Library·SOP·SEO
Compare a relevant competitor’s link profile, inspect representative referring pages, and choose realistic link-worthy work for your own site.
Last reviewed: · Editorial review
Clarity, structure and references reviewed in this version. This is not a test in a live account.
Compare a relevant competitor’s link profile, inspect representative referring pages, and choose realistic link-worthy work for your own site.
This SOP covers:
This does not cover:
A useful comparison starts with a peer serving a similar audience or need. Link counts alone do not make a site relevant to your business.
Make your own copy of the Semrush Competitor Backlink Pattern Review. The copy includes a Source URL count helper tab; leave its A1 Source URL header in place. In Start Here, enter your domain, the competitor domain, your audience and market, and one sentence about why the peer is useful to study. Record the domain scope and report date. Use Export Register for one row per downloaded file and separate totals for Backlinks and Referring Domains; use Sample Register for link-row and distinct-source-page counts.
Open Semrush Backlink Analytics or Backlinks and compare your domain with the competitor using the same scope. In Start Here, record the report date, each site’s backlink count and referring-domain count, and any visible trend period. A backlink is one reported link; a referring domain is one unique website that links. Keep the two counts separate.
Profile totals and charts describe the report. The source-page rows show where links appear and what they point to.
In the overview and detailed reports, compare the visible link growth over the same period, link types, link attributes, common anchors, target pages, and referring-domain categories. Record what you can see in Start Here. Keep the metric, date range, and your interpretation in separate notes.
Export the Backlinks and Referring Domains tables separately, and record the result count and coverage for each.
Open each table for the competitor and use the same domain scope and date range as your baseline. Clear the table filters, then enter each unfiltered result total in its own Backlinks or Referring Domains summary row in Export Register.
Use each table’s Export control to save CSV or Excel. Pro currently lists 10,000 rows per report; check the limit and access shown for your paid account because plan limits can change. Add one Export Register row per file with report name, exact filter names and values, filename, exported data-row count without the header, and imported data-row count after import. Keep the original files.
Export Register starts with 100 file rows, 11–110, shared by both reports. For another file, right-click row 111 and choose Insert 1 row above. Copy A110:G110 and use Edit > Paste special > Format only in A111:G111; copy A110 and use Paste special > Data validation only in A111. Extend C5, D5, C6, and D6 so their range endpoints reach your last file row, keeping the existing A5/A6 criteria and D/E sum columns. If row 111 is last, C5 is =SUMIF($A$11:$A$111,A5,$D$11:$D$111), D5 uses E, C6 uses A6 and D, and D6 uses A6 and E. Replace 111 with your actual last row and repeat for more files; the starter range is not a file limit.
If an export hits a row limit, use only filters shown in that report. Batches cover the full report only when they do not overlap and together include every row. Compare their combined rows with the unfiltered total. If you cannot establish that coverage, stop splitting and mark Capped. For an ordinary export whose counts disagree without a shown limit, use the Counts don't match check below.
After importing, choose a status for each report in Export Register. Use Complete only when displayed, exported and imported counts agree; Capped for a shown row limit or incomplete split; Unavailable when export is blocked. If an ordinary export's counts disagree, check the same domain scope, date, statuses, attributes and filters once. Reload the report once. If they still differ, select Counts don't match and record both displayed totals, file/import counts and dates in the notes. The cause remains unknown.
Import every downloaded Backlinks and Referring Domains file into its matching worksheet tab, then reconcile each report separately.
For the first Backlinks file, select Backlink Rows and import the file with the option that replaces only that sheet. Keep the original columns and record the imported data-row count, without the header, in that file’s Export Register row.
For the first Referring Domains file, select Referring Domains and import it with the option that replaces only that sheet. Keep the original columns and record the imported data-row count, without the header, in that file’s Export Register row.
For each remaining Backlinks file, append it to Backlink Rows if the import dialog offers that choice. Otherwise copy only its data rows below the last row in Backlink Rows. Remove any repeated header row and record that file’s imported data-row count in its Export Register row.
For each remaining Referring Domains file, append it to Referring Domains if offered. Otherwise copy only the data rows below the last row in that tab. Remove any repeated header row and record that file’s imported data-row count in its Export Register row.
For each report, compare its displayed total with exported and imported rows in Export Register. Keep the original displayed total and add the refreshed total to the notes, labeling approximate counts as approximate. If counts still disagree after the check above, use Counts don't match and note that full coverage is unconfirmed. Ask the report owner whether another report is needed to resolve it. Continue analysis of the downloaded rows, with that limit stated. Never remove raw backlink rows just to make totals match.
Use the Anchors report as a lead, then inspect matching Backlink Rows and pages. In Pattern Review, copy Anchor text as Semrush exports it and choose the Anchor class dropdown value: Brand/name, URL, Descriptive, Generic, Empty/image, or Mixed/other. Keep Link type and Link attribute separate: Link type is Text, Image, Form, or Frame; Link attribute is Follow, Nofollow, Sponsored, or UGC. In Observed source context, write what the inspected page shows around the link, such as a forum member’s reply. UGC as a link attribute is markup, not page context.
A profile can suggest where to look. Read the pages behind the links before you decide what the pattern means.
Plan the sample from the imported Backlink Rows by grouping link rows that share a feature that could change what the link means.
