Link building and digital PR
How many backlinks do I need to rank?
Written by the Top SEO Experts Team
Reviewed by Siddharth GangalLast updated 19 min read
Quick answer
No single number of backlinks works for every page. Google publishes none, and a 2026 industry study put median referring domains for top-ranking sites anywhere from 76 to 3,027. The workable answer is relative: count the sites linking to pages that already rank for your keyword in your market, then close the relevant gap.
If an agency has told you “you need more links” without saying how many or from where, this post is for you. We haven’t run our own ranking study on the question. Every example below is labeled when it’s hypothetical.
Is there a fixed number of backlinks you need?
No. Google’s SEO Starter Guide, accessed 22 October 2026, sets no target and says other sites linking to you “happens naturally over time.” Links still matter, though. Google’s ranking systems guide says PageRank “continues to be part of our core ranking systems.”
Put those together and you get the confusion. Links are a real input, and nobody outside Google can say how much of any one ranking they explain.
Key takeaway Google sets no target number for links. A target is therefore always an estimate drawn from competitors, never a rule.
For local businesses, Google’s Business Profile help describes prominence. It says prominence is “also based on info like how many websites link to your business and how many reviews you have.”
Links sit beside reviews, distance and relevance. Our local SEO ranking factors guide shows the wider picture.
Definition A referring domain is a separate website that links to yours. Many links from one site still count as one referring domain.
What do published studies say about how many links rank?
They say top results usually carry far more links than a new local site has, and they disagree about how many. Different keywords, years, filters and tools produce different answers. Here are the three studies we could open, plus one local survey.
| Source | Sample | What it reported | Main limit |
|---|---|---|---|
| WebFX, 2026 Backlink Study by Industry (April 2026; fieldwork March 2026) | 1,462 top-ranking domains, 150 service-intent keywords, 15 industries | Median 907 referring domains overall (the same as its Jobs & Education median, so treat with caution); Home & Garden median 461; “roof repair near me” median 144 | Ten keywords per industry; authors sell SEO services; shows correlation only |
| Internet Marketing Ninjas, via Search Engine Land (June 2024) | 1,113 sites in the top 10 for 200 commercial keywords | Over 96% had more than 1,000 referring domains; none had 50 or fewer | Link counting method not explained; large brands in the sample |
| Ahrefs, backlink growth study (18 September 2018) | About 200,000 pages, 10,000 keywords | Top-3 pages generally gained more new referring domains | 2018 data; found no clear link between growth pace and ranking change |
| Whitespark, Local Search Ranking Factors 2026 (6 November 2025) | Practitioner survey | Link signals “appear to have dropped a bit” in local pack weighting | The share-of-weight values sit in a chart we couldn’t read as text |
Why do the studies disagree?
Read the rows side by side and the problem is plain. One study says nearly every top-10 site has over 1,000 referring domains.
Another puts the median at 907, and Home & Garden at 461. The 907 matches the Jobs & Education median in the same WebFX table, so treat it with caution. A plumber who borrowed the first number would plan years of work the market doesn’t demand.
Watch out The WebFX study doesn’t list HVAC, plumbing or roofing as categories. Home & Garden is our guess at the closest fit for home services, and the study itself only says to pick the nearest match.
How do the WebFX medians compare?
The chart below shows the three WebFX medians next to each other. The overall median is 907, Home & Garden is 461, and “roof repair near me” is 144.
Why does counting competitors’ links mislead?
Because a link count describes who already ranks, not what made them rank. Five distortions show up in the benchmarks we’ve looked at.
- Reverse causation. Pages that rank get found, and pages that get found collect links. The Ahrefs author says the relationship probably runs both ways, and Search Engine Land says the same about the Internet Marketing Ninjas data.
- Brand sites in the sample. Amazon appeared in the top 10 for 164 of the 200 keywords in that study. A local contractor isn’t competing with that link profile.
- Quality ignored. The WebFX authors acknowledge that referring domain counts say nothing about how authoritative the linking sites are.
- Page versus domain. A sitewide total can hide that the ranking URL has three links while the homepage has three hundred.
- Tool differences. In our experience each backlink index crawls differently, so numbers from two tools aren’t comparable. Compare competitors inside one tool only.
None of this makes a benchmark useless. It makes the raw total the wrong thing to chase.
Our view We use competitor counts to set a rough floor. Then we spend our effort on which links are missing, not how many.
