How To Find Competitor Backlinks - 4 Discovery Phases

Competitor backlink discovery is the first phase of every link-building campaign we run. Done well, it produces a list of 300 candidate domains that real publishers in your niche have already chosen to link to. Done badly, it produces a spreadsheet of high-DR sites that look impressive in a deliverable and place none of the links the campaign actually needs.

This post documents our discovery process across 4 phases. The output is a qualified candidate list that hands off cleanly to our link prospect qualification workflow for scoring. The two posts together cover the full prospect funnel.

Why Competitor Backlink Discovery Comes First in Every Campaign

Competitor backlink discovery comes first because it answers the only useful question at the start of a link campaign: where are real publishers in your niche actually linking? Every other discovery method (broken-link mining, resource-page outreach, HARO) produces narrower candidate sets that depend on this baseline.

A site that ranks above you organically has already passed editorial review at the publishers you need links from. Those publishers will not link to a fitness blog about a legal SaaS topic. They will link to the legal SaaS that earned a relevant placement. Looking at where those competitors got linked is the cheapest, fastest map of the editorial territory worth chasing.

The mistake operators make is treating "competitor" as "the brand my client mentions during sales calls." Sales-mentioned competitors are often not organic-search competitors. The brands ranking above you for the queries you sell against are the real targets - and they are usually 1-3 names the client has never named in a meeting.

The 4-Phase Discovery Process at a Glance

The 4-phase process runs in order. Phase 1 identifies the right competitors (typically 5). Phase 2 extracts the backlink intersection. Phase 3 filters to the pursuable subset. Phase 4 hands off the candidate list to the prospect qualification workflow. Total operator time: 3-4 hours for a fresh campaign.

• Phase 1: Identify the right competitors. Tool: Ahrefs Organic Competitors report. Output: 5 organic competitors selected from the algorithmic list, validated against the client's actual SERP territory.

• Phase 2: Extract the backlink intersection. Tool: Ahrefs Link Intersect. Output: domains linking to 2 or more of the 5 competitors. Typical raw output: 300-600 candidates.

• Phase 3: Filter to the pursuable subset. Tool: Ahrefs Site Explorer + manual review. Output: roughly 300 candidates passing the 4-criteria filter (organic traffic, topical relevance, indexation, placement context).

• Phase 4: Hand off to qualification. Tool: our weighted scorecard. Output: the candidate list enters the qualification workflow which scores down to ~60 actionable prospects.

Phase 1: Identify The Right Competitors

Phase 1 identifies the 5 organic competitors whose backlink profiles we will intersect in Phase 2. The list comes from Ahrefs Organic Competitors filtered against the client's target SERPs. Sales-named competitors get included only if they also appear in the organic data.

We open Ahrefs Site Explorer for the client's domain, then Organic Search > Competing Domains. Ahrefs ranks competitors by shared keyword count and traffic overlap. We pull the top 20 and cross-reference each against the SERPs for the client's 5-10 highest-value commercial queries. A competitor that ranks top-10 for at least 3 of those queries qualifies.

We deliberately limit to 5 competitors because Ahrefs Link Intersect (the Phase 2 tool) handles 5 cleanly with the "linking to 2+" filter applied. More than 5 dilutes the intersection threshold. Less than 3 leaves the candidate list too small.

The output of Phase 1 is a list of 5 competitor domains plus a short note per domain on which queries they win. This becomes the input for Phase 2 and the reference document for the campaign overall. It also informs the topical-relevance scoring later in the link-prospecting qualification workflow.

Phase 2: Extract The Backlink Intersection

Phase 2 produces the raw candidate list using Ahrefs Link Intersect with all 5 competitor domains entered and the filter set to "linking to 2 or more." Typical output: 300-600 referring domains. The threshold matters: domains linking to only 1 competitor are coincidence; domains linking to 2+ are pattern.

We open Ahrefs Link Intersect, enter the 5 competitor domains in the "Show me who is linking to these targets" field, leave the "Excluding the targets" field with the client domain, and set the intersection filter to "Linking to 2+." Ahrefs returns the unique referring-domain list with each domain's DR, traffic, and the number of competitors it links to.

We export the full list to CSV. Sorting by "linking to" count descending reveals the publishers serving the niche most consistently - those linking to 4 or 5 of the 5 competitors are the highest-priority Phase 3 candidates. Domains linking to only 2 are still worth keeping but get less attention.