The sheet labels each group “Pattern / stratum”; a group means link rows that share a feature that could change what the link means. In Start Here → Sample Register, add one row for each recurring group you want to inspect. Name the feature, such as source context, target-page purpose, anchor class, link type or attribute, or new/lost status.
Population means all matching rows or pages in a group before you choose a sample. Filter Backlink Rows to one Sample Register group and count its individual link rows. Enter that number in Backlink rows in population. Do not use a domain or source-page count. For Capped or Counts don't match, count imported rows only and note the unknown remainder in unchecked scope.
Keep Backlink Rows filtered to one Sample Register group. In Source URL count, clear values from A2 through the last filled row from the previous group, leaving A1 intact. In Backlink Rows, find the column headed Source URL. Copy only its matching visible URL cells, then paste them beginning in A2. Do not copy other columns. Leave Backlink Rows and Referring Domains unchanged; never deduplicate backlink rows by referring domain.
On Source URL count, select A1:A[last pasted row]. Choose Data > Data cleanup > Remove duplicates. In the dialog, indicate that the range has a header row, select only the Source URL column, and click Remove duplicates. Sheets also merges URLs that differ only by capital letters. Keep the original rows so you can check those variants separately; they may lead to different pages.
On Source URL count, select the remaining filled URL cells from A2 through the last remaining URL. In the bottom-right calculation area, click its current label (for example, Sum) and choose Count. In Start Here → Sample Register, enter that number in Unique source pages in population for the same group; this is the count before selecting pages. In the row’s notes, state that Sheets merges URL letter-case variants, so this count may be lower than the number of separate pages.
In the same Sample Register row, explain why this group matters and what remains unchecked. Update Source pages inspected and Source pages left unchecked after inspection. For Capped or Counts don't match, state that population counts cover imported rows only and that other rows may be missing.
Sample Register starts with 20 rows, 25–44. When you need another group, right-click row 45 and choose Insert 1 row above. Copy A44:G44 and use Edit > Paste special > Format only in A45:G45; copy E44 and use Paste special > Formula only in E45. For each further group, insert above the next blank row and copy the last completed row’s format and E formula into it. The starting rows are not a sample limit.
Inspect representative pages for the groups recorded in Sample Register and record each link you actually check in Pattern Review.
Open at least one distinct Source URL page for each recurring Sample Register group. The Source URL is the page on another website that contains the link; the Target URL is the page the link points to. Read the text around the link, then open the Target URL. Inspect another page when a different target or context could change the interpretation. Do not describe a page you did not open as verified.
In Pattern Review, add one row for each inspected link. Record its Source URL, Target URL, exported Anchor text, Anchor class, Semrush Link type and Link attribute, visible context, and matching Sample Register group. If several checked links share one source page, record each link but count that source page once in Source pages inspected for that group.
Pattern Review starts with 200 rows, 5–204. For another sample, right-click row 205 and choose Insert 1 row above. Copy A204:N204 and use Edit > Paste special > Format only in A205:N205. Copy E204, J204, and K204 one at a time and use Paste special > Data validation only in the matching cells of row 205. For later rows, insert below the last completed row and copy its formatting plus E/J/K validations into the new row. The starter range is not a sample limit.
Check unusual link patterns and confirm the worksheet shows how much of each source-page population you inspected.
Review available outliers that could change your interpretation, such as unusual anchor groups, repeated sitewide links, sources with many links, new or lost patterns, or pages linked to a visible trend change. If you find a new recurring pattern, add a Sample Register row and inspect a representative page.
For each Sample Register group, count distinct Source URLs you checked and enter the count in Source pages inspected; a page counts once per group even if you inspected several links there. The worksheet calculates Source pages left unchecked from population minus inspected pages. Record why the selected pages matter and what remains unchecked.
Finish the available-row review when each recurring pattern and important outlier has a checked representative or a recorded access limit, and one full review of remaining imported rows adds no new pattern. For Complete, this covers the confirmed export. For Capped or Counts don't match, state that conclusions cover only downloaded rows and checked pages. Keep the full-report coverage question open and identify patterns whose pages could not be checked; do not describe unavailable rows.
For each checked link, choose the exact Pattern label dropdown value for the page you saw: Editorial/resource (a useful editorial reference), Image or data citation, Directory/listing (a business or organization profile), Product/review, Guest/author page, UGC/forum/comment (a user-written forum post or comment), Partnership/sponsorship, Sitewide/footer (a repeated sitewide placement), or Unknown when the page cannot be classified. Then choose Evidence confidence: Directly observed, Partial context, or Unverified hypothesis. A forum comment can have any Link attribute; its visible page context is why it may fit UGC/forum/comment.
For each useful pattern, ask whether your business can serve the same audience with a better page, original information, a real tool, or an eligible and accurate listing. In Opportunity Plan, choose Work next when the fit and required expertise are clear; Work later when the idea fits but another task or resource must come first; Needs facts when the business owner must confirm eligibility, expertise, or cost; or Do not pursue when the audience or offer does not fit. Explain the evidence, the missing capability, and the next internal action.
If your internal plan is to create or substantially improve a reader-focused page or resource, record its audience, purpose, unique information, and subject expert. Follow the linked SOP to prepare a writer-ready brief.
