from it. Our salary negotiation guide has more scripts.
The hiring manager interview
The hiring manager wants to know whether you can do the work they have, with the team they have. This is where the prep sheet earns its keep. If you are asked why should we hire you, the sheet's top three requirements are the answer.
One week out. Build the sheet. Read what the team actually ships: the product, recent releases, the company's last earnings call or annual report if it is public, and the manager's own public writing. Draft your five or six stories, and a short answer to where you see yourself in five years.
The day before. Say each story out loud once, timed. Answers that run long on paper run longer when spoken. Cut background, keep actions and the result.
The hour before. Reread the top three requirements, your three best stories and your two best questions.
In the room, answer the question asked, then stop. If a story does not fit, say so and offer a closer one. When you do not know something, say how you would find out. For the weakness question, our strengths and weaknesses guide has a formula.
Technical, case and skills rounds
These rounds test whether you can do the work, in a format you should practice in advance: a live coding problem, a case, a take-home task, a writing sample, a portfolio review or a role play.
Know the format before you prepare for it. Bain says consulting applicants typically go through "a digital assessment and then case interview(s)", and that its cases "reward precision, but also creativity" (Bain). Amazon tells research scientist candidates "We won't ask brain teasers", and its interview overview says some roles include a writing assignment. Ask the recruiter what the round covers, how long it is, what tools you can use and whether you will share a screen. Recruiters usually answer.
One week out. Practice the exact format under a timer, and talk while you work, because the interviewer scores your reasoning as well as the answer. Do at least one session with another person watching.
The day before. Review lightly. Confirm the coding environment, logins, allowed materials and the take-home deadline if there is one.
The hour before. Do one easy warm-up problem. Test camera, audio and screen sharing.
Panels and onsite loops
A loop is several interviews back to back, often with different interviewers assigned different topics. At Amazon, each research scientist interviewer "typically" asks two or three behavioral questions, and Amazon's interview overview says the loop will likely include a "Bar Raiser" (an objective third party, usually from another team).
The main risk in a loop is repetition. Interviewers compare notes, and the same story told four times looks like you only have one. Map your stories to the schedule so each interviewer hears different ones, and prepare a different question for each person.
One week out. Ask for the schedule and each interviewer's role. Assign stories and questions to each slot.
The day before. Plan travel or your video setup, food, water and the breaks between sessions.
The hour before. Read the names and roles in order. Eat something.
The data has one reassuring point. In Ashby's numbers, once candidates are past the onsite, 95% move on from the post-onsite stage and 81% from the offer stage. Getting through the loop is most of the work.
Video, recorded and AI-led interviews
AI interviews are now common. In a Greenhouse survey released on 1 May 2026, "nearly two-thirds (63%) have now been interviewed by an AI", among 1,200 US job seekers. Seventy percent "were never clearly told upfront that AI would be evaluating them", and 38% "have walked away from a hiring process because it included an AI interview". Of those who completed one, 28% moved forward, 13% were formally rejected and 51% never heard back.
The research on outcomes is thin but not all negative. In a randomized field experiment by Jabarian and Henkel across 70,884 applications to 48 customer service postings in the Philippines, AI-led interviews increased job offers by 12%, and 78% of applicants chose the AI interviewer when offered the choice. Humans still made the hiring decisions. It is one setting, so do not generalize it far.
How to prepare for either kind:
- Treat it like a real interview with a person. HireVue's own candidate tips say to prepare as you would for any job interview, research the job and company, and have short stories about challenges, actions and results ready.
- Use any practice questions the platform offers, and read the invitation for time limits and whether retakes are allowed. That varies by employer. Our HireVue interview guide covers prep time and retakes on that platform.
- Look at the camera, not the screen. Light your face from the front.
- Answer in STAR form and keep to the time limit. A recorded answer cannot be rescued by a follow-up question.
- If you would rather speak to a person, ask the recruiter whether that is an option, and whether a human reviews the result.
- Think twice about tools that feed you answers during a live interview. Our Cluely review covers what employers say about them.
The day before and the hour before
Whatever the round, the last 24 hours are for logistics and rehearsal, not new research.
The day before:
- Confirm the time, time zone, format, link or address, and the names of the interviewers.
- Rehearse your opening answer and your three best stories out loud.
- Print or open the posting, your resume as submitted and your prep sheet.
- For video, test the link on the device you will use, with the same headphones.
- For in person, plan the route with a buffer, and know where you will wait.
- Lay out what you will wear. Our guide to what to wear to an interview covers each industry and format.
The hour before:
- Reread the posting's top requirements and your prep sheet.
- Close everything else on the screen and silence notifications.
- Have water, a pen and paper, and your questions within reach.
- Join a video call a few minutes early. Arrive in person 10 minutes early, not 30.
After the interview
Within the hour, write down the questions you were asked, what you would answer differently, and anything the interviewer said about the timeline. This is the material for the next round, and for the next company if this one says no.
Send a short thank-you note the same day or the next morning, one per interviewer if you have their emails. Our thank-you email guide has templates.
Then wait the time they gave you, plus a couple of days, before following up. Ashby's medians from first to last interview, 14.4 days for business roles and 17.9 for technical roles, are a useful reference for how long the interviews take. What is normal after that, and when silence means a no, is in how long to hear back after an interview.
Keep applying while you wait. With only about a third of candidates passing the first screen, a single process is not a plan.
Interviews come after applications, and ApplyEze works on the application side. It writes a resume for each posting, adds a cover letter when the form asks, and submits applications from your own machine. It has no interview feature. If you used a tailored resume, reread the version you sent before the interview, because interviewers ask about what is on the page. ApplyEze brings the posting's named skills and tools onto your skills line, even when your own record does not list them. There is no step to remove one before a live run, so use a dry run to see what it writes, and be ready to discuss anything on the page.