Most SEO interview question lists are memorisation dumps. Eighty questions, eighty short answers, and no explanation of what the interviewer is actually trying to find out.
That is a problem, because a good interviewer is not checking whether you can recite a definition. They are checking whether you understand why something works, and whether your knowledge is current.
This guide is written for the person being interviewed. Each question comes with a solid answer and a note on what is being tested, plus a section on the questions where the textbook answer is now wrong, which is where most candidates lose the room.
How SEO Interviews Are Usually Structured
Knowing the shape helps you prepare the right things.
- Fundamentals, to confirm you know the vocabulary and are not bluffing.
- Technical, to see how deep you go and whether you can talk to developers.
- Scenario questions, where they describe a problem and watch how you think.
- Tools and measurement, to check you can actually do the daily work.
- Current awareness, to see whether your knowledge stopped updating two years ago.
The last two are where experienced candidates separate themselves. Anyone can learn definitions in a weekend.
Fundamental SEO Interview Questions
1. What is SEO?
Improving a site so it appears for the searches its audience actually makes, through content, technical health and reputation. SEO earns visibility rather than buying it, which is what separates it from paid search.
Being tested: whether you can explain it without jargon. If you cannot explain SEO to a client, you cannot sell your work internally either.
2. What is the difference between on-page, off-page and technical SEO?
On-page is what sits on your pages: content, titles, headings, internal links. Off-page is what happens elsewhere: links, mentions, reviews, reputation. Technical is whether search engines can crawl, render and index the site at all.
Being tested: that you see them as one system. A strong answer adds that technical problems make the other two irrelevant, because content nobody can crawl does not rank.
3. How does Google rank pages?
It crawls pages, indexes what it judges worth keeping, then ranks the indexed set for a query using relevance, quality signals, usability and context such as location and device.
Being tested: whether you understand crawl, index and rank as three separate stages. Candidates who blur them struggle with diagnosis later, because the fix for “not crawled” is nothing like the fix for “not ranking”.
4. What is keyword intent and why does it matter?
Intent is what the searcher wants: to learn, to compare, to buy, or to reach a specific site. It matters because a page that answers the wrong intent will not rank however well optimised it is.
Being tested: whether you check the SERP before writing. The honest answer is that Google has already shown you the intent, in the results it chose to rank.
5. Why does internal linking matter?
Internal links do three jobs: they help search engines discover pages, they pass authority from strong pages to weaker ones, and the anchor text tells Google what the target page is about.
The practical version is that a page nothing links to is a page the site itself is saying is unimportant. That is usually why it sits unindexed.
Being tested: whether you treat internal links as a lever you control. Unlike backlinks, this is entirely in your hands, which makes it the cheapest fix available on most sites.
6. What makes a good title tag?
It describes the page accurately, includes the term the page targets near the front, reads like something a person would click, and is short enough not to be truncated.
Being tested: whether you mention that Google frequently rewrites titles. Candidates who talk about exact character counts as if they were rules are working from old advice.
On-Page SEO Interview Questions
7. Walk me through your keyword research process.
Start from the business, not the tool. Understand what the company sells and who buys it, then build a seed list from the language customers actually use.
Expand with a keyword tool, group by topic and intent rather than by volume, check the SERP for each group to confirm what Google thinks the query means, then prioritise by a combination of intent, achievable difficulty and business value.
Being tested: whether you start with a tool or with the business. Candidates who open Semrush first usually produce keyword lists nobody asked for.
8. How do you write a meta description?
As ad copy, not a summary. It should describe what the page delivers and give a reason to click, using the query language where it fits naturally.
The nuance to add: meta descriptions are not a ranking factor, and Google rewrites them most of the time. They earn clicks, not positions.
Being tested: whether you claim they help rankings. That is a common and easy tell.
9. How should headings be structured?
One H1 that matches what the page is about, then H2s for main sections and H3s beneath them, in order. Headings are for structure and scanability, not for stuffing keywords.
Being tested: whether you talk about accessibility and readers, or only about crawlers. The best answer mentions both.
10. What is keyword cannibalisation and how do you fix it?
Two or more pages competing for the same query, so Google has to choose and neither page gets the full benefit.
Diagnose it in Search Console by filtering to the query and checking the Pages tab. If several URLs appear over time for the same term, you have it.
Fix it by consolidating: merge the weaker page into the stronger one and redirect, or differentiate them so they target genuinely different intents.
