What link velocity actually measures
The literal definition is simple. Link velocity is the rate at which a website gains or loses backlinks over a defined window, most commonly expressed as new referring domains per month or per 90 days (source: Linkbuildingjournal.co.uk). The arithmetic is trivial: forty new referring domains in 30 days yields a velocity of 40 in that window.
What matters for an SEO operating in 2026 is what the metric actually tells you, and what it tells Google. Link velocity is a time series, not a count. Every link-aging analysis Google has shipped since Penguin treats the shape of your backlink curve as a signal in its own right. A stagnant flat line that suddenly spikes reads differently from a gradual ramp matched by content publication. Google does not need to measure your velocity explicitly. The velocity is implied in any time-aware view of your backlink graph, and an unnatural shape sets off downstream classifiers long before any single link is judged on merit.
This is why velocity sits at an awkward intersection. It is not a ranking factor in the sense that more velocity equals more rankings. It is closer to a behavioral envelope: stay inside the band that looks plausible for a site of your age and authority, and you operate under the radar. Step outside, and you invite a closer look at every link you have.
How to actually measure it in 2026
Three tools surface link velocity directly: Ahrefs (referring domains chart with monthly granularity), Majestic (historic trust flow charts back to 2012), and Semrush (backlinks over time view). Pick one as your reference and stick to it. The absolute numbers will differ between tools because their crawlers cover different subsets of the web, but the shape of the curve is what you read.
The useful unit is a rolling 30-day count of new referring domains, complemented by a 90-day rolling average for context. The 30-day window catches recent campaign activity. The 90-day average smooths noise and reveals whether you are trending up, flat, or losing ground. A site at zero new RDs for three months that suddenly hits 40 in a week is more suspicious than a site averaging 10 per month that bumps to 25.
Loss velocity is the half nobody tracks, and the half that often explains ranking erosion. If you gain 10 RDs and lose 12 in the same month, your net velocity is negative regardless of how good the new links are. The Content Gravity Model methodology scored 2,500+ assets and correlated CGM scores with 90-day referring-domain velocity from Ahrefs and Majestic data: scores in the 25-34 band showed roughly 3x baseline velocity, scores 35-44 showed 9x, and scores 45-50 showed 20x (source: DigitalApplied.com, The Content Gravity Model 2026). The takeaway is not the multiplier itself but the structural point: velocity is downstream of asset quality. Pages that earn links earn them at a rate proportional to how linkable they are.
To calibrate a baseline, audit three to five direct competitors at similar age, authority, and vertical. Map their 90-day velocity and use the median as your floor. Anything above the top of that range is the band you should not exceed without a matching content or PR justification.
Where it matters in a netlinking operation
The single decision velocity drives in a paid netlinking program is cadence. If you have budget for 12 placements, the question is whether you book them in two weeks or spread them across 90 days. The answer depends on the site receiving the links and on what it has been doing in the months prior. A site averaging 8 organic RDs per month can absorb 12 placements over 90 days without distorting its curve. A site at zero RDs for half a year cannot, regardless of placement quality.
This is where spreading a campaign over a realistic window matters more than any individual metric on the linking site. We see operators panic-buy 15 placements at the start of a quarter to use the budget, then go silent for two months. The pattern reads like a paid burst because it is one. The same 15 placements split into batches of 3 to 5 per month, integrated with editorial publication on the receiving site, produce the same authority gain without the silhouette.
Velocity also dictates how you stage anchors. The anchor profile of newly acquired links matters more than the lifetime average, because Google reads recent acquisition patterns more aggressively than historical ones. A spike of 8 commercial-anchor placements in two weeks is the textbook over-optimization signal. The same 8 placements with diversified anchors, spread over 8 weeks, lands in safer territory.
For sites under 18 months old, the constraint is tighter. Young sites have shallow trust and proportionally less tolerance for velocity spikes. The pragmatic rule is to stay below 2x the median velocity of comparable sites in your vertical at your age, and to never exceed 1.5x your own trailing 90-day average without content or PR justification.
What we see go wrong in audits
The most common pattern in audit work is the launch spike. A site rolls out, pushes 20 to 40 RDs in the first month through a press release, agency burst, or platform mass-buy, and then settles into near-zero acquisition for the next quarter. Six months later the site stagnates, the operator blames the algorithm, and the Ahrefs chart shows the exact moment the trajectory broke. The fix is not more links. It is rebuilding a steady baseline of 2 to 5 RDs per month for two quarters before the next campaign.
The second pattern is the burst-plus-anchor mistake. Operators decide their commercial keyword is weak and book 10 placements pointing money-keyword anchors at the same target page in the same month. Even if every linking site is clean, the combined pattern of velocity plus anchor concentration triggers downgrades that take months to recover from. A defensible default is no more than 30% commercial anchors in any 90-day window on the same target, paced to whatever the site was doing organically.
The third is ignoring loss. A negative SEO attack or a publisher cleanup that strips 50 RDs from your profile in one month is a velocity event in the negative direction. Most operators only notice when rankings drop. The defensive move is monitoring loss velocity weekly and disavowing or replacing high-loss-rate clusters before they compound.
The Search Engine Roundtable cited statistic that 66.31% of pages have zero backlinks (source: Backlinkier.com) is useful framing here. Most pages on the web do not earn links at all. The velocity discussion only applies once you are above the floor. Below it, the conversation is about acquiring any links, not pacing them.
An operational playbook
Map your trailing 90-day velocity in your tool of choice and write down the median monthly figure. That is your baseline. Map the same number for three to five competitors at comparable age and authority. The band between your baseline and the competitor median is your safe operating window.
For paid placements, match the campaign size to the window. Twelve placements over 12 weeks is almost always safer than the same 12 over 4 weeks, regardless of where the links sit. We source placements directly from owned publishers rather than through marketplaces precisely because owned inventory lets us pace to the campaign rather than to platform availability.
Synchronize netlinking with content publication. A site that publishes 8 articles in a month and acquires 6 new RDs in the same month produces a coherent signal: the site is producing assets and the assets are getting cited. A site that publishes nothing and acquires 6 RDs produces a less coherent signal. The discipline is to never run a netlinking phase without a matching content phase, even when the links target older pages.
For larger programs, the media catalog open without registration is useful precisely because you can pre-plan a 90-day calendar of placements rather than discovering availability one batch at a time. This kind of forward visibility is what turns velocity from a constraint into a parameter you set.