TL;DR: There is no number. Links are one input among many into a ranking system nobody outside Google can observe, so a competitor's link count tells you about their position rather than what yours requires. The useful version of the question is a work rate: how many genuinely earned mentions a month can you sustain, and from which sources.
Key Takeaways
Link counts do not determine rankings, so a gap figure is not a target.
The competitor audit is still worth running, for the sources it reveals rather than the arithmetic.
The gap formula is pseudo-precise. Every input is an estimate and the output looks like a measurement.
Relevance and editorial independence are what distinguish a link that helps.
Velocity concerns are mostly folklore, except where they describe genuinely manufactured acquisition.
Set a work rate, not a count. Ten to twenty individual approaches a month is a sustainable programme.
Anchor targets assume your CMS generates slug IDs from heading text. Confirm on staging and adjust the hrefs if your platform slugifies differently.
Why there is no number
The question assumes links are a threshold: reach the count and the ranking follows. Rankings do not work that way.
Related reading: How Many Backlinks Do I Need? Data-Driven Guide
What the question assumes
Assumption | Reality |
|---|---|
Links are the deciding input | One signal among many |
A count is comparable across sites | Two sites with identical counts can rank very differently |
Competitors rank because of their links | They rank for reasons that include links |
The gap is closable by arithmetic | The gap includes age, brand, content and intent match |
The correlation point
Analyses of large result sets have repeatedly found links associated with ranking, which is real and frequently over-read. Association across millions of results tells you something about the population; it does not tell you what your page needs.
A widely cited study of millions of results found authority metrics correlating more strongly with rankings than raw link counts. That specific figure needs verifying before publication, and the conclusion it supports is directional rather than prescriptive.
Why correlation misleads here
Sites that earn links tend to be sites people reference, which tend to be sites with better content, stronger brands and longer histories. The links travel with the causes rather than replacing them.
Building links without the underlying reasons replicates the symptom rather than the condition.
What you can say honestly
More relevant, independently earned mentions tend to help. How many you need to rank for a specific term is not knowable, and anyone quoting a figure is estimating from a model they cannot validate.
The reframe
Stop asking how many and start asking which. The sources that matter for your category are findable, finite, and a much better basis for planning than a number.
What the audit tells you
Running it is worthwhile. Just not for the arithmetic.
What it genuinely reveals
Finding | Why it matters |
|---|---|
Which specific sites link to competitors | Your prospect list, with evidence |
Which publications cover your category | Frequently not the ones you assumed |
What those pages are about | Roundups, reviews, resource lists, news |
Where you are absent from a page listing competitors | The clearest gap available |
Which directories and platforms recur | Cheap wins, usually self-service |
Every row is a name. That is what makes the audit useful: it converts a vague ambition into a finite list of places.
The page-level finding
The most valuable single discovery is a page that lists several of your competitors and not you. That is not a link to build, it is an omission to point out, and publishers generally correct omissions when asked politely.
Those pages are usually a small minority of the profile and worth more than the rest combined.
Referring domains, not links
The brief is right about this. Count distinct domains rather than total links, since many links from one site add little.
That is standard practice and it does not make the resulting number a target.
What to record
A list of domains, what the linking page is, and whether you could plausibly appear there. Three columns, and the third is the one that turns research into a plan.
How long it takes
An afternoon for one competitor set, using any standard backlink tool. The output is a prospect list you can work for months.
What to ignore
The totals. Write them down if you like, and do not set them as objectives.
The gap formula problem
The formula in circulation looks rigorous. Take the top three competitors, filter referring domains by an authority threshold, average them, subtract yours, add a buffer.
It produces a precise number from inputs that cannot support one.
Where each step breaks
Step | Problem |
|---|---|
Top three competitors | Results differ by location, personalisation and over time |
Filter by authority above thirty | A third-party estimate, and thirty is arbitrary |
Average the remainder | Averaging three sites' positions tells you about them |
Subtract yours | Assumes parity on everything else |
Add ten percent | A made-up adjustment to a made-up figure |
The authority threshold
The sharpest issue. Domain authority scores are third-party estimates of ranking strength, produced by tools with their own models. They are not Google metrics and no threshold in them has any standing.
Filtering by a number from one vendor's model and treating the result as a requirement imports that model's assumptions wholesale.
What the output actually is
A description of how many domains link to three competitors, adjusted by a guess. That is a fact about them, not a requirement for you.
Why it persists
Because it converts an uncomfortable answer, nobody knows, into a number you can put in a plan. That is a real organisational need and the number does not meet it honestly.
The defensible alternative
Count the specific pages where competitors appear and you do not. That is also a number, it is checkable, and every unit of it is an action rather than an abstraction.