Being tested: whether you can diagnose it with data rather than guess from titles.
11. How do you win a featured snippet?
You have to rank on page one first, since Google pulls snippets from results it already ranks. After that, answer the question directly and early, in the format the current snippet uses: a short paragraph for a definition, a list for steps, a table for comparisons.
Look at what is in the box now, because Google has already told you the format it wants.
The part worth adding is the cost. Winning a snippet can reduce your clicks, because the searcher gets the answer without visiting. It is worth having for visibility and for brand, but treating it as free traffic is wrong.
Note too that AI Overviews have absorbed a lot of this territory, so snippets matter less than they did three years ago.
Being tested: whether you know you must already rank, and whether you understand the click trade-off.
12. What is People Also Ask and how do you use it?
The expanding box of related questions in the results. Answer one of them clearly on a page that already ranks and you can appear in it.
Say what it is actually worth, though. A PAA appearance rarely sends much traffic, because the answer opens in place and most people never click.
Its real value is research. PAA is the best free source of the exact questions people ask around a topic, in their own words, and it updates as you click through. It tells you what to write about far more reliably than it sends visitors.
Being tested: whether you treat SERP features as traffic sources or as intent data. The second answer is the more useful one.
13. How do you optimise images?
Descriptive filenames, alt text that describes the image for someone who cannot see it, modern formats, appropriate dimensions, and lazy loading below the fold.
Alt text exists for accessibility first. It happens to help image search, but writing it for keywords rather than for people is doing it backwards.
14. When would you update old content instead of writing new?
When a page already ranks but has slipped, when its facts have aged, or when it sits just off page one and the gap is coverage rather than authority.
Updating an established page is usually faster than earning a new one, because the page already has whatever authority it accumulated.
Being tested: whether you understand that publishing volume is not the only lever.
Off-Page SEO Interview Questions
15. What makes a good backlink?
Relevance first, then authority. A link from a modest site in your field is usually worth more than a stronger link from an unrelated one.
After that: editorial placement inside content rather than a footer or sidebar, a page that itself has some authority, natural anchor text, and a site that does not link out to everyone.
Being tested: whether you lead with a metric. Candidates who answer “high DA” and stop are describing a tool score, not a link.
16. How would you build links for a new site with no authority?
Start with what you can control: reclaim links pointing at your own broken URLs, claim unlinked brand mentions, and get listed where your business genuinely belongs.
Then earn rather than ask. Publish something worth citing, offer expertise to journalists and podcasts, and write for publications your audience actually reads.
Say what you would not do, too. A fixed monthly link quota gets filled with directory and profile links that pass nothing.
17. What is anchor text and how should it look?
The clickable text of a link. It tells search engines what the target page is about, which is why over-optimising it is a risk.
A natural profile is mixed: brand names, URLs, generic phrases like “this guide”, and only some exact-match terms. A profile where most anchors are your target keyword looks built rather than earned.
Being tested: whether you know Google’s spam policies name “links with optimized anchor text in articles, guest posts, or press releases” as a link scheme.
18. How do you analyse a competitor’s backlink profile?
Pull their referring domains, sort by relevance rather than raw count, and look for patterns: the publications that cover this space, the resource pages that list companies like yours, the assets that attracted the links.
The useful output is not a list of links to copy. It is understanding why those links exist, so you can create something with the same reason to be cited.
19. Is guest posting still worth doing?
Yes, with a caveat. Writing for a publication your audience reads, because you have something to say, is fine and always has been.
Producing guest posts purely to place keyword-rich anchors is what Google’s policy describes as a link scheme. The activity is the same; the intent is what differs.
20. What would you do about negative SEO?
Usually nothing, and say so calmly. Google’s systems are built to ignore spammy links pointed at a site, precisely because otherwise anyone could damage a competitor.
Act only if there is a manual action in Search Console, or if you can see real harm you can trace. Panic-disavowing on the strength of a third-party toxicity score is the wrong reflex.
Technical SEO Interview Questions
21. What is a canonical tag and when do you use it?
It tells search engines which version of similar or duplicate pages is the one to index. Common uses are parameter URLs, printer versions and syndicated content.
Being tested: whether you know it is a hint, not a directive. Google can and does choose a different canonical, which is why the Page Indexing report has a “Google chose different canonical” status.
22. What is the difference between robots.txt and a noindex tag?
Robots.txt controls crawling. Noindex controls indexing. They are not interchangeable.
