Good SEO Backlinks vs. Bad Backlinks - A Review - Episode 114

Grumpy SEO Guy

Abstract

Thirty-five minutes on which backlinks help and which waste money, built on a single move: think like a search engine whose job includes stopping people from manipulating the results. Almost every specific rule in the episode is derived from that rather than listed, which is deliberate — understanding why removes the need for a checklist, and wanting just the checklist is named as the reason people fail at this.

• Four requirements in priority order, and authority is the one the whole episode is about.
• Backlinks are the only source of authority — not speed, not content, not user experience, not dwell time.
• Anything spammy, anything automatable, anything easy to do probably will not help, because ease is what gets a pattern discounted.
• Good backlinks are one of two kinds: citations, or contextual links inside real content.
• Beyond the link type you have to judge the linking site — its IP neighbourhood, its outbound links, its inbound ones, its countries and languages.
• The master test behind every rule is simply whether the thing makes sense.

Everything below follows the episode in order, from the four requirements through the derivation, into the bad-link types taken one at a time, and out to how a linking site is judged. For anyone building a personal brand the transferable part is the method: a rule you can derive is a rule you can extend to a case nobody has covered.

Chapter summaries

00:39 - A recurring confusion is which backlinks help and which waste money

A compact review covers a recurring confusion — which backlinks help and which waste money or harm you.

"I'm the grumpy SEO guy. Let me tell you why I'm grumpy."

— 00:39

The confusion is recurring and expensive: people learn they need links, acquire the wrong ones, and conclude the mechanism does not work. A second aside is worth noting for its register — most content promising that machine answers will upend search says nothing actionable, and the change is granted as real while the commentary is dismissed as hype.

05:36 - Four things are needed to rank, in a priority order

Four things are needed to rank, in priority order: no penalties, content, relevancy, authority.

"There are four things you."

— 05:36

Four things are needed to rank, in priority order rather than in parallel. The list is the same one that runs through every episode in this series, and the repetition is the method: a closed list, restated until a listener can apply it without the episode in front of them.

05:45 - A penalised site does not rank however good everything else is

No penalty is the highest priority; however good the SEO, a penalised site does not rank.

"Number one, no penalties. Your highest priority is not having."

— 05:45

A penalty overrides every other signal, which is what makes it the first of the four and not one among them. Nothing further up the list is reachable while it stands.

06:02 - Content quality is not a ranking factor, and writing well produces nothing by itself

Content quality is not a ranking factor; writing good content does not by itself produce rankings.

"Two, content. Now, content. It's not about how good your content is."

— 06:02

The complaint it answers is stated directly — months of great content and still no position — and the answer is that quality was never the lever. The machine reads topic, not merit.

07:21 - Content is still needed for two reasons: presence, and relevancy

Content is still needed, for two reasons: to have something on the site at all, and to create relevancy.

"Thing you need is content, but it's not about the quality of."

— 07:21

Content is still needed, for two reasons that are easy to confuse with the first. It gives the site something to be, and it creates relevancy. Neither of those is about how well it is written.

08:00 - Relevancy is the topic of the site, and one reason a search surfaces it at all

Relevancy is the topic of your website — one reason a search would surface you at all.

"Content creates relevancy. Relevancy means what is the topic of your website?"

— 08:00

Content is how relevancy gets made, which is the honest role content plays in the model: an input to eligibility rather than to position.

08:46 - Authority is how much power and trust a site carries

Authority is how much power and trust a site has — the reason yours should appear over all the others.

"More thing that you need, and that is authority. Number four, authority."

— 08:46

Authority is the tie-breaker among sites that are all relevant, which is why the question is put as: of all the sites online, why should yours appear. Power and trust are the two words used for it.

09:17 - Backlinks are the only way to get authority

The only way to get authority is backlinks — not loading speed, content, user experience, dwell time, title tags or meta descriptions.

"The only way to get authority is from getting backlinks."

— 09:17

Authority is conferred by other sites linking and never by anything on the page, which the episode establishes by running through the on-page factors that do not confer it. There is no internal route to it at all.

10:00 - A link passes a share of what the linking site holds

Every site carries some authority; picture a 1–100 scale, and a link passes a share of the linker's authority to you.

"Here's how authority works. Every website on the internet."

— 10:00

Picture a scale of one to a hundred: what a link passes is a share of whatever the linking site holds, so the same link is worth different amounts depending on who sends it. If the first three requirements are handled and a site still does not rank, authority is the missing lever.

11:42 - Once the first three are handled, backlinks become the biggest part of the work

Once penalties, content and relevancy are in place, backlinks become the biggest part of SEO.