If leadership wants a figure
Give them that one: there are eleven pages in our category listing competitors and not us, and we are approaching them at a rate of four a month. Concrete, honest and trackable.
What makes a link help
Four things, none of them a score.
1. Editorial independence
Someone chose to mention you. Google's link spam policy addresses links that are bought, exchanged or placed by arrangement, and the distinguishing question is whether an editor made a decision.
This is the first filter and most of the rest follows from it.
2. Relevance
A mention on a page genuinely about your category, read by people in it. Relevance is about the page and the audience rather than the domain's overall metrics.
A small specialist publication in your field is worth more than a large general site that covers your topic once.
3. Context
What the sentence around the link says. A link with a describing sentence is doing work that a bare link is not, which matters for human readers and, as the outreach article covers, for anything composing an answer about your category.
4. Real readership
People actually read the page. This is the simplest test of whether a site is a publication or a link farm, and it is visible without any tool: look at whether the content is written for someone.
What does not determine it
Factor | Why not |
|---|---|
A third-party authority score | An estimate of ranking strength |
Dofollow versus nofollow | Google treats the attribute as a hint |
Anchor text optimisation | Manipulating it is the pattern being detected |
Total link count on the domain | Says nothing about this page |
The practical test
Would you want this mention if search engines did not exist? If the answer is yes, because the readers matter, it is a good link. If the only reason is the link itself, it is the kind of acquisition the policies describe.
Link velocity, examined
The idea that acquiring links too quickly triggers penalties is widely repeated and mostly misstated.
What is actually true
Google's systems identify patterns characteristic of manipulation. A site acquiring hundreds of links from unrelated low-quality sources in a week displays such a pattern.
What is not established
That a rate exists above which legitimate acquisition is penalised. A product launch, a funding announcement or a piece of research going round the industry can produce a large number of genuine links very quickly, and that is normal.
If speed alone triggered penalties, every successful launch would be penalised.
What the concern actually describes
The pattern | Why it is flagged |
|---|---|
Many links, unrelated sites, short window | Characteristic of purchased or automated acquisition |
Identical anchor text across sources | Characteristic of coordination |
Links from sites with no readership | Characteristic of a network |
All appearing with no corresponding event | Nothing happened to cause them |
It is the combination that identifies manipulation, not the rate.
Why the folklore matters
Because it produces the wrong instruction. Teams are told to slow down, when the actual requirement is to stop doing the thing that looks manufactured.
Slowly acquiring manufactured links is still manufactured link acquisition.
The honest advice
Earn mentions for real reasons and the rate takes care of itself. If you published research that got picked up widely in a month, that is a good month rather than a risk.
The one case to be careful about
Inheriting a profile built by a previous agency with manufactured links. That is a cleanup question rather than a velocity one, and the disavow discussion belongs there.
Where automation belongs
A real distinction, and the one this whole category gets wrong.
Reasonable to automate
Task | Why it is fine |
|---|---|
Finding candidate sites | Research at scale |
Finding contact details | Clerical |
Tracking who you contacted and when | A CRM function |
Monitoring acquired links for breakage | Maintenance |
Reporting | Administrative |
Not reasonable to automate
Task | Why not |
|---|---|
Choosing which sites to approach | The judgement that makes it targeted |
Writing the pitch | The part that works or fails |
Sending at volume | Where it becomes the pattern policies describe |
The line, stated plainly
Google's link spam policy names automated link building among the practices it treats as spam, and the policy addresses the activity rather than the vocabulary around it.
A tool that writes personalised emails and sends them at a volume no person could reach is automated outreach, whatever the personalisation layer is called.
Why volume fails on its own terms too
Publishers filter aggressively and the pattern is recognisable. Response rates on generic outreach are low enough that the activity consumes more time than targeted work while producing worse placements.
Twenty well-chosen approaches outperform three hundred generic ones on total placements, not just on quality.
The version that is defensible
Software finds prospects and tracks the pipeline. A person chooses the targets and writes each message. Ten to twenty a month, each genuinely specific to its recipient.
That is a real workflow with real tooling, and it is what the useful half of this category's products actually support.
The question to ask of any such tool
Does it make a person's judgement faster, or does it replace it? The first is leverage. The second is the thing being described in the policy.
The target to set instead
A work rate and a source list, which are both honest and both trackable.
The targets worth setting
Target | Example |
|---|---|
Approaches made | Fifteen individually written approaches a month |
Pages where competitors appear and you do not | Eleven identified, four approached this quarter |
Corrections requested | Every inaccurate mention found |
Something worth linking to | One piece of original research a quarter |
Placements accepted | Reported, not targeted, because it is not yours to control |
Why a work rate rather than an outcome
Because the outcome depends on publishers' decisions. You control the quality and the number of approaches; you do not control acceptance.