The trap follows: if you block a page in robots.txt, Google cannot crawl it, so it cannot see the noindex tag. To remove a page from the index you must let it be crawled.
Being tested: exactly that trap. It is the single most common technical misunderstanding in the field.
23. What is crawl budget and who needs to care?
The number of pages a search engine will crawl on your site in a given period, shaped by demand and your server’s capacity.
The honest answer includes who it does not apply to. For a site of a few thousand pages it is rarely the constraint, and worrying about it is usually a distraction from a real problem.
Being tested: whether you apply concepts proportionately or repeat enterprise advice at small sites.
24. What is hreflang and what usually goes wrong with it?
Hreflang tells search engines which language and regional version of a page to serve. It comes up for any role touching international sites.
What goes wrong is usually return tags. Every page in a hreflang set must reference every other page in that set, including itself. Miss the return reference and the whole cluster can be ignored.
Being tested: whether you have implemented it or only read about it. The return-tag answer is the tell.
25. What is schema markup and what does it actually do?
Structured data describes your page’s content to search engines in a format they can parse, using vocabulary from schema.org. Common types are FAQ, Product, Article, Recipe, LocalBusiness and Review.
Be precise about the benefit. It does not directly improve rankings. What it can do is make you eligible for rich results, which changes how your listing looks and therefore your click-through rate.
Being tested: whether you say “it boosts rankings”, which is the wrong answer. Eligibility is the right word, since Google decides whether to show a rich result at all.
26. What are Core Web Vitals and how much do they matter?
They measure loading, interactivity and visual stability: Largest Contentful Paint for how quickly the main content appears, Interaction to Next Paint for responsiveness, and Cumulative Layout Shift for how much the page jumps around.
They are a genuine ranking signal, but a weak one compared with relevance. A fast page with the wrong content will not outrank a slower page that answers the query.
Being tested: proportion. Candidates who claim speed is a top-three ranking factor are overstating it, and candidates who dismiss it entirely miss that it affects conversions regardless of ranking.
27. What is the difference between a 301 and a 302 redirect?
A 301 is permanent and consolidates signals to the new URL. A 302 is temporary and tells Google to keep the original indexed.
Use 301 for anything permanent: moved pages, merged content, protocol and domain changes. Use 302 only when the original genuinely returns, such as a page temporarily out of stock.
Being tested: whether you know the consequence of getting it wrong. Temporary redirects left in place for years stop the new URL ever taking over properly.
28. What is an XML sitemap and what should be in it?
A list of the URLs you want indexed, which helps discovery on large or poorly linked sites.
What matters is what you leave out. It should contain only canonical, indexable, 200-status URLs. Including redirects, noindexed pages or 404s sends contradictory signals.
A sitemap is a hint, not a guarantee, and it does not compensate for a page nothing links to.
29. How do you handle a site migration?
Crawl and record the old site first, so you have a baseline of URLs, titles and traffic. Map every old URL to its closest new equivalent, and redirect them one to one rather than to the homepage.
Launch, then verify: check redirects resolve in a single hop, re-crawl for errors, submit the new sitemap, and watch indexing and rankings daily for the first weeks.
Being tested: whether you plan the rollback and the measurement, not just the redirect map.
30. How does JavaScript affect SEO?
Google renders JavaScript, but rendering is a second pass that happens after the initial crawl and can be delayed. Content that only exists after JavaScript runs is discovered later, and sometimes not at all if rendering fails.
The practical check is to view the rendered HTML in Search Console’s URL Inspection, not the browser view, and confirm your main content and links are actually there.
31. What does log file analysis tell you that other tools cannot?
Server logs record what Googlebot actually did, rather than what a crawler simulates or what Search Console samples. That is the whole value.
You can see which URLs get crawled and how often, which get ignored entirely, how much crawl activity is wasted on parameters, feeds and paginated archives, and which response codes Googlebot really receives.
Be honest about scope, though. On a site of a few hundred pages this is rarely worth the effort, and Search Console will tell you enough. It earns its place on large sites, ecommerce catalogues and after migrations, where crawl waste is a real cost.
Being tested: whether you know it exists and whether you apply it proportionately. Claiming you run log analysis on every small site suggests you have not actually done it.
32. What is mobile-first indexing?
Google indexes and ranks based on the mobile version of your page, not the desktop one.
The consequence people miss: if your mobile version hides content, drops structured data, or has fewer internal links, that reduced version is what gets indexed.