"That's why backlinks is a big part of SEO."

— 11:42

The first three are preconditions rather than levers: each one only has to be satisfied once, and after that the work that moves position is link work. That ordering is the whole shape of the discipline — three cheap requirements to clear, and then one expensive contest that never ends.

11:57 - Not all backlinks are equal, and rubbish ones do nothing

Not all backlinks are equal; people learn they need backlinks, then acquire rubbish ones that don't help and may hurt.

"Are good backlinks and there are bad backlinks, because not all backlinks."

— 11:57

People learn they need links, go and acquire rubbish ones, and get nothing for it — which is the specific failure the episode exists to prevent.

12:33 - Optimisation is a competition rather than a checklist, and wanting the list is why people fail

SEO is a competition, not a checklist; wanting "just the list" is why people fail at it.

"Backlinks. And I know people are going to try and jump forward."

— 12:33

A checklist optimises items in isolation, and ranking is relative to competitors, so the two cannot be reconciled. The listener who just wants the list is named directly — and checklists, he says, do not really exist.

12:55 - Understanding why removes the need for checklists

If you understand why something is the case you won't need checklists — the successful minority operate this way.

"Listen. If you understand the — this is what I."

— 12:55

Understanding generalises where a memorised list does not, which is the whole argument for the harder route. He puts it at roughly ten per cent of listeners — the successful ten per cent — offered as a belief rather than a measurement.

14:14 - Part of an engine's purpose is to stop people manipulating the results

Think like a search engine: part of its purpose is to stop people manipulating the results.

"So if you were a search engine, if you."

— 14:14

That single move generates most of the specific rules that follow, because every rule is an answer to what an engine would have to do about a cheap trick.

14:35 - Optimisation is manipulation, even at its most transparent

SEO is manipulation — even the most transparent white-hat SEO is trying to manipulate the results.

"Manipulation. When we're ranking a client's website, we are manipulating the results."

— 14:35

Optimisation is manipulation, even at its most transparent — ranking a client's site is manipulating the results, and the admission is made without defensiveness. It matters because the engine's countermeasures only make sense once you accept which side of them you are on.

15:56 - An engine has to build a process that separates good links from bad

Because backlinks confer authority, a search engine must build a process to separate good backlinks from bad.

"Engine, that back links provide authority. So you need to come up with a process to."

— 15:56

The obligation is structural: a signal that valuable and that cheap to fake has to be policed, or it stops working as a signal.

16:43 - Anything spammy, automatable or easy probably will not help

The quick rule: anything spammy, anything automatable, or anything easy to do probably won't help.

"Relatively brief list of good backlinks versus bad backlinks."

— 16:43

Ease is the tell, because anything easy gets done at scale and anything done at scale gets discounted. The rule is usable without knowing any of the specifics, which is what makes it the one to remember.

17:37 - Good backlinks are citations or contextual links

Good backlinks are one of two kinds: citations (a business mention with name, address and phone) or contextual (a link inside a sentence in a post).

"All right, whatever. Good back links are one of two kinds of backlinks."

— 17:37

Everything else in the episode is an application of that pair: a citation proves the business exists, a contextual link proves somebody thought it worth mentioning mid-sentence.

18:19 - Blog-comment links are easy to do, so they get done at scale and carry nothing

Blog-comment backlinks are spammy because they're easy to do, so they get done at scale and carry little value.

"So what is not contextual? Blog comments are not."

— 18:19

Easy to produce means mass-produced, and mass-produced means discounted. A blog comment could in principle be contextual and is treated as spam anyway, because the class is what gets judged.

18:44 - Forum-profile links are mass-created by tools and carry nothing

Forum-profile backlinks are spammy: tools mass-create hundreds or thousands of profiles all carrying your link.

"Are spammy. Forum profile backlinks are spammy. Why are forum profile?"

— 18:44

Forum-profile links are mass-created by tools — hundreds or thousands of profiles all carrying the same link — which is the same failure with a different surface.

19:22 - A footer link repeats on every page, which is why it reads as spam

Footer links are spammy: a footer repeats on every page, so a 70-page site links to you 70 times, which looks unnatural.

"What else? Footer links. Do footer links?"

— 19:22

A footer repeats on every page, so a seventy-page site links to you seventy times from one decision. The volume is the tell: nothing a person chose to do once produces that pattern.

20:28 - Blogroll links repeat site-wide for the same reason

Blog-roll links are bad for the same reasons: sidebar links repeat site-wide and are easy to do.

"More about that later. Anyways, and blogroll links is another kind."

