| What this covers What a Referring Domain Actually Is Why the Count Rises While the Authority Does Not The Only Link Types That Change Anything How Many Links Does a Business Actually Need What Standard Link Advice Leaves Out Anchor Text Is Where Campaigns Go Wrong Questions to Ask Before You Pay What to Do If the Profile Is Already Full of Junk The Half of Search That Ignores Links Entirely Where Links Sit in the Bigger Picture The Number Worth Watching Instead |
Two websites can have the same number of referring domains and be in completely different positions. One ranks. One does not rank for a single term.
That gap is where most link building budgets quietly disappear.
Referring domains – how many separate websites link to you – is the number almost every report leads with. It is a reasonable thing to measure and a terrible thing to measure on its own, for a reason that should be said out loud more often: a link only passes value if the site giving it has value to pass.
What a Referring Domain Actually Is
A referring domain is one website that links to yours, counted once no matter how many times it links.
That distinction matters more than people expect. Fifty links from one blog count roughly as one referring domain. One link each from fifty different sites counts as fifty. The second is worth far more, because the signal being measured is how many independent parties vouched for you, not how many times one party did.
So counting referring domains rather than raw backlinks is already an improvement. It is just not the finish line.
Why the Count Rises While the Authority Does Not
Junk links do not usually drag a site down. They simply do nothing. The referring domain count goes up, the report looks busier, and the site’s actual standing has not moved.
This is why a business can buy a package advertising hundreds of links, watch the number climb monthly, and see no change in rankings. The links were counted and disregarded.
The practical consequence: if the count is rising and nothing else is, the links are not being credited. That is a signal to change approach, not to buy more of the same.
The Only Link Types That Change Anything
Not every link is trying to do the same job, and confusing the two is the most expensive mistake in this whole discipline.
| Link type | What it actually does | Worth paying for? |
|---|---|---|
| Cheap network links | Helps a new page get discovered and indexed | Only for indexing, never for rankings |
| Directory listings | Confirms your business details are consistent | Yes, but once, and mostly free |
| Local sponsorships | Real editorial endorsement from a real organization | Yes – the strongest option for most local businesses |
| Chamber and trade memberships | Credible, relevant, verifiable | Yes, and cheap |
| Guest articles on real publications | Context, relevance, a genuine audience | Yes, if the publication is real |
| Bulk link packages | Inflates a count | No |
The first row gets misrepresented most. Those links have a legitimate purpose – helping a new page get found. They are not a ranking strategy, and any vendor selling them as one is selling the wrong thing.
How Many Links Does a Business Actually Need
Fewer than most people are told, and the honest answer depends on the market rather than a formula.
Kansas City is a useful illustration because the metro spans two states. The Census Bureau defines metropolitan areas by economic integration rather than by state borders, and the metropolitan area definitions put Kansas City, Missouri and Kansas City, Kansas in a single market. A business in Kansas City, Kansas competes against firms in Kansas City, Missouri, Overland Park and Olathe – one metro to a searcher, separate jurisdictions for licensing, service areas and citations. A searcher typing “near me” does not observe the state line. Directories, licensing bodies and chambers of commerce do.
That has a direct effect on link strategy. A business sitting on a metro boundary has two regional ecosystems available to it – two sets of chambers, two sets of local media, two sets of community organizations. Most competitors only work the side they are registered in.
Where competitors have thin or no websites, a handful of links is often enough and sometimes none are needed. Where they sit on pages two to four, a small number of quality links moves things. Where they dominate both the map and organic results, expect sustained effort and everything else done well alongside it.
The instinct is to assume more is always better. Ten credible links outperform two hundred forgettable ones, and keep outperforming them, because the two hundred were never counted.
What Standard Link Advice Leaves Out
Three things get repeated constantly and deserve qualifying.
“Relevance is essential.” Relevance helps. It is not a hard requirement. A strong link from a credible general publication still counts. Turning down good opportunities because they are not a perfect topical match is a self-inflicted limitation.
“Links are the main ranking factor.” They are roughly half the picture. Content and, once you are near the first page, user experience carry the rest. Sites with strong links and weak pages stall, and the owners usually respond by buying more links, which does not address the actual constraint.
“Disavow the bad ones.” Usually unnecessary. Search engines are built to ignore low-quality links rather than penalize sites for them, which is why most junk simply sits there doing nothing. Reserve disavow for genuinely toxic cases – bought schemes, or links from sites removed from the index entirely. Used carelessly, it strips out links that were helping.