33. How would you diagnose a page that is not indexed?
Start with the Page Indexing report in Search Console and read the status, because the label names the cause.
- Discovered, not crawled, a discovery or crawl priority issue. Improve internal links to it.
- Crawled, not indexed, Google looked and declined. That is a page quality or duplication problem, and resubmitting will not fix it.
- Duplicate, Google chose different canonical, it was merged with another page.
- Excluded by noindex, you are blocking it yourself.
Being tested: whether you reach for a tool or a process. The strong answer is that the status tells you the fix, so you never guess.
Analytics, Tools and Measurement Questions
34. Which SEO tools do you use, and what for?
Name them with a job attached, because a list on its own says nothing. A crawler for technical audits, a keyword and backlink tool for research, Search Console for what Google actually reports, and a rank tracker for movement over time.
The strong version adds what you would do without them. Search Console and a crawl will diagnose most problems; the paid tools mainly save time and add competitor data.
Being tested: whether you depend on tools or use them. Candidates who cannot work without a subscription struggle at companies that will not buy one.
35. Which Search Console reports do you use, and what for?
Performance for queries, pages, positions and the branded versus non-branded split. Page Indexing for what Google has and has not kept. URL Inspection for a single page, because it shows Google’s rendered version rather than your browser’s.
Then Core Web Vitals, and Manual Actions, which you check first whenever something drops suddenly.
Being tested: whether you use it daily or only when something breaks. Naming Performance and Page Indexing as routine is the practical answer.
36. Search Console shows pages are indexed but impressions are very low. How do you investigate?
Indexed is not the same as ranking. Work through it in order.
- Check the average position. Pages sitting at 40 or 60 are indexed and simply never seen.
- Check whether the queries exist at all. A page can be perfectly optimised for something nobody searches.
- Compare the page against what ranks. If the intent is different, no amount of optimisation fixes it.
- Look for cannibalisation, where your own pages compete for the same term.
- Then consider authority, which is usually the answer when everything else looks right.
Being tested: whether you separate indexing from ranking. Candidates who suggest resubmitting the sitemap have not understood the question.
37. How do you find and fix crawl errors in Search Console?
The Page Indexing report groups URLs by reason, so the label names the fix. Server errors point at hosting, 404s at broken internal links or removed pages, redirect errors at chains and loops.
Fix the cause, then use Validate Fix so Google recrawls that set rather than waiting.
One judgement call worth stating: not every 404 is a problem. A page that is genuinely gone should return 404, and redirecting it somewhere loosely related creates a soft 404 instead.
38. Why do Search Console and Google Analytics numbers never match?
Because they measure different things at different points. Search Console counts what happened at Google: impressions and clicks. Analytics counts what happened on your site, after the page loaded and the tag fired.
So clicks will normally exceed sessions. People leave before the page loads, block scripts, decline consent, or get redirected on the way in.
Search Console also aggregates by query and day and omits very rare queries for privacy, so its totals are not a simple sum.
Being tested: whether a mismatch makes you suspect broken tracking or whether you know it is expected.
39. How do you get organic keyword data in GA4?
You do not, not from GA4 alone. Keyword-level data has been withheld since 2011, which is why organic terms show as not provided.
You link Search Console to GA4, which adds query data in a limited set of reports. For anything serious you work in Search Console directly, or export both and join them.
40. How does GA4 differ from Universal Analytics?
GA4 is event-based rather than session-based, so everything is an event and sessions are derived rather than primary. Conversions are configured rather than assumed, and data retention is limited by default.
The metric people trip over is bounce rate. GA4 launched with engagement rate instead, and its bounce rate is simply the inverse of that, so it does not mean what it meant in Universal Analytics.
41. Do you use Bing Webmaster Tools, and how is it different from Search Console?
Yes, and the honest reason is not Bing’s traffic share. It is free, it takes ten minutes, and it gives you things Search Console does not.
Its keyword research tool shows actual search volumes rather than the ranges Google gives away, and its reports are often more generous with data.
It also supports IndexNow, which lets you notify search engines the moment a URL changes rather than waiting to be crawled. Google’s own sitemap ping endpoint was retired in 2023, so this is one of the few push mechanisms left.
Being tested: whether you dismiss Bing on market share alone. The volume data and IndexNow are the reasons to have it regardless.
42. How do you set up and use GA4 for SEO?
Link it to Search Console so query data sits alongside behaviour, define conversions that reflect the business rather than every click, segment organic traffic properly, and use landing page reports rather than site-wide totals.