— 20:28

A ninety-page site produces ninety identical links, and no site links to another ninety times for honest reasons. The early-blog-era sidebar blogroll is the origin of the pattern.

21:32 - The linking site has to be judged, not just the link

Beyond the link type you must judge the linking site itself — which is why SEO is simple but not easy, and why good SEO is expensive.

"There's more to it than that. You also have to view."

— 21:32

Link quality is a property of the source site rather than of where on the page the link sits. The concepts explain quickly; the work of judging a site does not.

21:57 - Good backlinks come from sites on unique IP addresses

Good backlinks come from sites on unique IP addresses, distinct at the first or second number (the A or B class).

"So look, you have to look at the website too."

— 21:57

Addresses that share their first or second number tend to share a host, and a cluster of links from one host is one voice repeated rather than many voices agreeing. The check exists because a network of sites on one address is one decision wearing many hats.

24:07 - A good site does not link out to gambling, pharmacy or adult sites

A good site does not link out to gambling, pharmacy or adult sites — check its outbound links before requesting a link.

"Here's another thing that good websites display. Good websites do."

— 24:07

Checking the outbound links before requesting one costs a minute and tells you what neighbourhood the link would be arriving from. The neighbourhood is readable and it is read.

24:44 - A good site has no backlinks coming from those neighbourhoods either

A good site also has no backlinks coming from gambling, pharmacy or adult sites; check with a backlink-analysis tool.

"Here's another thing. We also do not want the."

— 24:42

A good site also has no links coming from those neighbourhoods, which is the same check run in the other direction. Both are cheap to run and both are routinely skipped.

25:55 - The countries of a site's backlinks have to make sense

Check whether the site's backlinks come from unrelated countries; foreign links aren't automatically bad — they have to make sense.

"We have other episodes explain it. So another thing."

— 25:55

Foreign links are not automatically bad — a single one is fine and even welcome. What the check is looking for is a large unexplained mass of them, which no honest site accumulates by accident.

26:58 - The languages of the inbound links have to make sense too

Likewise check the languages of the inbound links; the language mix has to make sense.

"So another thing that you want to check is the website."

— 26:58

A foreign-language flood with no business reason behind it may be adult spam rather than an audience. Many foreign-language links is not a verdict — it is a prompt to check whether there is a reason for them.

28:00 - The master test behind every rule is whether the thing makes sense

The master test behind every rule is "does this make sense?" — most link questions resolve to it.

"Has to make sense. This is — people seem to not understand this."

— 28:00

The master test behind every rule is whether the thing makes sense. Most link questions resolve to it, and he notes that people seem not to understand this — which is the same complaint as wanting the checklist. The heuristic outlives every specific rule.

29:55 - Good sites show normal content rather than spun text

Good sites show normal-looking content, not spun content — thesaurus-swapped text that reads as nonsense.

"They have normal-looking content. So when I started talking about this."

— 29:55

Machine-mangled prose signals a link farm, and spun content is exactly that: a program swapping every word through a thesaurus, grammatical and meaningless. If the articles do not read right, avoid the site.

30:30 - Bad backlinks fail two ways, and doing nothing is the neutral case

Avoid backlinks from sites that are obviously AI content — offered as a roughly 70%-likely future call, not established fact.

"I'm also going to tell you now, I'm 70% sure that this."

— 30:30

Links from sites that are obviously machine-written are to be avoided, offered explicitly as a roughly seventy-per-cent-likely prediction rather than as a current rule. The hedge is stated as a number, which is more than most predictions in the field get.

32:34 - Good backlinks probably help, which is as strong as the claim gets

Bad backlinks fail two ways: worst case a penalty, neutral case they do nothing — and nothing is still a loss, because you spent money or time.

"Will help you. Bad backlinks will not help you."

— 32:34

Low-effort, automated or spammy links are eventually classed bad, which is why the positive claim stays hedged at probably help. The asymmetry is the point: the downside is certain and the upside is not.

Personal Branding Lessons

The episode is about links and the method is about reasoning. Every rule below is derived from one question — what would an engine have to do about a cheap trick — rather than memorised from a list.

