A content comparison becomes useful when it helps someone make a specific editorial decision. A store owner might need to find which product descriptions explain materials clearly. A developer might compare service pages before a redesign. An editor might check how a topic is represented across several articles. In each case, the important question is what the collected documents actually say and how confidently that evidence supports a change.
The CrawlNFix comparison tool accepts bulk keyword lists, groups of selected page URLs and bounded website crawls. Saved HTML or text can also be analyzed locally after the tool loads. You can work without a public account, a paid plan or an external provider API key. Online fetching still uses a server, while local analysis uses your device.
Give every collection a clear purpose
Begin by identifying comparable material. Group a shop's related product pages together and place another shop's equivalent product pages in its own competitor group. For a service business, compare pages describing similar services and audiences. For an editorial project, use documents addressing a common question in the same language. Different collection sizes should remain visible because a large collection naturally offers more opportunities to contain a phrase.
CrawlNFix supports your collection and up to three competitor groups. Each group can contain several page URLs, one per line, so two pages from the same competitor stay together rather than appearing as two unrelated businesses. The report retains per-page evidence alongside the aggregate view. This makes it possible to investigate whether an apparent collection-wide advantage comes from one document or is spread across many pages.
Import a keyword list with a reason behind it
Paste keywords one per line or import a TXT/CSV file. The first CSV column holds the keyword, and a conventional keyword header can be omitted from the list. A run accepts up to 100 unique phrases, with a maximum phrase length and a bounded page-keyword processing budget. These technical budgets protect processing time and memory; they do not create a paid upgrade or a daily subscription quota.
Start with language your audience would reasonably use for the page's subject. A phrase can be useful even when its exact words are absent from a well-written page. Conversely, repeating a phrase many times does not demonstrate usefulness. The tool measures normalized exact matches, including Unicode text, within sentence and text-block boundaries. It does not join the last word of one card to the first word of another, interpret intent or predict semantic equivalence.
Select known pages or discover a website section
Selected-page mode requests the URLs you supplied. It is appropriate for an intentional sample such as several product descriptions, service pages or articles. You control the sample directly, and a repeat scan can use the same URLs to investigate a manual edit. Up to 25 submitted URLs fit each group, subject to the run's source and time budgets.
Website mode starts from the group's seed URLs and discovers permitted same-origin pages through HTML links and bounded sitemap reads. Choose a crawl page budget and inspect the resulting coverage. Robots rules, excluded paths, response errors, source size and elapsed time may stop or reduce discovery. A budget-limited crawl is an observed sample and should not be presented as a complete inventory.
Use the site crawler when the task requires a broader link graph, inbound-link evidence or a focused investigation of discovery. The robots tester and sitemap validator help explain why a relevant section may be missing from a comparison.
Read totals through the per-page evidence
The selected keyword controls placement metrics for the current view. Body occurrences are summed across analyzed pages. A positive title, heading or description result means the phrase was observed on at least one page in that collection. It does not establish that all pages are optimized or that every document should contain the same phrase.
Main-content words and H1 counts also aggregate across the collection. Navigation, footer and detected short-link menus are separated from primary text; form controls and option lists are excluded. Repeated/chrome occurrences appear separately, so twenty footer links do not turn two body mentions into twenty-two. Detection uses source structure and heuristics, so inspect unusual templates manually. Internal-link totals may count a shared destination on several pages. The coverage list shows each page and its selected-keyword occurrences, so you can check where the result came from. Failed competitor requests remain unavailable evidence instead of turning into imaginary zero-word documents.
Use filters to narrow an investigation
Change the keyword selector to inspect another phrase without fetching the same source again. Select a competitor to focus the table and phrase candidates on a particular group. Measured-area filters separate placement evidence from content size, headings, links and schema. Phrase search and minimum-occurrence controls narrow the observed phrase list.