Know the platform differences that trip people up: GA4 is event-based rather than session-based, and its “engagement rate” is the inverse of the bounce rate people expect from Universal Analytics.
Being tested: whether you can actually operate the tool the company reports from, not just name it.
43. Do bounce rate and dwell time affect rankings?
This one is closer to a trap than it looks. Google has repeatedly said it does not use bounce rate or time on page as ranking signals, and both are unreliable measurements anyway.
A single-page visit where someone got their answer and left is a success, and it counts as a bounce. That is why the metric is a poor quality proxy.
What is fair to say is that the underlying experience matters. Pages that satisfy people tend to earn links, repeat visits and brand searches, and those do influence rankings.
Being tested: whether you repeat a popular claim or know Google’s actual position.
44. How do you measure SEO success?
Against the business goal, and with a hierarchy. Revenue or qualified leads from organic at the top, then conversions and non-branded organic sessions, then impressions, rankings and indexing as diagnostics.
Separate branded from non-branded, because branded search rises with any marketing activity and quietly flatters an SEO report.
Local SEO Interview Questions
45. How is local SEO different from regular SEO?
It optimises for a map pack and a physical or service area, not only for blue links. The ranking inputs differ too: relevance, distance and prominence, with the Google Business Profile doing much of the work.
46. What would you do to improve a local business’s map pack visibility?
Complete and verify the Google Business Profile properly, choose accurate primary and secondary categories, keep name, address and phone consistent across citation sites, earn and respond to reviews, and build location-relevant pages that are genuinely different from each other rather than one template with the town name swapped.
Being tested: whether you name the profile and categories first. Candidates who jump to backlinks are applying national SEO to a local problem.
47. What is NAP consistency and does it still matter?
NAP is your business name, address and phone number, appearing identically everywhere: your site, Google Business Profile, directories and old listings.
The honest answer is that it matters less than it did. Google is far better at recognising that two slightly different listings are the same business than it was five years ago.
Where it still bites is contradiction: an old address on a directory you forgot about, or a phone number from a previous owner. Those create genuine doubt, and doubt is what costs you.
48. How would you handle a business with multiple locations?
A separate Google Business Profile for each location, and a separate page on the site for each, with genuinely different content: staff, parking, local landmarks, the services that branch actually offers.
The trap is producing one template with the town name swapped. That is the doorway page pattern Google’s policies describe, and it is the most common way multi-location SEO goes wrong.
Being tested: whether you know where location pages cross into doorway pages.
49. How do you track local rankings accurately?
You cannot use a normal rank tracker, and saying so is the answer. Local results change with the searcher’s physical position, so one number for “plumber in Bristol” is meaningless.
You need location-specific tracking, ideally a grid that samples rankings from multiple points across the service area. A business can rank first outside its own door and nowhere three miles away.
50. What do you do about a negative or fake review?
Respond to genuine negative reviews professionally and publicly, because the reply is read mainly by future customers rather than the complainant.
Fake reviews that breach Google’s policies can be reported, though removal is slow and not guaranteed. Honest negative reviews cannot be removed, and trying is wasted effort.
Being tested: whether you promise to get bad reviews deleted. That answer tells an interviewer you have not done this work.
51. What changes for a service-area business with no storefront?
The address is hidden and service areas are defined instead. That removes proximity-to-address as a lever, so relevance and prominence carry more weight.
It also changes citations, since many directories expect a street address. You are working with fewer signals than a business with a shopfront, which is worth setting as an expectation up front.
The 2026 Questions: AI Search
This is the fastest way to show that your knowledge is current.
52. How have AI Overviews changed SEO?
They have separated rankings from clicks. A page can hold its position and still lose traffic, because the answer appears above it.
Bring numbers if you can. Pew Research, tracking the browsing of 900 US adults in March 2025, found people clicked a result far less often when Google showed an AI summary, and clicked a source inside the summary in about 1% of visits.
Being tested: whether you would spot a traffic drop that is not a ranking drop, and explain it to a client without panicking.
53. How do you get cited in AI answers?
Broadly the same things that earn links: clear structure, specific and citable facts, original data, and being a source others already reference.
The honest addition is that AI answers lean heavily on community platforms. Reddit threads and Quora answers surface as sources far more often than their link value would suggest.
