Link Velocity: How Many Backlinks a Month Is Natural?

Ask ten agencies how many links a month is safe and you’ll get ten numbers, all delivered with total confidence. Five. Ten. “No more than twenty on a new site.” None of them can show you where the number came from, because it doesn’t exist. Google has never published a velocity threshold, and the systems that judge links don’t work by counting to a limit and then flipping a switch.

What does exist is a pattern. Links accumulate in a shape, and that shape either looks like something an audience produced or something a purchase order produced.

What does link velocity actually mean?

Link velocity is the rate at which a site gains (or loses) referring domains over time. It’s usually measured monthly, counting new referring domains rather than raw links, because a hundred links from one site is a single editorial decision, not a hundred.

The distinction matters. If a news site links to you and the article gets syndicated across their network, your tool might report 40 new backlinks in a day — but that’s one domain, one decision, one relationship. Track domains. Treat the raw link number as noise.

Velocity also has a direction people forget about: links die. Pages get pruned, sites get sold, posts get consolidated. A profile that shows 12 new domains a month and 9 lost domains a month is growing at 3, not 12. We wrote about that decay in more detail in our piece on backlink lifespan, and it changes the maths on any velocity plan.

Is there a safe number of backlinks per month?

No — not as an absolute. Anyone who tells you “never exceed X links per month” is repeating folklore, because the same X can be perfectly ordinary on one site and glaring on another. The number that matters is the one relative to your own baseline and your competitors’ baseline.

Consider two sites in the same niche. One is a four-year-old recipe blog that has organically picked up between 8 and 20 new referring domains every month since launch. The other is an eleven-month-old affiliate site with 6 total referring domains, four of which are directories. Give both of them 15 new links next month. On the first site that’s Tuesday. On the second it’s a 250% increase in the total profile inside 30 days, from a standing start, with no traffic growth or brand mentions to explain it.

Same number. Completely different signal. That’s why the question “how many backlinks per month” has no useful answer without the context of the site asking it.

Why is the right pace relative rather than absolute?

Because link acquisition is supposed to be a byproduct of something real — publishing, promotion, press, product launches, people talking about you. Those things produce a rate that’s roughly proportional to a site’s size and activity level. A big active site produces a lot of link-worthy events; a small dormant one doesn’t.

So the useful frame isn’t “is 20 links safe” but “what would have to be true about this site for 20 links a month to make sense?” If the answer is “it would need to be publishing weekly, running a newsletter, getting mentioned in industry roundups, and shipping things people write about” — and your site does none of that — then 20 is a number you’re asserting rather than earning.

The second reference point is the competition. If the pages holding positions 1 through 5 for your target term are pulling in 25 to 40 new referring domains a month between them, then your 6 a month isn’t cautious, it’s a losing pace. Velocity planning cuts both ways. Under-building is the more common failure by a wide margin.

How do you work out a sensible pace from competitors?

Pull the referring-domain growth history for the three to five sites that actually rank for your target terms, look at their monthly new-domain counts over the last 12 months, and use the median as your ceiling rather than the top performer as your target. Then check whether your own site’s history could plausibly support that pace.

1. Pick the right competitors

   Not the biggest brands in the niche — the sites currently occupying positions 3 to 8 for the specific queries you want. Those are the ones you have to displace. Amazon’s link velocity is irrelevant to you.

2. Export referring domain growth, monthly

   Most backlink tools give you a new/lost referring domains chart by month. Take 12 months if you can get it. You’re looking for the typical month, not the best one.

3. Find the median, ignore the outliers

   One competitor will have a month with 90 new domains because they got picked up by a large publication. Strip that out. The median month is what a normal operating pace looks like in your niche.

4. Compare against your own baseline

   Pull your last six months of new referring domains. If your median is 2 and the competitor median is 18, you don’t jump to 18 next month. You climb toward it over a couple of quarters, which also happens to be roughly how long it takes to see whether the links are doing anything.

5. Set a range, not a target

   Plan a band — say 6 to 10 a month — rather than a fixed quota. Quotas produce the exact pattern you’re trying to avoid: identical counts every month, which no natural profile has ever produced.

What monthly range makes sense for a site like mine?

The table below gives planning ranges by site profile. Read these as judgement calls, not measured thresholds — the reasoning column matters more than the numbers.