A separate page-type filter narrows the coverage list to products, categories or other pages. The interface explains that collection totals still cover the full analyzed sample. This separation prevents a filtered list of three products from silently rewriting a collection total that also includes a category and several informational pages. Downloads retain the full measured run.
Interpret commerce fields cautiously
Public e-commerce pages can contain Product, ProductGroup, Offer or AggregateOffer structured data. CrawlNFix reports observed product schema, offer schema, SKU, price, currency and availability values when they can be read from that source. URL patterns and structured types help classify product and category evidence; classification is not a comprehensive platform certification.
A missing field does not prove that a product has no price or that inventory is unavailable. JavaScript, regional pricing, variants and platform templates may affect the fetched response. The comparison excludes cart, checkout, payment, account and login paths. It does not submit an order or validate a merchant feed. Follow product evidence back to the actual public page before making a change.
Separate a phrase candidate from a recommendation
Observed phrase gaps describe repeated wording found in a competitor collection and absent from yours. Brand-like terms and overlapping phrases are reduced to make the list easier to review. Shared candidates appear once per competitor collection and are independent of the selected keyword. The list remains a statistical content observation. It does not prove search demand, a business opportunity or a missing product feature.
For example, a phrase about a backlink database should not be added to a product that does not provide backlink data. An unrelated pricing promotion should not become a paragraph in your service explanation. Investigate whether a genuinely relevant topic would help your own audience, then write an original explanation grounded in the features you actually offer.
Keep an online and a local workflow
Online comparison reads public HTML with a source budget per page. Scripts are not executed, so the response may omit content that appears after browser rendering. A truncated document is explicitly marked as partial. If you have permission to review a rendered document, saving its HTML can provide a useful local comparison source.
Saved-content mode keeps the input on your device and treats HTML as data. Uploaded scripts are not executed. It works after the tool and its assets have loaded. Fresh offline startup and packaged app availability are separate capabilities and should be verified independently. Live URL fetching and website discovery require a network connection.
Compare the available approaches
| Approach | Evidence | Useful for |
|---|---|---|
| Selected page groups | Known product, service or editorial samples | Only submitted URLs within the run budget |
| Bounded website comparison | Keywords across discoverable public pages | Robots, source, page and time limits remain visible |
| Saved HTML/text | Local comparison of supplied documents | Source versions may differ from the current public site |
| Manual review | Meaning, relevance and factual accuracy | Requires a person to read and judge the evidence |
These approaches can support the same investigation at different stages. Choose the source that answers your question and record what remains unknown.
Using CrawlNFix for this task
Use SEO Audit for diagnostic evidence alongside the content comparison. The meta tag checker and heading checker can help investigate a title, description or heading decision. Make a relevant change in the source page or template, deploy it, then use Rescan comparison with the same inputs.
Clear data resets input lists, uploaded selections, errors and the current result before a new project. JSON and CSV preserve the full measured run. The compact PDF contains coverage, keyword totals, your page findings and one shared phrase list per competitor. Enable its detailed appendix for the full page-by-keyword matrix, including title, H1, description, URL and repeated/chrome evidence. Clickable URLs and page numbers help reviewers follow each finding. For the operational checklist, read the bulk comparison guide. Keep evidence and editorial judgment connected rather than treating an automated list as permission to publish changes.
A missing exact phrase is a question to investigate
A guide can explain canonical decisions using relevant language without containing the exact phrase “canonical URL checker.” Zero exact matches means those normalized words were not found together in the analyzed body. It does not mean the subject is uncovered, the page is broken or a keyword must be inserted. First inspect source coverage, then read the document for meaning. If the phrase naturally names the task, a clearer heading may help. If it describes a feature you do not offer, leave it out.
For a product-and-category comparison, follow the focused e-commerce workflow. Keep product pages and category pages in separate runs when their purposes differ.
Further reading
Google documents crawling, rendering and JavaScript SEO as separate processing concerns. Its structured-data guidelines explain why correct markup does not guarantee a rich result. Use those platform documents alongside the measured source evidence.