54. What are GEO and AEO, and are they different from SEO?
Generative Engine Optimisation and Answer Engine Optimisation are newer labels for optimising to be cited in AI-generated answers rather than only to rank in blue links.
The honest answer is that the underlying work overlaps heavily with good SEO: clear structure, specific citable facts, entity clarity and being a source others reference. Treat them as an emphasis, not a separate discipline, and be sceptical of anyone selling them as one.
55. What is llms.txt and should we implement it?
A proposed file, in the spirit of robots.txt, that offers AI systems a curated guide to a site’s content.
The current honest position is that adoption is limited and no major AI provider has committed to using it as a ranking or citation signal. It costs little to add, and claiming it will improve AI visibility is overstating what is known.
Being tested: whether you can discuss something new without either dismissing it or overselling it.
56. How would you track visibility in AI answers?
Search Console now reports AI Overview performance within Search data, so start there. Beyond it, a category of AI visibility tools monitors whether your brand is mentioned and cited across assistants.
Set expectations honestly: this data is far less complete than rank tracking, results vary between sessions, and anyone promising precise AI ranking positions is selling something.
57. How should we optimise for voice search?
This question is often a check on whether you repeat old predictions. Around 2016 the claim that half of all searches would be voice by 2020 was repeated everywhere. It did not happen, and voice never became the separate discipline the industry expected.
What is true is narrower. Voice queries tend to be longer and more conversational, they often have local intent, and assistants usually read back a single answer, so being that one answer matters more than ranking third.
In practice the work is the same work: answer questions directly, structure content clearly, keep business details accurate for local queries. Voice has largely folded into AI assistants, which is where the same optimisation now pays off.
Being tested: whether you say “optimise for conversational long-tail keywords” and stop, or whether you can put the trend in proportion.
58. How do ChatGPT, Perplexity, Gemini and Copilot differ for SEO?
They source answers differently, which changes what you can influence.
- Perplexity runs live searches and cites sources visibly, so being a citable source on a crawlable page matters most.
- ChatGPT answers from both training data and its own live search, which is a separate system with a separate crawler.
- Gemini sits closest to Google’s own index, so ordinary Google visibility carries over.
- Copilot leans on Bing, which is a practical reason to keep Bing Webmaster Tools set up.
Being tested: whether you treat “AI search” as one thing. Naming which engine draws on which index is the answer that shows you have looked.
59. How would you get your content cited in ChatGPT specifically?
Start with the crawler, because most sites get this wrong. OpenAI runs two separate bots and blocking one has no effect on the other.
- GPTBot collects content that may be used for model training.
- OAI-SearchBot is what surfaces sites in ChatGPT’s search answers.
The trap: people block “AI bots” wholesale in robots.txt and remove themselves from ChatGPT search entirely without meaning to. The common position is to block GPTBot and allow OAI-SearchBot, so you stay visible in answers without feeding training.
After that it is the same work as being cited anywhere: clear structure, specific facts worth quoting, and a presence on the platforms these systems lean on, which includes Perplexity and the community sites they cite heavily.
60. Is SEO dying because of AI?
No, but the metrics are shifting. Impressions and brand visibility matter more, raw click counts matter less, and reporting that only shows sessions will increasingly tell a misleading story.
Being tested: whether you can handle an opinion question without either panicking or dismissing it. Both extremes read badly.
Trap Questions Where the Textbook Answer Is Now Wrong
These are the ones that separate candidates. Each has an answer that was correct years ago, is still repeated in older guides, and will cost you the room.
61. “How do you handle toxic backlinks?”
Outdated answer: run a toxic link audit and disavow them regularly.
Current answer: almost certainly do nothing. Google’s John Mueller has said the disavow tool “is not something that you need to do on a regular basis. It’s not a part of normal site maintenance. I would really only use that if you have a manual spam action.”
Google’s systems generally ignore junk links, because otherwise anyone could damage a competitor. The “toxic” scores that prompt most disavow files come from third-party tools, not Google.
62. “How do you improve Domain Authority?”
The correct move is to gently correct the premise. Domain Authority is a Moz metric. Google does not calculate it, use it, or see it.
It is reasonable for comparing two sites and a poor target to chase, since you can raise it without affecting rankings at all.
Being tested: sometimes deliberately. Some interviewers ask this to see whether you will repeat an agency talking point.
63. “Is duplicate content penalised?”
No. There is no duplicate content penalty, and there has not been one for over a decade. What happens is filtering: Google picks one version to show and the others do not appear.
