SEO Manager Interview Questions

A technical SEO audit has to be systematic rather than opportunistic: it is too easy to chase the most visible issue rather than the one with the highest impact. I start with a crawl using Screaming Frog or Sitebulb to map the entire site structure, then layer in data from Google Search Console and Google Analytics to understand which issues are actually affecting pages that generate traffic. My audit framework covers six areas: crawlability (can search engines access everything they should?), indexability (are the right pages being indexed?), page speed and Core Web Vitals, internal linking structure, structured data implementation, and mobile usability. For each area I produce a prioritised list of issues scored by impact and implementation effort. The highest priority fixes are almost always crawl budget issues on large sites, Core Web Vitals failures on high-traffic pages, and canonical tag problems creating duplicate content. I also interview the engineering team before the audit to understand the technical stack, because different platforms have different failure modes and some fixes require developer time to schedule.

Sustainable content strategy starts with understanding the user's intent at each stage of the funnel, not just keyword volume. I map content needs to three levels: informational content for users who are learning about a topic, comparative content for users evaluating options, and transactional content for users ready to act. Within each level I conduct keyword research using a combination of Google Search Console data, keyword research tools like Ahrefs or Semrush, and "People Also Ask" analysis to understand the actual questions users have. I prioritise topics using a scoring model that weighs monthly search volume, keyword difficulty, and commercial relevance to the business. I also look at the current ranking position for each topic: a page ranking at position 6 to 15 with an existing article often needs on-page optimisation rather than new content, while a topic with no existing content requires a net-new piece. Content calendars are planned at least 90 days out, and I track each piece of content through its lifecycle from brief to ranking.

Link building quality has changed fundamentally over the past decade. Tactics that worked in 2015 are now liabilities. My approach focuses on two categories: earning links through content that naturally attracts attention, and building relationships that lead to editorial placements. For earned links I invest in creating original research, data visualisations, and tools that journalists and content creators in our space have an incentive to reference. I track mentions of our brand without links using tools like Ahrefs alerts and convert them into links by reaching out with a polite request. For relationship-based links I identify the sites that our target audience reads and that have editorial standards, and I build genuine relationships with their writers and editors before pitching placement. A valuable link has four characteristics: it comes from a site with genuine traffic and authority, it is contextually relevant to our content, the anchor text is natural, and the linking site has editorial standards. I never buy links or participate in link schemes, both because of the algorithmic risk and because link schemes typically produce low-quality placements.

International SEO has several layers of complexity beyond just translating content. The first decision is URL structure: subdomain, subdirectory, or country-code top-level domain. Subdirectories are the most common and manageable approach for most businesses because they consolidate domain authority and are easier to maintain. The second critical element is hreflang implementation: every page must correctly reference every language and region equivalent, and errors in the hreflang implementation are extremely common and result in the wrong language version appearing in search results. I audit hreflang using Screaming Frog and validate with Google Search Console's international targeting report. The third consideration is content localisation, not just translation: a French-language version of a page optimised for French search behaviour will significantly outperform a direct translation of English content, because user intent, terminology, and search behaviour differ meaningfully by market. I use local keyword research tools and native-speaker review for each market.

I led a site architecture restructure for a B2B SaaS company that had accumulated 12 years of content without a coherent structure. The site had significant keyword cannibalisation: 40% of our target keywords had multiple competing pages that were splitting ranking signals. I ran a content audit identifying which pages to consolidate, redirect, and delete, built a new internal linking architecture to concentrate authority on priority pages, and coordinated the developer work for a URL restructure with 301 redirects. The project took four months. Within six months of completion, organic traffic had grown 34% and our ranking for core commercial terms had improved significantly. The main learning was about communication with the engineering team: the first restructure proposal I presented was deprioritised because it was too large to schedule in one sprint. I learned to break site architecture work into phases that can be shipped incrementally.

During the 2023 HCU (Helpful Content Update) rollout, a client's organic traffic dropped approximately 28% over two weeks. My first step was to identify exactly which pages had lost visibility using Search Console and Ahrefs data: the losses were concentrated on broad informational articles that had been written primarily to rank rather than to provide genuine utility. The pattern matched what Google was targeting: thin content with limited original insight that mimicked articles already ranking. My response was a content quality programme that prioritised the top 30 impacted articles. For each one I conducted a gap analysis against the top-ranking competitors, identified the specific original value we could add, and rewrote rather than edited. Over the following six months, 18 of the 30 articles recovered and organic traffic overall returned to within 8% of the pre-update baseline. The lesson was that AI-generated content at scale without a quality review process is a structural vulnerability.

Early in a new role I inherited a content team producing eight articles per week. When I analysed the content performance data, I found that 70% of articles published in the past 12 months had received fewer than 50 organic sessions in their first six months. The team was producing a high volume of content on topics with very low search demand or high competition, while under-investing in the pages already ranking on pages 2 and 3 that could be pushed to page 1 with optimisation. I proposed cutting new content production by 50% and redirecting that capacity to optimising existing near-ranking content. The proposal was met with scepticism because "writing less" is counterintuitive in content marketing. I presented the data on the expected ROI comparison: optimising a page already ranking at position 12 typically generates results in six to eight weeks, while a new article takes four to twelve months. The team agreed and within three months the optimised pages had generated 40% more organic traffic.

Core Web Vitals are a direct ranking signal and also a genuine user experience proxy, so I treat them as a joint engineering and SEO responsibility. The three metrics are LCP (Largest Contentful Paint), INP (Interaction to Next Paint), and CLS (Cumulative Layout Shift). For LCP, the most common causes of poor scores are large unoptimised images, render-blocking JavaScript and CSS, slow server response times, and no preloading of the LCP element. My approach is to use PageSpeed Insights and the Chrome UX Report field data to identify which pages have real user LCP issues, then work with engineering to implement image format optimisation (WebP or AVIF), lazy loading, and CDN caching. For CLS, the most common cause is elements that shift layout when ads, fonts, or images load. INP (which replaced FID in 2024) measures interactivity responsiveness: heavy JavaScript execution on the main thread is the primary cause, and I use the Chrome DevTools Performance profiler to identify long tasks.

Search Console and GA4 answer different questions and combining them gives a much richer picture than either alone. Search Console tells me about impressions, clicks, average position, and click-through rate for each query and page: it is my primary tool for identifying ranking opportunities, diagnosing position drops, and understanding which queries are triggering pages I may not have optimised. GA4 tells me what happens after the click: engagement rate, pages per session, goal completions, and revenue. I combine the two by exporting Search Console data by landing page and joining it with GA4 behavioural data to identify pages with high impressions but low CTR, pages with high traffic but low engagement, and queries where I have strong click-through but the landing page is not converting. I also use GA4 to set up custom events for SEO-specific actions like newsletter signup from organic traffic, which lets me measure organic channel value beyond sessions.

AI-generated content at scale has changed the SEO landscape significantly and the risks are real. Google's Helpful Content guidance explicitly targets content created primarily for search engines rather than for users. My approach is to treat AI as a production accelerant for specific content tasks, not as an autonomous content factory. Tasks where AI adds genuine value without risk include generating outlines, expanding bullet points into prose drafts, translating content, and producing structured data markup. Tasks where AI introduces risk include writing original thought leadership, producing content that requires experiential authority, and generating statistics or citations that need to be verified. I apply a human editorial layer to every piece of AI-assisted content, which means fact-checking, adding original examples and perspectives, and ensuring the content has a distinctive point of view. I also monitor the ranking performance of AI-assisted content separately from hand-written content so I can detect any differential impact from future algorithm updates.