How do you benchmark backlinks in your own city?
Count referring domains for the top three organic results and the top three map listings for one keyword in one city. Strip out the noise, then compare your page to the median.
It takes an afternoon. We found nobody publishing this for service businesses, so the method below is our own.
- Pick one keyword and one place. Use the money term, such as “water heater repair Tulsa.” Search from that city, signed out, and note the top three organic results and the top three map listings.
- Drop what you can’t compete with. Remove directories, review platforms, franchisors’ national pages and marketplaces. Keep local businesses and regional operators.
- Pull referring domains for each ranking URL. Use Ahrefs, Semrush, Majestic or Moz, but one tool only. Record the page-level count and the domain-level count separately.
- Remove obvious junk. Delete scraped directories, link farms and foreign-language spam from each list. Google’s own definition of link spam is links created “primarily for the purpose of manipulating search rankings.”
- Tag what’s left. Mark each source as local, trade-related, press, supplier or association, or general. Local and trade-related sources are your relevant set.
- Take medians, not averages. One well-linked competitor drags an average up. The median of your three to six comparables is a sturdier floor.
- Subtract your own numbers the same way, and write down the gap for the total, the filtered set and the relevant set.
- Pick one keyword and one placeNote the top three organic results and top three map listings.
- Drop what you can't compete withRemove directories, review platforms, national pages, marketplaces.
- Pull referring domainsUse one tool only. Record page-level and domain-level counts.
- Remove obvious junkDelete scraped directories, link farms and foreign-language spam.
- Tag what's leftMark each source as local, trade-related, press, supplier or general.
- Take medians, not averagesOne well-linked competitor drags an average up.
- Subtract your own numbersWrite down the gap for total, filtered and relevant sets.
Here’s the arithmetic for a hypothetical roofer. These numbers are invented to show the method, not measured.
| Site (hypothetical) | Referring domains, raw | After removing directories and junk | Local or trade-relevant |
|---|---|---|---|
| Competitor A | 118 | 71 | 22 |
| Competitor B | 65 | 40 | 14 |
| Competitor C | 42 | 30 | 9 |
| Median | 65 | 40 | 14 |
| Your site | 19 | 12 | 3 |
| Gap | 46 | 28 | 11 |
Read the three columns from left to right. The raw column is everything the tool lists. The middle column drops directories and junk, and the last keeps only local and trade-relevant sources.
Key takeaway The raw gap of 46 looks daunting. The relevant gap of 11 is a plan.
Spread over six months, it’s under two a month (11 divided by 6 is 1.8). It also tells you which kind of source to hunt for.
The chart below shows how the gap shrinks as the junk is removed. Its numbers are invented, like the table.
Compare that with the 48 new referring domains a month the WebFX sample averaged. The roofer’s real job is about one twenty-sixth of that, which is exactly why a national figure is a bad target.
Our competitor analysis guide covers more inputs for the same exercise.
When are links the bottleneck, and when are they not?
Links are the bottleneck only when the page is already indexable, relevant and complete, and the relevant gap is large. These are our rules of thumb. We haven’t tested the thresholds, so use them as a first cut.
- If the page can’t be indexed or doesn’t match the search intent, fix that first. Links to a broken page are wasted. Run the website SEO checker before you spend a dollar on outreach.
- If you’re chasing the map listings, compare reviews and profile completeness before links. Google lists both under prominence, and reviews are quicker to move.
- If the filtered gap is under about a quarter of the median, work on content and internal links before outreach. Our internal linking guide covers the cheapest links you control.
- If the relevant gap is large and the page matches competitors on content, plan link work over months. Nothing here is a quick fix.
- If every competitor in the top results is a national brand or aggregator, find the smallest local business in the top 10. Use it as your floor.
Run rule 3 on the hypothetical roofer. A quarter of the filtered median of 40 is 10, and the filtered gap is 28. That’s well over the line, so link work is justified, assuming rules 1 and 2 pass.
Build authority worth earning.
Not sure where to start? Get a plan for your own business.
How fast should you add backlinks?
At a pace you can sustain with real sources. Published velocity figures describe national samples.
What do the studies report?