The exception is content duplicated deliberately to manipulate rankings, which falls under spam policies rather than a duplicate content penalty.
64. “Do nofollow links help SEO?”
They do not pass ranking credit, but since 2020 Google treats nofollow as a hint rather than a directive, so it may use those links for crawling and ranking.
They also bring referral traffic and make a link profile look natural. A profile of 100% followed links looks engineered.
65. “Does AI-written content get penalised?”
Not for being AI-written. Google’s policy since March 2024 is scaled content abuse, which targets pages produced mainly to manipulate rankings, and it is explicitly method-agnostic.
AI, automation or a room of freelancers, the policy judges the purpose of the pages rather than the tooling.
Scenario and Situational Questions
These matter more than definitions, because they show your thinking. There is rarely one right answer, and the interviewer is watching your order of operations.
66. “Traffic dropped 40% last month. What do you do?”
Work outward from the cheapest explanations first.
- Check for tracking breakage before assuming a real drop, since a broken tag looks identical to a collapse.
- Check Search Console for manual actions and for indexing changes.
- Separate branded from non-branded, and compare against the same period last year for seasonality.
- Check whether a core update landed, and whether competitors moved too.
- Only then look at site changes, migrations and technical faults.
Being tested: whether you rule out the boring causes before reaching for the dramatic ones.
67. “A client wants to rank number one for a head term in three months.”
Say no, and say it usefully. Explain what is realistic given their authority, propose the long-tail terms you can genuinely win in that window, and set a measurable interim goal.
Being tested: whether you will overpromise to please someone. Interviewers ask this because overpromising is how agencies lose clients.
68. “You inherit a site with 5,000 pages and no documentation. Where do you start?”
Crawl it, then compare that crawl against what is actually indexed and what actually earns impressions. The gap between those three numbers is your roadmap.
69. “How would you prove SEO is working to a CFO?”
In their terms, not yours. Organic-attributed revenue or qualified leads, cost per acquisition against paid channels, and the compounding nature of the asset.
Rankings and impressions are diagnostics, not results. Saying so is the answer.
The Questions You Cannot Revise For
Everything above can be prepared. These cannot, and a good interviewer knows it, which is why they are often the questions that decide the outcome.
There is no model answer to memorise here. What follows is what each one is really asking, so you can think about it before you are sitting in the room rather than during.
70. “What is something most SEOs believe that you think is wrong?”
This separates people who read the field from people who repeat it. Have an actual position ready, with a reason behind it. Any defensible view beats a safe non-answer.
71. “Tell me about a time you advised against doing SEO.”
They want to hear that you know when the channel is wrong: no search demand, a launch that needs paid, a product nobody is looking for yet. If you have never advised against it, say why honestly rather than inventing an example.
72. “How do you decide what not to do?”
Every audit produces more recommendations than anyone has time for. Answer with how you weigh effort against impact, because working down a checklist in order is the answer they are hoping you do not give.
73. “A developer disagrees that your fix matters and says it will take three weeks. What happens next?”
Technical SEO lives or dies on this. Show that you can quantify the impact, negotiate, and accept a smaller version of the fix rather than escalating or dropping it.
74. “What did you believe about SEO that turned out to be wrong, and how did you find out?”
Have a specific one ready. This tests whether your knowledge updates, which matters more than what you currently know, and a vague answer wastes the best opportunity in the interview.
75. “Show me a keyword you chose not to target, and tell me why.”
Anyone can point at a keyword they went after. Being able to explain one you turned down, and why, is what shows judgement.
76. “What is in your reporting that the client never asks for?”
This is asking whether you report what is comfortable or what is true. Saying that you include what went badly, and why, is the answer that earns trust.
Questions by Experience Level
For freshers
Expect fundamentals, plus one question you cannot fake: what have you actually done?
Nothing beats a real site. A blog you ran yourself, with Search Console screenshots showing what changed and why, outperforms any certificate. If you have one, bring the data.
You will also be asked how you keep up. Naming specific sources you genuinely read is fine; claiming to read everything is not.
For experienced roles
Expect scenarios, prioritisation and disagreement. A common one is describing a time you were wrong, which is testing whether you notice and correct yourself.
Have one migration story, one recovery story and one thing that failed. The failure story, told with what you learned, tends to land better than a run of successes.
Questions to Ask the Interviewer, and the Red Flags
The 73 questions above are ones the interviewer asks you. These are the ones you put to the interviewer, which is why they sit outside that numbered list.