Anchor Text Is Where Campaigns Go Wrong
Anchor text – the clickable words in the link – is the fastest way to make a link profile look manufactured.
A natural profile is mostly your business name, a decent share of generic phrases, some partial keyword matches, and only a small fraction of exact-match keyword anchors. When that last category dominates, the pattern is obvious, because real editors do not link to a plumber using the exact phrase “emergency plumber Kansas City” fifty times.
A useful rule: if you would be uncomfortable showing the anchor list to a search engineer, it is already too aggressive.
A profile that reads as natural tends to break down roughly like this:
| Anchor type | Example | Rough share |
|---|---|---|
| Branded | the company name | Largest single share |
| Generic | “read more”, “this guide” | Substantial |
| Partial keyword | “their roof repair work” | Substantial |
| Exact keyword | “roof repair Kansas City” | Very small |
| Naked and variations | the raw address, near-misses | Modest |
The exact-match row is the one that gets abused, because it is the one that feels most like it is doing something. It is also the one that makes a profile look bought. Restraint there is not caution for its own sake – it is what keeps the rest of the profile credible.
Questions to Ask Before You Pay
Anyone selling links should be able to answer these without hesitating.
- Which specific pages will these links point to, and why those pages?
- Can I see examples of publications you have actually placed on?
- What anchor text will be used, and what is the spread?
- Are these permanent, or do they disappear when I stop paying?
- What is the plan when the easy placements are used up?
Vagueness on the second and fifth questions is the most telling. Anyone who has done this work has examples, and anyone thinking past the first ninety days has an answer for what comes next.
What to Do If the Profile Is Already Full of Junk
Most of the time: nothing dramatic.
Start building genuinely good links and let the junk sit there being ignored. The ratio improves as real links accumulate, and no risky intervention was required. The exception is a profile with links from de-indexed sites or an obvious paid scheme bought previously – that is worth a careful review.
Where a business does need help untangling this, working with a team that handles off-page SEO for small businesses rather than a bulk link vendor is the meaningful difference – the first is auditing what you have before adding to it.
The Half of Search That Ignores Links Entirely
Here is the part that makes 2026 different from the years when link building advice was written.
AI assistants do not use backlinks the way search engines do. When someone asks an assistant to recommend a plumber or a marketing agency, the answer is assembled from mentions, citations, reviews and structured data – not from a link graph. A business can have an excellent backlink profile and still be invisible in that answer.
The inputs that matter there are different:
| Signal | Classic search | AI assistants |
|---|---|---|
| Backlinks | Roughly half the picture | Very little weight |
| Brand mentions without a link | Minor | Substantial |
| Review content and specificity | Affects clicks | Directly quoted |
| Structured data on the page | Helpful | Heavily relied on |
| Presence in forums and communities | Minor | Frequently cited |
Two consequences follow, and both change how a link budget should be spent.
An unlinked mention now has real value. For years the industry treated a mention without a link as a near-miss worth chasing down and converting. In AI search it counts as it stands. That reframes what a placement is worth, and it follows the broader shift in how people find information online that Pew Research tracks – answers arriving in place, rather than as a list of destinations.
Informational queries are worth less than they were. AI Overviews answer a growing share of “what is” and “how do I” searches directly, which means ranking first on those terms increasingly produces an impression rather than a visit. Commercial and transactional searches – the ones where someone is choosing a supplier – still send clicks. Weighting a content and link plan toward those is now the sensible default.
None of this makes links obsolete. It makes links one of two systems worth building for, and only one of them counts them.
Where Links Sit in the Bigger Picture
Links are necessary and insufficient at the same time.
A new page needs at least one external link so it gets found and treated as real. Beyond that, links compete with everything else the site is doing. A page that loads slowly or was clearly written for a search engine will not be rescued by more links pointing at it.
The sequence that works: make the page worth linking to, get it indexed, then earn a small number of credible links. The reverse – links first, quality later – is how budgets get spent with nothing to show.
The Number Worth Watching Instead
Stop leading with referring domain count. Ask three questions instead:
- Are these links from places a real person might actually encounter?
- Is the count going up while nothing else is?
- Would I be comfortable explaining how every one of them was acquired?
Those three tell you more about a link profile than any single figure on a dashboard. The count is a symptom. What produced it is the thing that matters.