Rewrite or rebuild: making the call with SERP data

One of your posts drops from position 4 to 14 over three months. Most teams reflexively open the doc, swap a few paragraphs, update the date, and shove it back into the CMS. In 60% of the audits I ran across Ahrefs and Search Console, that cosmetic update never moved the needle because the content was not the problem - search intent had shifted. The right question is not 'how do I rewrite this?' but 'is this URL still the right answer to that query?'. Skip that step and you burn hours of editorial work to claw back two positions that do not convert anyway.

The first filter is looking at the current SERP with fresh eyes. Run the head query in an incognito window, map the top 10 results, and label each one: guide, comparison, calculator, video, listicle. If 7 of the 10 changed format since you published, that is intent drift - and an update will not fix it. This exercise links directly to Search intent: 4 types and how to map them on the SERP, because reclassifying the query is non-negotiable before you optimize for a game that already ended. In parallel, run Content decay: spotting the posts quietly losing traffic to separate ordinary decay from structural change.

After the SERP, go to internal data. In Search Console, compare average CTR over the last 28 days against the expected curve for that position - use CTR benchmark by position: updated 2026 data as the reference. If you sit at position 6 with a 1.2% CTR while the benchmark is 6%, the problem is not ranking, it is the snippet. Tuning Title tags that convert: 7 patterns tested on real SERPs and Does meta description still matter? What CTR data shows can recover traffic without touching the body. Update wins here because the cost is low and the improvement signal shows up in GSC within two weeks.

Now the trigger to rebuild from scratch. If content depth sits below the top-5 median (measured by entities covered in Surfer or MarketMuse), if the original angle aged out, or if you wrote for a persona that has since shifted - rewriting paragraph by paragraph leaves scars. New URL, new brief, new structure. The old post becomes a 301 or turns into a support page inside a larger cluster, following Topical authority: how to build clusters that rank. Rebuilding costs 3x more but avoids the worst outcome: a Frankenstein that ranks neither as the old version did nor as the new one should.

There is an underused middle path: modular surgery. Instead of treating the post as one atom, break it into blocks (intro, definition, step-by-step, FAQ, comparison) and replace only the blocks that lost relevance. It works beautifully on long pages where 70% of the content is still valid. Pair it with a review of Headings H1-H6: the structure Google actually reads and Smart interlinking: the internal authority map pointing into the new blocks, and you signal real freshness - not just a swapped timestamp. Tools like Clearscope help spot which entities vanished from your text but appear in today's top 3.

Cadence matters as much as the decision. Updating a pillar every 21 days burns editorial budget and confuses Googlebot, which crawled it last week - check Crawl budget: when to worry and how to measure it and Content refresh: the right cadence by page type to calibrate. My rule: commercial evergreen, quarterly reviews; news or regulatory posts, monthly; informational pillars, twice a year with a deep audit. Log the date and hypothesis of every update - without that record you cannot run A/B testing in SEO: methodology that resists noise later to learn what actually moved the metric.

Practical takeaway: before any rewrite, fill three columns in a sheet - current SERP intent, depth gap vs. top 3, and CTR vs. position benchmark. If all three say 'update', do modular surgery. If two or more say 'new', kill the old URL with a 301 and start clean. Rewriting on autopilot is the most expensive sin in content SEO in 2026.