Do not treat them as a formality at the end. What you choose to ask is read as evidence of whether you have done the job before, and the interviewer’s answers tell you whether the role can actually succeed.
Worth asking:
- How is SEO success measured here? If nobody can answer, nobody will be able to defend your budget either.
- Who signs off on content, and how long does it take? This predicts more about your results than any tool.
- How much developer time does SEO get? Technical recommendations that nobody implements are just documents.
- What happened to the last person in this role? A new position and a replacement are very different situations to walk into.
- What has been tried already, and what did not work? This saves you repeating something the company has already decided it dislikes.
- What is the current link building approach? The answer tells you whether you will inherit a problem.
Red flags in the interviewer’s answers
- Guaranteed rankings promised to clients. You would be expected to deliver them.
- A fixed monthly link quota. That number is usually hit with directory and profile links that do nothing.
- No developer access, paired with an expectation of technical results.
- Success defined only as traffic, with no line to revenue or leads.
- An unexplained gap where the last SEO left after a few months.
- Content sign-off with no named owner. If nobody owns approval, your work sits in a queue indefinitely.
How to Prepare in a Week
- Pick one site you can talk about in detail, ideally your own, and know its numbers.
- Prepare three stories: something that worked, something that failed, and a disagreement you handled.
- Refresh the current answers to the trap questions above, since those are where dated knowledge shows.
- Audit the company’s own site before you go in and bring two specific observations. Not a full audit, two things you noticed.
- Prepare your own questions, from the list above.
That last one changes the tone of the conversation more than anything else on this page.
Frequently Asked Questions
What are the most common SEO interview questions?
The difference between on-page, off-page and technical SEO, how Google crawls and indexes, what canonical tags and robots.txt do, how you approach keyword research, and how you measure results. Beyond those, expect scenario questions such as diagnosing a traffic drop, since they show how you think rather than what you memorised.
How do I prepare for an SEO interview with no experience?
Run a real site, even a small one. A blog with Search Console data you can talk through beats any certificate, because you can explain what you changed, what happened and what you would do differently. Learn the fundamentals properly, and be honest about what you have not done yet.
What technical SEO questions should I expect?
Canonical tags, the difference between robots.txt and noindex, crawl budget, structured data, Core Web Vitals, redirects and hreflang. A frequent trap is whether you know that blocking a page in robots.txt stops Google seeing its noindex tag, so the page can stay in the index.
Should I still talk about Domain Authority in an interview?
Only with the caveat attached. Domain Authority is a Moz metric that Google does not calculate or use. It is fine as a rough way to compare two sites, but presenting it as a Google ranking factor is a common way candidates reveal dated knowledge.
What do interviewers ask about AI and SEO in 2026?
How AI Overviews affect traffic, how to get cited in AI answers, and whether AI-written content is penalised. On the last one, the current answer is that Google’s scaled content abuse policy judges the purpose of pages rather than how they were produced, so AI writing is not penalised for being AI.
What questions should I ask in an SEO interview?
Ask the interviewer how SEO success is measured, who signs off on content and how long that takes, how much developer time SEO gets, and what the current link building approach is. The answers tell you whether the role can succeed, and asking them signals that you have done the job before.
What SEO tools should I learn first?
The free ones, in this order: Google Search Console, then Google Analytics 4, then a crawler such as Screaming Frog, whose free tier covers 500 URLs. Those three will diagnose most problems on most sites.
Learn a paid keyword and backlink tool after that, not before. Employers can teach you their subscription in a week, but they cannot teach you how to read Search Console in a week.
Which SEO tools should I list on my CV?
Only the ones you can be questioned on for ten minutes. Listing a tool you opened twice is a common way candidates create a hole in their own interview.
It is stronger to name three tools you genuinely use, with what you use each one for, than to list eight and be vague about all of them.
Do I need to know Ahrefs or Semrush to get an SEO job?
Usually not. Most employers already pay for one and will give you a login, and the two are similar enough that knowing either transfers.
What is far harder to fake is knowing Search Console properly, since that is the data Google itself reports and every employer has it. Interviewers notice the difference between someone who has read about a tool and someone who has used it to solve a problem.
How long should I prepare for an SEO interview?
A week is enough if you already work in SEO. Spend it refreshing the questions where the accepted answer has changed, preparing three specific stories from your own work, and auditing the company’s site so you can bring two concrete observations.