The table sets the two studies side by side.
| Study | What it found |
|---|---|
| WebFX, top-ranking sites | About 48 new referring domains a month over a three-month look-back |
| WebFX, apparel | Near 15 a month |
| WebFX, Finance & Insurance | Near 101 a month |
| Ahrefs, 2018 | Most number-one pages gaining followed referring domains at roughly 5% to 14.5% a month |
Neither figure is a target for a plumber with 19 referring domains. Compare that with the hypothetical roofer above, whose relevant gap works out to under two links a month. That pace is a small fraction of the national samples.
What does Google say about speed?
Google hasn’t published a speed limit. Its spam policies define link spam by purpose, not by count or timing.
Our view A steady handful of editorial links a month looks like a business being talked about. A sudden jump of 80 links from unrelated sites looks like a purchase.
How do we vet link prospects?
We ask six questions about every source, and a bad answer to any one stops us. This is our own process, not an industry standard. Our full vetting post goes deeper.
- Real readers. Do people in the service area actually visit the site?
- Editorial control. Did a person choose what appears, or can anyone pay for a spot?
- Fit. Is it about the trade, the place, or the client’s customers?
- Nofollow test. Would the link still be worth earning if it were qualified with nofollow?
- Context. Does it sit in running text or a real profile, not a footer or a block of unrelated links?
- Risk. Does the site sell links openly, or look like part of a network?
- Real readersDo people in the service area actually visit the site?
- Editorial controlDid a person choose what appears, or can anyone pay?
- FitIs it about the trade, the place, or the customers?
- Nofollow testWould the link still be worth earning if it were nofollow?
- ContextDoes it sit in running text or a real profile?
- RiskDoes the site sell links openly or look like a network?
We also confirm that a live, recent page exists and that the site is still active. The sources that pass are usually ones a person would mention unprompted. For the kinds that survive this test for trades, see our post on contractor link building and the guide to local link building.
What do we refuse to do, and which policies back that up?
We refuse tactics built to manufacture links, because Google’s spam policies list several of them and because they fail the vetting above. If a vendor’s proposal includes any row in this table, that’s your cue to ask harder questions.
| We do not | Why | Policy or source |
|---|---|---|
| Sell bulk link packages | Volume placed for rankings | Google spam policies, link spam |
| Build or buy from private blog networks | Sites created to pass links | Our policy; Google’s link spam list does not name private blog networks |
| Buy links without qualification | Paid links need rel="sponsored" or rel="nofollow" |
Google spam policies; Google link qualification docs |
| Place paid advertorials or articles with links that pass ranking credit | Paid articles that pass ranking credit | Google spam policies, link spam (advertorials and native advertising) |
| Trade links at scale | Excessive link exchanges | Google spam policies |
| Use automated link creation | Programs that create links | Google spam policies |
| Promise a ranking or an authority score | Nobody outside Google controls either | Our no-guarantee position |
Google’s spam policies say violations are found “both through automated systems and, as needed, human review,” and that affected sites “may rank lower in results or not appear in results at all.” Google’s link qualification page says paid placements should carry the sponsored value, with nofollow still acceptable.
The spam update documentation adds a consequence that surprises people. When Google’s systems discount spammy links, “any ranking benefit the links may have previously generated for your site is lost,” and it “cannot be regained.”
Watch out You can’t buy a count and later undo it.
Definition A nofollow link carries a tag that qualifies it as paid or untrusted. Google says such links are generally not followed.
Our link building service follows this list. If you’re weighing a proposal, the guide to buying backlinks explains how paid links differ from sponsorships.
What should you ask a vendor who says you need more links?
Ask for specifics before you agree to anything. A claim that you need more links should come with numbers and sources.
- How many links, and from what kinds of sites?
- Which competitors and which keywords did you compare?
- Which tool did you use, and on what date?
- Are paid placements qualified with rel=sponsored or rel=nofollow?
- Does any row in the table above appear in your plan?
Quick check If a source quotes a precise link count as needed for a position, ask for the sample, the keywords and the date.
A clear answer is a good sign. A vague one is your cue to ask harder questions.
What mistakes do we see when people chase a link count?
Most mistakes come from treating the count as the goal. These are the ones we flag most often.
- Waiting for a number before fixing the page. If your page is thinner than the three above it, extra links won’t close that. More of these appear in our link building mistakes post.
- Buying a package to hit a figure. Packages deliver volume, and volume is what the spam policies describe.
- Heavy exact-match anchor text on a thin profile. It looks engineered. Our anchor text guide covers a safer mix.
