An industrial-filter website is preparing more content. Its editor finds three articles containing “filtration precision.” One colleague proposes combining everything; another argues that different titles justify keeping every page. Neither has yet examined the task each page serves.
This is an explicit teaching simulation, not a customer performance case. The pages, review findings and proposed treatment are constructed to demonstrate a method. No real rankings or enquiry improvements are claimed. The topic follows recent discussions of keyword cannibalisation; the scenario and analysis are independently designed.
Put the three pages on the table
| Page | Illustrative title | Actual task in the scenario |
|---|---|---|
| A | How to choose industrial filter precision | Collect the operating inputs needed for selection |
| B | A guide to filtration precision selection | Repeat a similar process using the same inputs |
| C | How to check filter material compatibility | Organise compatibility evidence for engineering review |
In this setup, A and B overlap in reader, evidence and next action. C uses related terminology but supports a different decision. A keyword-only comparison could wrongly send C into the same consolidation.
Do not diagnose from two visible URLs alone
In a real project, first hold market, device and time range consistent, then examine which pages appear for a query group. Several URLs receiving impressions do not independently demonstrate harmful competition. Intent, geography and normal changes can explain variation.
Use Search Console’s query-to-page analysis workflow as a source of clues, then inspect the actual content. This exercise does not have a company’s private performance data, so it does not manufacture an impression table or ranking graph.
Write one outcome sentence per page
After A, the reader should be able to assemble selection inputs. In the scenario, B enables essentially the same action. After C, the reader should understand what compatibility information to gather and what requires technical approval.
Now label paragraphs as unique evidence, necessary explanation, repeated wording or unrelated material. A newer publication date does not automatically make B the better destination. Compare completeness, stable access, relevant links and maintainability before choosing the retained page.
Choose the destination before moving content
For this exercise, assume A already contains a clear input table while B’s only unique contribution explains how to follow up on incomplete customer information. Move that useful explanation into A and improve its unclear steps, rather than keeping two nearly equivalent guides.
Keep C separate because its decision differs. A can link to C where material compatibility becomes relevant, and C can refer back to A when discussing selection inputs. Their boundaries make them complementary rather than redundant.
The decision is conditional. If a real review finds B addresses a separate operating situation with different evidence, revise the scope instead of applying the simulated merge mechanically.
Handle the old address only after the content is ready
Once A genuinely serves B’s previous task, plan a permanent redirect from B to A and update internal links to the final address. The sitemap should reflect retained canonical pages.
Do not delete B first and hope the relationship will be inferred. Do not canonicalise C to A merely to concentrate signals: different tasks justify distinct pages. The bilingual archive repair case offers a useful separation between data, relationships and actual entry points.
What result can this exercise establish?
It produces an editorial plan: A owns selection inputs and absorbs B’s unique material; B’s old entry point receives equivalent treatment; C retains its compatibility task. That is a concrete structural outcome, not evidence that rankings rose.
After a real release, verify the old URL destination, retained content, internal links and indexing signals. Then observe whether relevant queries reach the appropriate page and whether readers can continue toward a product or enquiry. If the result disappoints, reconsider intent rather than attributing every problem to waiting for an algorithm.
| Layer | Immediate checks | Continued observation |
|---|---|---|
| Technical | Redirect, canonical and internal links | Search processing of the new structure |
| Content | Unique evidence retained and task clear | Remaining unanswered reader questions |
| Business | Product and enquiry paths work | Qualified enquiries from relevant visits |
Apply it to fifteen topics per category
Before commissioning each article, name the reader, decision, evidence and next action. If all four closely match an existing page, improvement may be better than another URL. If they differ materially, a separate article may be justified.
The aim is not the smallest possible page count. Each page needs a reason to exist. Google’s helpful-content assessment similarly asks whether material contributes substantial value. A continuing SEO / GEO workflow can assign evidence and review owners.
Questions after the exercise
Do similar titles always require a merge?
No. Examine tasks, evidence, readers and performance. Clarifying a title may be enough without changing its URL.
How soon should consolidation improve rankings?
No reliable gain or timetable can be promised. Verify the structure first and observe real data. This exercise supplies neither invented uplift nor a recovery deadline.