One deliberate omission from that table: a “maximum safe” column. We don’t have one, nobody does, and inventing one would be the most useful-looking and least honest thing we could put on this page.

What actually looks unnatural?

Volume on its own rarely looks unnatural. Pattern does. The profiles that stand out are the ones where every variable moves in lockstep, because real link acquisition is messy and paid link acquisition tends to be tidy.

Here’s what we’d flag if we were auditing a profile:

• The flat line that spikes. Eighteen months at one or two new domains a month, then 40 in March, then back to two. Nothing about the site changed to explain the March.
• Everything landing in one week. Thirty links across four days and then silence. Editorial decisions from thirty independent publishers don’t synchronise.
• One tier, over and over. Every new domain between DR 42 and DR 48, every one a general-interest blog with a “write for us” page, every one linking from a listicle. Real profiles are lumpy — a forum, a university page, a competitor’s comparison post, a Reddit thread, a DR 71 trade publication.
• Commercial anchors on everything. If 30 new links arrive and 24 of them use a money keyword, the velocity is the least of the problems. Sensible profiles skew heavily branded — we target roughly 40-50% branded, 25-35% partial or topical, 10-20% naked URL or generic, and only 5-10% exact-match. The full reasoning is in our breakdown of anchor distribution.
• Links with no traffic behind them. Thirty new referring domains and not one referral session. Links from pages nobody reads are the cheapest kind to buy and the easiest kind to spot.
• All pointing at the same URL. A natural profile spreads across the homepage, a few strong articles, and whatever page happened to get cited. Twenty-five links all aimed at one commercial landing page is a campaign, not a following.

Notice how few of those are about the count. That’s the point. You can acquire 30 links in a month across mixed tiers, mixed anchors, mixed target pages and a spread of dates, and it will read more naturally than 8 links that all look like siblings.

Can a sudden spike ever be natural?

Yes, and this is why raw velocity limits are the wrong model. Genuinely viral content produces enormous spikes — a piece hits Hacker News or a subreddit or gets cited by a large publication, and a site that averaged 4 new domains a month picks up 300 in a fortnight. That happens constantly and it isn’t a problem.

What makes it read as legitimate is that the spike has all the fingerprints of a real event. The links arrive from wildly different tiers, from DR 12 personal blogs to DR 88 news sites. The anchors are messy: the article title, “this study”, “here”, the bare domain. They almost all point to one specific piece of content rather than a commercial page. There’s a corresponding jump in direct traffic, branded search, and social mentions. And the tail decays the way attention decays — a hard peak, then a long slope down over weeks.

A purchased burst has none of that. Uniform tiers, tidy anchors, commercial targets, no traffic signature, and it stops dead rather than decaying. So the honest version of the rule isn’t “don’t spike.” It’s “don’t manufacture a spike that has no event behind it.”

How should you pace a bought or outreach-driven campaign?

Spread placements across the month rather than delivering them in a batch, vary the target pages, and keep the mix of source types deliberately uneven. If you’re buying, the vendor’s scheduling matters as much as their site list.

This is one of the practical arguments for a link insertion (also called a niche edit) over a batch of freshly published guest posts — placing links inside articles that are already indexed and already have readers means the placement lands in existing context. An existing article has its own age, its own traffic, and its own link profile; a link added inside it inherits that context instead of arriving as part of an obvious cohort. Our own turnaround runs 7-14 days per placement, which naturally staggers delivery rather than dumping everything on day one.

A few pacing habits worth adopting regardless of who’s doing the work:

• Vary the monthly count. Six, then eleven, then seven. Never the same number twice in a row.
• Point at least a third of new links at informational pages, not just the pages that make money.
• Keep earning some links outside the campaign — digital PR, a genuinely useful free tool, original data. Those are the links that make the bought ones look ordinary.

How long before velocity changes show up in rankings?

Plan on a quarter minimum before you judge anything. Links need to be crawled, evaluated, and given time to influence pages that are themselves competing on content and intent — and in competitive niches the lag between acquisition and measurable movement commonly runs 8 to 16 weeks.

This is the single most common reason campaigns get abandoned early. A client buys three months of links, sees nothing by week six, panics, doubles the order, and now has both a velocity spike and an unanswered question about whether the first batch worked. Set the pace, hold it, measure at the end of the quarter.