- Treating a vendor’s authority score as the target. Our view is that it’s a convenience metric. Our domain authority guide explains why Google doesn’t use it.
- Benchmarking once and never again. Competitors keep earning links, so re-run the seven steps each quarter and compare against the last gap.
What does this mean for a local business?
Start with your own city, not a national study. Count the referring domains for the top three organic results and the top three map listings, then compare your page to the median.
Next, look at the relevant gap, not the raw one. In our hypothetical roofer example, a raw gap of 46 shrinks to a relevant gap of 11 once directories and junk are removed.
Plan that gap over months and add links at a pace you can sustain. Re-run the seven steps each quarter. If the page can’t be indexed or doesn’t match the search intent, fix that first.
What does the data not show?
Why can’t these studies prove cause?
It doesn’t show that adding a given number of links will move a given page. Every study above profiles sites that already rank, so none can prove cause.
The WebFX authors call their formula a starting point, not a guarantee. The Ahrefs author says his experiment shows correlation, not cause, and found no clear tie between link growth pace and rank change.
What gaps remain?
We also have no result of our own to cite. Our evidence pages contain no link-building case from a home services or other local business, and we won’t pretend otherwise. If a source quotes a precise link count “needed” for a position, ask for the sample, the keywords and the date.
Two more gaps. Tool counts differ, so a benchmark made in Ahrefs won’t match one made in Semrush. And we couldn’t read the Whitespark chart values as text, so we don’t quote a percentage for links in local weighting.
Watch out We haven’t verified the claim, repeated in several write-ups, that Google has said pages need “very few links.” Search Engine Land mentions it, and we haven’t traced it to a Google document. Treat it as unconfirmed.
Frequently asked questions
How many backlinks do I need to rank on Google?
There’s no published number. Google’s documentation gives no target, and studies disagree because they sampled different keywords, years and tools.
Count the referring domains of pages already ranking for your keyword in your market, then plan to close the relevant part of that gap. Treat the result as an estimate.
How many backlinks do I need to rank locally?
Local results mix organic pages and map listings. Google says prominence is partly based on how many websites link to your business and how many reviews you have.
Compare links and reviews against the top three businesses in your own city before choosing a link target. We don’t have a universal local figure.
Is one good backlink better than ten weak ones?
Often, but we can’t prove a ratio, because studies that count links ignore quality. Our view is that one editorial link from a relevant local site is worth more than several from sites that exist to sell links. The weak ones can also carry risk under Google’s spam policies.
How many backlinks should I build per month?
Match the pace to what you can earn. A WebFX study found top-ranking sites averaged about 48 new referring domains a month in early 2026, but that’s a national sample across 15 industries.
Google defines link spam by purpose, not speed. We’d plan a steady handful of vetted links rather than a spike.
Do nofollow links count toward the number I need?
Google says nofollow and sponsored links are acceptable ways to qualify paid or untrusted links, and that such links are generally not followed. Counting tools often include them, so your benchmark may run high. We still pursue some, because the audience, referrals and credibility can be worth having on their own.
Can I fix a low backlink count by buying links?
Buying links to influence rankings violates Google’s link spam policy unless the links are qualified with rel=sponsored or rel=nofollow. Google also says benefit from links its systems discount is lost and can’t be regained. We don’t sell bulk links and we advise against buying them to close a gap.
Sources
- Google Search Central: SEO Starter Guide, accessed 22 Oct 2026
- Google Search Central: A guide to Google Search ranking systems, accessed 22 Oct 2026
- Google Search Central: Spam policies for Google web search, accessed 22 Oct 2026
- Google Search Central: Qualify your outbound links, accessed 22 Oct 2026
- Google Search Central: Spam updates, accessed 22 Oct 2026
- Google Business Profile Help: Improve your local ranking on Google, accessed 22 Oct 2026
- WebFX: 2026 Backlink Study by Industry, accessed 22 Oct 2026
- Search Engine Land: Study on links in Google’s top 10 (Internet Marketing Ninjas), accessed 22 Oct 2026
- Ahrefs: Backlink growth study, 18 September 2018, accessed 22 Oct 2026
- Whitespark: Local Search Ranking Factors 2026, accessed 22 Oct 2026
Run your own page through the website SEO checker first, then Get Your Organic Growth Plan.
Build authority worth earning.
Not sure where to start? Get a plan for your own business.