Targeting accepted placements creates pressure to lower the bar, which is how teams end up buying links.
The source list as the real asset
A maintained list of every publication, roundup, directory and community that matters in your category, with your status on each. That document is worth more than any count and it compounds.
The outreach article covers building it from what actually gets cited rather than from authority scores.
The thing that makes approaches land
Having something worth publishing. Original data is the strongest version, and it converts the request into an offer.
That is why a research cadence belongs in a link-building plan, which is not where most plans put it.
How to report it
Approaches made, corrections secured, inclusions achieved, and the source list with its coverage. No gap figure, no authority average, no projection.
The honest expectation
A sustained programme at this rate produces steady accumulation over quarters. It does not produce a step change, and anything promising one is describing a different activity.
Six mistakes
1. Treating a competitor's link count as your requirement. It describes their position, which they reached for reasons that include but are not limited to links.
2. Using the gap formula. Every input is an estimate and the output looks like a measurement. Count the pages where competitors appear and you do not instead.
3. Filtering prospects by authority score. It is one vendor's estimate of ranking strength. Relevance and editorial independence are what distinguish a useful mention.
4. Worrying about velocity instead of method. Acquiring manufactured links slowly is still manufactured acquisition. The pattern is what gets identified, not the rate.
5. Automating the pitch. Google's link spam policy names automated link building, publishers recognise it, and it performs worse than targeted work on total placements.
6. Targeting placements rather than approaches. You control the quality and number of approaches. Targeting acceptances creates pressure to lower the bar.
The pattern
Most come from wanting a number where there is a judgement. The judgement is which sources matter in your category, and that is answerable.
The one to do first
Find the pages that list several competitors and not you. Those are omissions rather than links to build, and publishers correct omissions when asked.
Usually a short list, usually the highest-value thing in the whole audit, and usually nobody has looked for it.
FAQs
How many backlinks do I need to rank?
There is no number. Links are one input into a system nobody outside Google can observe, and a competitor's count describes their position rather than your requirement. Set a work rate instead.
Is the competitor link gap formula useful?
No. Every input is an estimate, including the authority threshold, which comes from a third-party model rather than from Google. The output looks like a measurement and is not one.
Should I count links or referring domains?
Referring domains, since many links from one site add little. That is standard practice and it does not make the resulting figure a target.
What makes a link actually valuable?
Editorial independence, relevance to your category, a describing sentence around it, and real readership. Not an authority score, and not the dofollow attribute, which Google treats as a hint.
Can I build links too fast?
The concern is usually misstated. What gets identified is a pattern of manufactured acquisition rather than a rate, and legitimate events can produce many genuine links quickly. Acquiring manufactured links slowly is still manufactured.
Can outreach be automated?
Discovery, contact finding and tracking, yes. The pitch and the sending, no. Google's link spam policy names automated link building, and targeted individual approaches outperform volume on total placements.
What should I target instead of a link count?
Approaches made, pages where competitors appear and you do not, corrections secured, and one piece of original research a quarter. All controllable, all trackable.
What is the highest-value finding in a link audit?
Pages listing several competitors and not you. Those are omissions rather than links to build, and publishers generally correct them when asked politely.
Find the omissions
Run the competitor audit, then look for one thing specifically: pages that list several of your competitors and not you.
Those are not links to build. They are omissions, publishers generally correct them when asked politely, and almost nobody looks for them.
Then
Build the source list: every publication, roundup and directory that matters in your category.
Set a work rate, not a count. Ten to twenty individual approaches a month.
Lead with corrections, which are the fastest and most welcome.
Write each pitch yourself. Automate the finding and the tracking, not the writing.
Publish something worth linking to, ideally original data, once a quarter.
Report approaches and corrections, not a gap figure.
The thing that makes it all easier
Having something publishers want. That is a content problem, and Blazly SEO covers research, planning, drafting and direct publishing from about $39 a month.
Summary
There is no number of backlinks that produces a ranking, because links are one input among many into a system nobody outside Google can observe, and a competitor's count describes their position rather than your requirement. The gap formula in circulation looks rigorous and is built entirely from estimates, including an authority threshold taken from a third-party model that Google does not use. Run the competitor audit anyway, for the specific sites it names rather than the arithmetic, and look especially for pages listing competitors where you are absent. What makes a mention valuable is editorial independence, relevance and a sentence describing you, not a score. Velocity worries are mostly misstated: the pattern of manufactured acquisition is what gets identified, not the rate. Set a work rate of individually written approaches, and report those rather than a projected gap.