Top 40 Google Interview Questions (2026)

Google grades every candidate on four attributes: general cognitive ability (GCA), role-related knowledge, leadership, and Googliness. Every question you get — whether it sounds technical, behavioral, or hypothetical — is feeding a score on one of those four. Prepare by attribute, not by memorizing individual questions, and you’ll be ready for variations you’ve never seen.

How Google Actually Grades You

Each interviewer scores you independently on a 1–4 scale against the four attributes, then writes detailed feedback with specific evidence — quotes, examples, reasoning. That packet goes to a hiring committee: Googlers who were not in your interviews and who make the hire/no-hire recommendation based only on the written record. Google’s own re:Work materials describe this structure publicly, and it has two implications:

1. Your answers need to be quotable. Vague answers produce vague notes, and vague notes die in committee. Specific numbers, named trade-offs, and clear decisions survive.
2. One weak interview isn’t fatal, but one “no evidence” attribute is. If nobody could gather evidence of leadership, that gap no amount of charm fixes.

Interviewers use structured interviewing — the same questions and rubric across candidates for a role — which means the categories below are stable and preparable.

GCA: General Cognitive Ability Questions

GCA is not brainteasers (Google retired those after finding they predicted nothing). It’s how you structure ambiguous problems: clarify, decompose, weigh trade-offs, land on a defensible answer.

1. How would you improve Google Maps for a user in a city with unreliable GPS?
2. Our product’s user engagement dropped 15% last quarter. Walk me through how you’d investigate.
3. You have two competing feature requests and resources for one. How do you decide?
4. How would you estimate the number of daily YouTube uploads, and what would you do with that number?
5. A metric you launched is being gamed by users. What now?
6. Design a system for a library to track overdue books, then tell me where it breaks at 100x scale.
7. Tell me about a time you had to make a decision with incomplete information.
8. If you joined this team tomorrow, what would you do in your first 30 days to figure out what matters?
9. How would you decide whether Google should enter a new market?
10. Walk me through a complex problem you solved. What made it complex?

How to answer: think out loud, state assumptions, and structure before you solve. The rubric rewards process over the final number.

Role-Related Knowledge Questions

These vary by function — for engineers, coding and system design — but the pattern is constant: depth on what you claim, honesty about what you don’t know. Below are the cross-functional versions.

11. What’s the hardest technical (or domain) problem you’ve solved in the last two years?
12. Explain a complex concept from your field to me like I’m a smart non-expert.
13. How do you stay current in your field? Give me an example of something you learned recently and applied.
14. Tell me about a project where your first approach was wrong. How did you find out?
15. What tools or methods do you use that most of your peers don’t?
16. Critique a product or system you know well. What would you change and why?
17. Describe a time you had to learn a new domain quickly to deliver.
18. What’s a strong opinion you hold about how work in your field should be done?
19. Tell me about a time your expertise was challenged by someone senior. What did you do?
20. Where does your knowledge run out? What’s adjacent to your role that you’d need help with?

Question 20 is a trap for bluffers: interviewers probe until they find the edge of your knowledge, and candidates who acknowledge the edge score better than those who improvise past it.

Leadership Questions (Emergent Leadership)

Google explicitly looks for “emergent leadership” — stepping up without the title, and just as importantly, stepping back when someone else is better positioned. You do not need to have managed people.

21. Tell me about a time you led a project without formal authority.
22. Describe a time you saw a problem nobody owned and did something about it.
23. Tell me about a time you had to influence a decision above your pay grade.
24. Give me an example of stepping back and letting someone else lead. Why?
25. Tell me about a time you disagreed with your team’s direction. What did you do?
26. How have you helped a struggling teammate?
27. Describe a time you had to rally people around an unpopular decision.
28. Tell me about a time you took ownership of a failure that wasn’t entirely yours.
29. How did you handle a situation where two stakeholders wanted opposite things?
30. What’s the hardest feedback you’ve ever delivered?

Googliness Questions

Googliness sounds fuzzy, but the rubric is concrete: comfort with ambiguity, intellectual humility, bias to action, caring about users and colleagues. It is not “being quirky.”

31. Tell me about a time you were wrong about something important. How did you find out and what did you do?
32. Describe a time priorities changed suddenly. How did you adapt?
33. Tell me about a time you did something for a user or customer that wasn’t in your job description.
34. When did you last change your mind because of someone else’s argument?
35. Tell me about a time you worked with someone very different from you.
36. Describe a situation where the right thing to do conflicted with the easy thing.
37. Tell me about a failure. What did you learn, and what did you change afterward?
38. How do you handle working on something when the goal keeps shifting?
39. Tell me about a time you gave credit to someone else for work you contributed to.
40. What do you do when you receive feedback you disagree with?

