Compare backlink evidence

A larger backlink total can describe a different scope, date or counting rule. Make the inputs comparable and retain the rows behind each total before drawing a conclusion. A missing inventory row is a question to investigate.

Match the report scope

Save the original files with provider, report, target, scope, filters, collection dates and stated row limit. Exact page, folder, host and whole-domain reports answer different questions. Label deliberate mismatches.

Work on copies. Spreadsheets can change dates and URLs. Map source, target, observation time, anchor, attributes and status into a separate working table; retain original columns and mark missing values. A supplier row alone does not establish a current link.

Check the specific report's limit. Semrush's Backlink Audit limit applies to that report, not every Semrush export. Save any truncation or one-link-per-domain setting beside the file.

Choose what one count means

Keep four units distinct: link instances, source pages, hostnames and registrable domains. One page can contain two links; forty pages can repeat one footer link. Neither count establishes forty independent publishers.

Preserve the provider's domain field and derive your grouping separately. Use a documented rule and Public Suffix List version rather than taking the last two hostname labels. One organization may own several domains; one platform may host unrelated authors. Domain grouping cannot establish ownership.

Keep raw URLs, anchors, attributes and dates beneath the aggregate. Add normalized comparison values without replacing the originals: case, query parameters, redirects and trailing slashes can change resource identity. Date any redirect you use to group URLs; today's route does not establish historical equivalence.

Separate the dates

Discovery is when a system added a candidate; collection is when it retrieved evidence; export and import record when files moved; verification is your own check. An HTTP response date describes message timing, not when the backlink appeared.

A file downloaded today may contain old observations. Keep row-level dates, timezones and supplied precision. Do not turn a date-only value into invented midnight. Preserve the last complete observation when a later request fails, while showing that the latest attempt is unresolved.

Choose a refresh interval for the decision and budget. A weekly schedule alone cannot guarantee fresh evidence.

Keep exclusions consistent

State each exclusion's purpose, match scope and reason before applying it. Preview affected rows and whole groups across every compared inventory, including your own. Retain excluded evidence, rule version, owner and date.

Suppose one domain links to a competitor from a resource page and to you from a profile. Removing only your profile row manufactures a missing-domain gap. Either exclude the whole group consistently or compare exact source pages. Access failures and missing targets belong in an unresolved queue; they are not policy exclusions.

Read gaps and intersections

Backlink Gap and Link Intersect can help frame source comparisons. Declare your own grouping and threshold before reviewing results. Two-of-three competitors and all-three are different sets. Count each competitor once per group; repeated links cannot increase that threshold.

This table is a synthetic teaching example, not measured supplier or customer data:

B is shared research material. E cannot become a gap merely because its import is incomplete. A group present for two known competitors qualifies for “at least two” even when the third is unknown. The unknown value does not add a membership or support an all-three claim.

Freeze input identifiers, dates, competitor choices, thresholds, grouping and exclusions. Products can require approved comparison inputs; an arbitrary import may not qualify. Freshness limits can reduce a report to overlap-only analysis; see the methodology.

Explain changes between inventories

Keep the comparison conditions stable. Adding competitors, changing filters or broadening an exact URL to a domain starts a different series.

In a second synthetic example, 80 source groups become 86 through ten additions and four missing groups. The net gain is six; fourteen memberships changed. Newly discovered pages may be old. Missing rows may still contain live links. Report inventory additions and omissions separately from dated page inspections.

Choose the next inspection

Open the exact source, its destination and surrounding passage. A shared domain may contain an irrelevant archive, a repeated footer or a useful resource list. Inspect what earned the reference before proposing your own page.

Save the included, excluded and unresolved counts with reviewed mismatches. Then use opportunity review for reader fit or verification and repair for a placement you expected to exist. Product import steps belong in the import and verification docs.

Sources (5)

1. Learn more about the Public Suffix List · Public Suffix List · Not stated; accessed September 12, 2026
2. RFC 9110: HTTP Semantics · IETF · June 2022
3. Backlink Gap · Semrush · Not stated; accessed September 12, 2026
4. Link Intersect · Ahrefs · Not stated; accessed September 12, 2026
5. Why didn't Backlink Audit fetch all my backlinks and referring domains? · Semrush · Not stated; accessed September 12, 2026