Consolidating a Bloated Blog After a Subdomain Migration
Horizon HR, a B2B HR platform, moved its blog from blog.horizonhr.com to horizonhr.com/learn/ six months ago. The team redirected every old URL 1:1, kept all 3,800 posts live, and left 1,100 tag pages and 240 author archive pages indexable. Since the move, organic sessions to /learn/ are down 28% year over year, while GSC shows 42,000 URLs indexed under /learn/ and crawl requests to tag pages up 3x. Seventy percent of tag pages contain fewer than three posts; the top 40 posts drive 80% of blog traffic. Internal links in roughly 900 posts still point to the old blog subdomain, which now 301s. The VP of Marketing wants to 'just noindex all tags' and move on, but the SEO lead worries about losing category discovery and link equity. The company has one developer for two weeks and a content team of three.
1. Which URLs should be kept indexable, consolidated, or noindexed-and in what order should the team execute? Justify using the evidence in the setup.
Read the worked answer
Export every /learn/ URL from GSC with clicks, impressions, position, and last crawl. Segment into top 40 posts, long-tail posts with any clicks, tag pages with 3+ posts and impressions, thin tags, and author archives. Decision rule: keep posts and tags that earn clicks or impressions; merge tags by topic into 8-12 hubs; 301 thin tags to the nearest hub; noindex author and date archives, which have no unique query demand. Do not blanket-noindex all 1,100 tags: 70% are thin, but the other 30% may be the only internal path to older posts. Sequence by impact and effort: first, update the 900 body links from the old subdomain to final /learn/ URLs-the content team can do this in the CMS, and it removes 301 hops that waste crawl paths. Second, map tags to hubs and 301 thin tags. Third, noindex author archives. Fourth, add hub-to-post and post-to-hub links. Leave the top 40 posts unchanged until week 6 so you can isolate effects. If a tag page has more than 50 clicks in 90 days, keep it as a hub or merge it into one. Measure crawl requests to /learn/ and indexed count weekly; fix redirect chains if tag 301s create 404s.
2. The VP says 'just noindex all tags.' How do you respond, and what 90-day measurement plan would show whether the change helped or hurt?
Read the worked answer
Push back on the word 'all.' A blanket noindex is easy, but it treats a data problem as a switch. Pull GSC clicks by tag page first; if tags deliver less than 5% of /learn/ clicks, noindexing most is defensible. However, noindexing does not remove the URLs from internal navigation, so Googlebot still crawls them; that does not solve crawl waste. 301 thin tags into hubs both removes the page and passes link equity. For tags with clicks, keep or merge. The 240 author pages can be noindexed immediately-they are unlikely to satisfy search demand and add little unique value.
Measurement: baseline 90 days pre-change on (a) GSC indexed URLs under /learn/, (b) crawl requests by directory, (c) organic clicks to /learn/ and to the top 40 posts, (d) rankings for the top 40, and (e) sessions and demo requests from organic. Check at 14, 30, 60, and 90 days. Guardrail: if top-40 clicks fall more than 10% for two consecutive weeks, pause further consolidation. Because this is sitewide, you cannot A/B test; use matched cohorts-tags kept versus tags merged-and annotate the change date. Report indexed count and crawl distribution, not just sessions.
3. Design the internal linking model for the consolidated blog. What specific links would you add or remove first?
Read the worked answer
Use a hub-and-spoke model: one hub per cluster (for example, HR compliance, hiring, onboarding), each hub links to 8-15 posts with descriptive anchors; each post links up to its hub and sideways to 2-3 related posts. Remove the 900 old-subdomain links first because they are 301 hops; replace them with final URLs. Add breadcrumbs with BreadcrumbList schema. Prioritize links from the top 40 posts to commercial pages-those posts carry 80% of traffic and likely most of the link equity. Add a 'related resources' module with editorial selection, not an algorithmic tag cloud. Avoid sitewide footer links to every hub; they dilute context. Map anchor text to the target's primary query, but vary it naturally.
Measure link clicks in GA4 and crawl depth in server logs; target all posts within three clicks of /learn/. After 60 days, check whether hub pages rank for head terms and whether posts gain impressions for cluster queries. If a hub gets no impressions after 60 days, the problem is likely the hub's content or its internal links, not the number of tags.