Three Full Example Answers

”Tell me about a time you led a project without formal authority.” (Leadership)

“At my last company, our onboarding docs were so outdated that new hires were pinging senior engineers daily with the same questions — I counted 14 interruptions in one week. Nobody owned documentation, so I proposed a two-week docs sprint in our team channel. I built the outline, split it into eight chunks, and asked people to claim one — I took the two nobody wanted. To keep it moving without being anyone’s boss, I ran a 10-minute Friday check-in and publicly credited every completed section. We shipped in 12 days, and new-hire questions to seniors dropped about 70% the following month. The lesson: authority wasn’t the blocker — nobody had made the work small and claimable.”

Why this scores well: unprompted ownership, a mechanism, a quantified result, a generalized lesson — all easy to write down.

”Tell me about a time you were wrong about something important.” (Googliness)

“I pushed hard to build an in-house analytics dashboard instead of buying a vendor tool — I estimated six weeks and argued the vendor cost was unjustifiable. Three months in, we were still building, and the two engineers on it were pulled from roadmap work. A teammate showed me a cost breakdown proving the vendor would have been cheaper even at double the license price. She was right and I said so, in the same meeting. I recommended we kill the project, wrote the post-mortem myself, and we migrated to the vendor in three weeks. Since then, every build-vs-buy call I make includes a written estimate of fully loaded engineering cost, reviewed by someone incentivized to disagree with me. The dashboard was a bad call; the bigger error was framing the debate so nobody wanted to challenge me.”

Why this scores well: a real, costly mistake, fast public reversal, and a durable process change — textbook intellectual humility.

”Our product’s user engagement dropped 15% last quarter. Walk me through how you’d investigate.” (GCA)

“First, I’d verify the drop is real: tracking changes, metric definition changes, seasonality against last year. If it’s real, I’d segment — platform, geography, cohort, acquisition channel — because a uniform 15% drop and a 50% drop in one segment are different problems. Then I’d overlay a timeline of changes: our releases, pricing, competitor launches, platform policy. If the drop correlates with a release, I’d look at funnel data around the changed surface and consider a holdback test to confirm causation. If nothing internal explains it, I’d talk to churned users directly — ten interviews beat another dashboard. Throughout, I’d distinguish ‘engagement dropped’ from ‘value dropped’: if the release made a task faster, sessions falling might be fine. My deliverable: a ranked list of hypotheses with evidence and the cheapest next test for each.”

Why this scores well: structure first, and it questions the metric itself — a top-of-rubric move.

How to Prepare in the Two Weeks Before

• Build 8–10 stories, not 40 answers. One strong story about an ambiguous project can answer a leadership, Googliness, or GCA question depending on emphasis. Index your stories by attribute.
• Quantify everything. “Improved onboarding” becomes “cut new-hire ramp questions 70%.” Committee packets run on numbers.
• Practice out loud, under pressure. Reading answers silently doesn’t prepare you to deliver them; rehearse with a friend, or use Alex, an AI interview coach to run realistic mock interviews with follow-ups until your stories come out tight and quotable.
• Prepare for the drill-down. Google interviewers probe: “Why that decision? What would you do differently?” If your story collapses one layer deep, it wasn’t ready.
• Ask your recruiter what the loop covers. They’ll tell you the structure and attributes for your role. Use it.

FAQ

How many interviews are in the Google process?

Typically a recruiter screen, one or two phone/video interviews, then a loop of four to five 45-minute interviews. The hiring committee then reviews written feedback and makes the recommendation, which is why the process often takes four to eight weeks.

Does Google still ask brainteasers like “how many golf balls fit in a bus”?

No. Google publicly dropped brainteasers after internal research showed they didn’t predict job performance. Estimation questions still appear, but they’re graded on structure and reasoning, not the number.

What is Googliness, in plain terms?

Comfort with ambiguity, intellectual humility, bias to action, and caring about users and teammates. Show it with stories where you admitted a mistake, adapted to sudden change, or helped someone with no upside for yourself.

Can I reuse the same story in multiple interviews?

Across different interviewers, yes — a strong story told to two of five interviewers is fine even though notes are compared in committee. Within a single interview, give each question fresh evidence.