Common Phase 2 mistakes: forgetting to exclude the client domain (skews the data with self-references), using fewer than 5 competitors (intersection threshold too forgiving), and using the "Linking to all" filter (cuts the list to maybe 30 domains, losing useful pattern data).

Phase 2 takes about 45 minutes including the export and the initial CSV cleanup. The output is the raw candidate list that feeds Phase 3. For an alternative manual approach that does not require Ahrefs, see our manual backlink check sequence.

Phase 3: Filter To The Pursuable Subset

Phase 3 applies our 4-criteria filter to the raw 300-600 candidate list, dropping anything that fails on organic traffic, topical relevance, indexation status, or placement context. Output: roughly 300 pursuable candidates ready for the qualification workflow. Time: 90-120 minutes for the bulk filter. Before that slower manual pass, you can check Domain Rating across the whole list in one paste with our free tool, so the clearly-thin domains surface first and the 90 minutes goes to candidates worth reading.

The 4 criteria, applied in order from cheapest to most expensive:

• Organic traffic floor. Ahrefs traffic value above $500/month (or estimated organic traffic above 1,000/month for B2B niches). Zero-traffic domains are link farms regardless of DR. We drop these first because Ahrefs surfaces the metric in the same view we exported from in Phase 2.

• Topical relevance. Source covers your niche territory. The check: open the candidate's top 20 ranking pages and verify entity overlap with the client's target cluster. We score this 0-10; anything below 4 gets dropped.

• Indexation status. Linking page is indexed in Google. Use a site:domain.com/page query to confirm. Non-indexed pages do not pass signal.

• Placement context. The competitor link sits in editorial body content, not footer or sidebar. We sample 2-3 of each candidate's links to competitors to check placement pattern.

The 4-criteria filter is the same one we apply during full qualification. Phase 3 runs it at lower depth (5-10 seconds per candidate to drop obvious failures) so Phase 4 can run the full weighted scoring on a cleaner list.

Phase 3 drops about half the raw Phase 2 list - typical 300 candidates from 600. The kept candidates head into our weighted qualification scorecard where each gets scored 0-100 across 5 weighted criteria.

Phase 4: Hand Off To Outreach Qualification

Phase 4 hands the filtered candidate list to the prospect qualification workflow. We do not start outreach against the Phase 3 output - the list is still too coarse. Qualification scores each candidate 0-100 across 5 weighted criteria, surfaces the ~60 with composite scores above 65, and produces the outreach-ready batch.

The handoff is operational, not strategic. Phase 3 produces a list of "domains that pass our quality floor." Phase 4 (qualification) produces a list of "domains we will actively pursue this quarter." The distinction matters because outreach is the most expensive phase of the campaign, and burning outreach effort on lukewarm candidates is the dominant failure mode in agency link-building.

The qualification workflow runs the candidate list against topical relevance (40% weight), editorial quality (25%), backlink profile health (15%), traffic and engagement signals (10%), and outreach yield probability (10%). For the full workflow, see our link prospect qualification scorecard.

Tools We Use (and Which We Skip)

We run discovery in Ahrefs primarily, with one auxiliary tool for cross-validation. Most of the "competitor backlink tool" market sells features built on top of the same underlying data Ahrefs surfaces directly. We avoid the wrappers because they add cost without adding signal.

Those are the tools we reach for in discovery. If you are still choosing one, I compare the backlink checker tools on index size, gap analysis, and price in a separate guide.

• Ahrefs Site Explorer + Link Intersect. Primary tool. Handles Phases 1, 2, and the bulk of 3. The Link Intersect feature is the load-bearing tool of the entire process.

• Semrush Backlink Gap. Auxiliary tool for cross-validation. We open it only when Ahrefs surfaces an unusually small intersection list - sometimes Semrush picks up referring domains Ahrefs missed.

• Majestic. We use Majestic Trust Flow as a tiebreaker on borderline candidates in Phase 3. Not part of the core flow.

• Sitebulb. Used elsewhere in our audit work but not for competitor backlink discovery.

Top-10 ranking pages tend to carry more referring domains than positions 11-20. The competitor intersection method, the same approach Ahrefs documents in its competitor link analysis guide, exploits this directly: domains that link to multiple top-10 competitors are the highest-probability targets for your own outreach.

