Who Gets the Lifeboat? Scoring Every Page Before You Purge Content
Photo: Unknown authorUnknown author or not provided, Public domain, via Wikimedia Commons
Content pruning has become one of the few SEO recommendations a CFO will approve without argument, largely because it sounds like housekeeping. Delete the thin pages, consolidate the near-duplicates, prune the blog to its winners. The advice is common, tidy, and — applied without a scoring framework — one of the more reliable ways an e-commerce site loses revenue it never knew it had.
Consider a composite case drawn from patterns we see repeatedly: a mid-market retailer purged 1,800 URLs ahead of a replatform. Traffic recovered within six weeks. Revenue did not, because among the deleted pages were forty-one product variants that received almost no organic entrances but appeared in multi-session journeys at a rate three times the site average. Nobody had measured assisted conversions. Nobody had checked which category pages depended on those variants for internal links. The pages were thin. They were also load-bearing.
Why "Thin" Describes a Symptom, Not a Page
Word count and crawl frequency describe a page's appearance, not its function. A specification table with ninety words may be the only page on the internet answering a precise compatibility question. A 2,400-word buying guide may be read by nobody and linked by nothing.
Before a purge, every candidate needs a seat assignment. In aviation, the exit-row briefing exists because the passenger sitting there has a job that other passengers do not. Your URL inventory works the same way: some pages look ordinary and are responsible for an exit.
The Four Scores
Score each candidate from 1 to 5 on each dimension. Do not average them into a single number until the very end, and expect a handful of pages to score high on exactly one measure and low on the rest. Those are the ones that get cut by mistake.
Assisted Conversions
Look beyond last-click. Pull every URL that appeared at any point in a converting session over the trailing twelve months and weight earlier touches appropriately. A page that initiates 4,000 comparison journeys and closes forty is not thin; it is early-funnel infrastructure. This is the single most neglected score in pruning decisions, and the one most likely to reverse a verdict.
Internal-Link Dependency
Map which pages receive internal links and which pages your crawl depth relies upon. Removing a hub can orphan an entire cluster, and orphaned pages do not announce themselves — they simply stop appearing in reports until a competitor outranks them.
Seasonal Demand
A page that produces nothing in April may be your best performer in the second week of December. Compare three years of data where available, and flag anything with a demand curve peaked outside the current quarter. Purges scheduled for summer are the classic trap.
Replacement Cost
Estimate what it would take to recreate equivalent coverage: research, writing, photography, structured data, and the authority accumulated in the URL itself. Replacement cost is usually the number that ends the argument, because rebuilding a page is never as cheap as the original team remembers.
Reading the Board
Once scoring is complete, sort candidates into four tiers rather than one delete list:
- Keep — high across assisted conversions or replacement cost. Leave alone, even if the page looks weak in isolation.
- Fortify — strong function, poor presentation. Rewrite, add structured data, improve internal links. This is where most of the budget should go.
- Consolidate — genuine overlap with a stronger page. Merge the content, redirect deliberately, and verify the destination satisfies the original intent.
- Evacuate — low on every measure with no seasonal spike and no link dependency. These are the pages the standard advice was actually about, and they are usually a much smaller group than expected.
In most audits, evacuation candidates account for 15% to 25% of the flagged inventory — not the 60% that a thin-content filter suggests.
Redirects Are Not a Rescue Plan
A 301 preserves a fraction of link equity and almost none of the user intent that made the removed page useful. Redirecting forty-one product variants to a parent category tells a shopper that the specific answer they wanted no longer exists. If a page is worth redirecting, it is usually worth improving instead.
Also resist consolidating pages whose only shared trait is a keyword stem. Two pages can target adjacent queries and serve genuinely different shopping stages; merging them produces one page that serves neither well.
The Ninety-Day Watch
After any purge, track the removed set for at least one quarter: organic entrances to their redirect targets, assisted conversions across the funnel, and internal link equity reaching the destination. Set the threshold for reversal before the purge, not after, because the case for restoring a page is always harder to make once the decision has been announced.
The framework will not make pruning painless. It will make it defensible, which is what a content team needs when the CFO asks why the page count barely moved.