The definition, and the two parts that matter
A backlink — also called an inbound link — is any hyperlink on another website that points at a page on yours. That's the whole definition. The interesting part is what Google does with it.
The original idea was simple: if lots of pages link to a document, the document is probably worth reading, and links from important pages should count more than links from unimportant ones. That's PageRank, it's Google's own patent, and in one form or another it's still running.
Two components of any link carry meaning. The source — which site, which page, how trusted, how relevant to your topic. And the anchor text — the visible words the link is wrapped around, which tell Google what the destination page is about. A link from a respected industry publication reading "SEO pricing in India" says something specific. A link from a scraped directory reading "click here" says almost nothing, twice over.
The four rel attributes and what each one signals
Every link carries an instruction to search engines in its rel attribute, or carries none at all. The rules changed in 2019 and a lot of advice still hasn't caught up.
- The practical read: chasing dofollow links exclusively is a strategy from 2015. A profile with zero nofollowed links looks manufactured, because real brands get mentioned on platforms that nofollow by default.
- The full comparison lives in dofollow versus nofollow, including what each is worth in practice.
| Attribute | Where you find it | What Google does with it |
|---|---|---|
| dofollow (no attribute) | Editorial links in articles, resource pages, genuine citations. | Passes ranking signals. This is the link everyone means when they say "backlink". |
rel="nofollow" | Wikipedia, many news sites, forums, most social platforms. | Since March 2020, treated as a hint — Google may use it for crawling and ranking, or ignore it. Unpredictable, not worthless. |
rel="sponsored" | Paid placements, affiliate links, advertorials. | A declaration that money changed hands. Passes no ranking credit, and declaring it is what keeps you compliant. |
rel="ugc" | Blog comments, forum posts, user profiles. | Marks the link as reader-submitted rather than editorial. Discounted accordingly. |
The discount curve: where a link's value goes to die
This is the part nobody selling links wants discussed. A backlink's value is discounted — sometimes to zero — by a stack of factors that have nothing to do with whether the link exists. Here's roughly where the value drains out, worst first.
- Sitewide footer and sidebar links. Repeated across every page of a site, so they inflate your total link count enormously. Google collapses them; the practical value sits closer to a single link than to thousands. This is how agencies produce "we built you 12,000 backlinks" reports from four placements.
- Links on pages nobody links to. A link's power comes from the linking page's own authority. A brand-new page on a strong domain that no one has linked to yet passes very little, regardless of the domain's rating.
- Links on pages Google hasn't indexed. If the page isn't in the index, the link isn't being counted. A surprising share of purchased guest posts sit on unindexed pages, which is why they're cheap.
- Generic directory listings. Unmoderated "submit your business" directories exist to sell links. Google has been discounting them since roughly 2012.
- Blog comment and forum signature links. Nearly always
ugcornofollow, and pattern-obvious at scale. - Paid links that don't declare themselves. These violate Google's link spam policies. Google's systems now largely neutralise them, which means you paid for nothing — the more expensive failure mode than being penalised, because nothing visibly breaks.
- Private blog networks. Sites built to sell links to each other. They work until the network is identified, and then everything they touched loses its value at once.
Editorial, placed and bought
Every link falls into one of three categories based on how it came to exist. This is the classification that predicts value best, and it's the one that never appears on a link-building invoice.
| Type | How it happened | Realistic value |
|---|---|---|
| Editorial | A writer found your page useful and cited it, unprompted. | The highest, by a distance. Also the hardest to manufacture, which is exactly why it works. |
| Placed | You pitched — digital PR, an expert quote, original data, a partnership. A human editor said yes. | Genuinely valuable. Most legitimate agency link work lives here. |
| Bought | You paid a site or a broker for placement, undeclared. | Against Google's policies. Increasingly neutralised, meaning you paid for a link that does nothing. |
How many links a page actually needs
The honest answer is that it depends entirely on the query, and the honest method takes ten minutes rather than a guess.
Search your target keyword. Take the pages ranking in positions one to five — the specific pages, not the domains. Check how many referring domains each of those *pages* has. The median of those five is roughly your target. Not the domain totals, which are dominated by a homepage's history and tell you nothing about what this page needed.
- Long-tail informational queries frequently need zero external links. A well-structured page with good internal linking, on a site with any history, competes fine.
- Body-term service pages in a competitive Indian vertical typically need somewhere between a handful and a few dozen genuinely independent referring domains — and this is where most link budgets should go.
- National head terms in finance, insurance, travel or e-commerce need hundreds, built over years, alongside a brand people already recognise. If your budget is ₹40,000/month, this isn't the fight to pick this year.
- If the top five have fewer referring domains than you do and still outrank you, your problem isn't links. It's relevance, format or intent — and buying more links is the expensive way to avoid finding that out.