How Often We Re-Run The Discovery

We re-run the full 4-phase discovery once per quarter on every active client. Competitor profiles shift, new editorial publishers enter the niche, and old prospects either get exhausted (we placed there) or rotated out (their editorial standards shifted). Stale candidate lists drop conversion rates 30-50% over 6 months.

Between quarterly refreshes, we do not run discovery again. We work the qualified candidate list from the last refresh until it is exhausted or until conversion rates drop sharply. Re-running discovery more often produces marginally different lists at full operator cost - the time is better spent on outreach against the existing qualified pool.

Who Should Run This Process

This process is for agency outreach managers running 3 or more active client campaigns, in-house SEOs at brands with active link-building budgets, and freelance link operators handling multiple clients in parallel. Solo operators on a single campaign can use a lighter version focused on Phase 1 and Phase 2 only.

The structured process earns its weight at scale. A single campaign can be discovery-light if the operator already knows the niche's top publishers from prior work. Multiple campaigns across multiple verticals cannot - the only sustainable approach is the same documented workflow applied per client. The senior strategist running our monthly SEO retainer executes this 4-phase process during week 1 of each new client engagement.

What Goes Wrong with DIY Competitor Backlink Discovery

Three failure modes repeat: choosing the wrong 5 competitors (sales-named brands instead of organic competitors), running Link Intersect without the "linking to 2+" filter (too many false positives), and skipping the 4-criteria filter to move straight to outreach (the dominant failure that produces $200-per-link spreadsheets with zero placements).

• Wrong-competitor selection. Including a competitor the client mentions but who does not rank for the client's commercial queries pollutes the intersection list with off-target publishers. Validate every competitor against the client's real SERPs before Phase 2.

• Loose intersection filter. Running Link Intersect with "linking to 1+" instead of "2+" returns 2,000+ domains. Most are noise. The pattern signal of "multiple competitors got linked here" is what makes the discovery method work; relaxing the filter erases the signal.

• Skipping Phase 3. Operators under deadline pressure sometimes go straight from Phase 2 (raw 600 candidates) into outreach. The conversion math falls apart: a 5% outreach yield on 600 unfiltered candidates is 30 attempted placements with most failing on editorial quality grounds. The same 5% yield on 300 filtered candidates produces 15 placements, but the AVERAGE quality of those 15 is far higher.

• Treating the list as static. A candidate list discovered in Q1 and worked through Q2-Q3-Q4 progressively loses conversion as the warm prospects get exhausted. Quarterly re-runs are not optional.

Competitor backlink research is one input into the prospecting phase. For the end-to-end program this fits into, see Link Building: The Operations Guide.

Competitor Backlink Discovery FAQ

How long does the full 4-phase discovery take?

3-4 hours of operator time for a fresh campaign. Phase 1 is 30-45 minutes, Phase 2 is about 45 minutes, Phase 3 is 90-120 minutes (the longest phase because the manual filter takes time per candidate), and Phase 4 is the handoff to qualification (immediate).

Can we run this with fewer than 5 competitors?

Yes but the intersection threshold has to relax. With 3 competitors, change the filter to "linking to 2+" (which equals "linking to most") and accept a smaller candidate list. With 2 competitors, the method effectively becomes single-competitor analysis, which loses the pattern-detection benefit.

What if a client has no organic competitors yet?

Run discovery against the competitors of the closest commercially-adjacent vertical. A pre-launch legal SaaS targeting personal-injury attorneys can run discovery against the established personal-injury legal SaaS market and inherit the publisher list. Most editorial publishers cover the vertical, not the specific brand.

Does this process work for local businesses?

Partially. Phases 1-2 work the same. Phase 3 filtering relaxes the organic-traffic threshold because local-business publishers (community news sites, local directories) often have low Ahrefs-measured traffic but high local intent. Use local-keyword ranking as the substitute relevance signal.

Why exclude the client domain in Phase 2?

If you do not exclude the client domain, Ahrefs counts domains already linking to the client as part of the intersection. Those are not discovery targets - they are existing links. The whole point of the intersection method is finding NEW publishers, which requires removing publishers we are already on.

Is the candidate list confidential?

Treat it as confidential client work product. It is not a published asset. Sharing the candidate list with the client during scoping is fine; publishing it externally undermines the campaign by exposing target publishers to competitive outreach.