Work the four requirements in priority order

No penalties, content, relevancy, authority — in that order, not in parallel. A penalised site does not rank however good everything else is, which makes the first item a knockout rather than a contributor. The order is what stops effort landing on a variable that cannot matter yet, and it is restated in every episode precisely so it can be applied without the episode in front of you. 05:36

Stop writing more content and start counting links

Content quality is not a ranking factor, and the complaint it answers is stated directly: months of great content and still no position. Content is still needed — it gives the site something to be, and it creates relevancy — but neither of those jobs is about how well it is written. If the first three requirements are handled and the site still does not rank, authority is the missing lever and more prose will not supply it. 06:02

Judge the linking site rather than only the link

Beyond the type of link, the site sending it has to be assessed — which is why this is simple but not easy. The link-type rules can be applied from a list; this cannot. Check the IP neighbourhood, whether the site links out to gambling, pharmacy or adult sites, whether links come in from those neighbourhoods, and whether the countries and languages of its own links make sense. 21:32

Reject anything spammy, automatable or easy

The quick rule covers most cases: if a link is easy to get, it gets got at scale, and anything acquired at scale gets discounted. That single test disposes of blog comments, forum profiles, footers and blogrolls without needing four separate rules — the footer one is the clearest, because a seventy-page site links to you seventy times from one decision, and nothing a person chose to do once produces that pattern. 16:43

Check the outbound and inbound neighbourhood of a linking site

A good site does not link out to gambling, pharmacy or adult sites, and does not have links coming in from them either. Both checks are cheap, both are routinely skipped, and both are the same question asked in opposite directions. The neighbourhood is readable, which means it is read. 24:07

Check that the countries and languages make sense

Look at where a site's backlinks come from and what language they are in. Foreign links are not automatically bad — a single one is fine and even welcome — and a large unexplained mass of them is the signal. The test is not really about geography; it is about noticing incoherence between what a site claims to be and who is pointing at it. 25:55

Ask whether the thing makes sense before applying any rule

The master test behind every specific rule is simply whether it makes sense. Most link questions resolve to it, and the frustration voiced about people not grasping that is the same complaint as wanting the checklist. A rule you can derive extends to the case nobody covered; a rule you memorised does not. 28:00

Expect a bad link to do nothing rather than to help

Bad backlinks fail two ways: worst case a penalty, neutral case they do nothing — and nothing is the usual outcome. That framing matters because it removes the hope that a cheap link is at least a small win. Good ones, contextual or citations on legitimate-looking sites, probably help, which is as strong as the claim is allowed to get here. 32:34

Questions

Each answer ends at the moment in the recording where it is given.

What actually makes a backlink good or bad?

Good links come in two forms: citations, which are business mentions carrying a name, address and phone number, and contextual links, which sit inside a sentence in a real post. Everything else — anything spammy, automatable or simply easy to produce — is treated as low value, because ease means it gets done at scale and scale is what a search engine learns to discount. The dividing line is whether the link was cheap to fake. 17:37

Why isn't my great content ranking?

Because content quality is not itself a ranking factor. The recurring case is skilled writers producing excellent work for months, hiring experts, and still going nowhere — and the answer offered is that the prose was never the missing variable. Content does two structural jobs, giving a site something to hold and establishing its topic, but being found depends on authority earned elsewhere. The fix is to separate making the work from distributing it deliberately. 06:02

Do I really need backlinks in order to rank?

On this model, yes — authority is the deciding factor among relevant sites, and the only thing that confers authority is other sites linking in. Not loading speed, not content, not user experience, not dwell time, not title tags or meta descriptions. Links function as the leverage that credible outside sources lend you, and once penalties, content and relevancy are handled, they become the largest remaining lever. 09:17

Which kinds of backlinks should I avoid?

Four types are named, and each fails the same way — it is easy to produce, so it gets mass-produced and discounted. Blog-comment links, which anyone can drop in bulk; forum-profile links, which software creates by the thousand; footer links, which repeat on every page so a seventy-page site links out seventy times; and blogroll links in a sidebar, which repeat identically for the same reason. No genuine site links to another dozens of times over. 18:19

How can I tell whether a website is safe to get a link from?

Judge the site, not just the link. Check that it sits on a genuinely distinct server address, that it neither links out to nor receives links from gambling, pharmacy or adult sites, and that its foreign countries and languages have a business reason to be there. Underneath every check is the standing of the site itself, and the master question is simply whether the whole picture makes sense. 21:57

Are backlinks from other countries bad for my site?

Not automatically. A single link from abroad is fine and even welcome — you cannot control who links to you, and someone may simply have liked the site. The signal to watch is a large, unexplained share from one country or language with no reason behind it: a business that trades in a place will naturally attract links from there, but a flood with no logical connection can be spam. The test is whether the provenance makes sense. 25:55

Should I worry about links from AI-written websites?

This is offered as a hedged prediction, not an established fact — roughly seventy per cent likely, on anecdotal experience, with arguments running both ways. Whether search engines detect and devalue machine-written "slop" is called a contested question. The practical stance is cautious: if a site's pages look obviously machine-generated, it may be worth thinking twice before seeking a link from it, while acknowledging the call is not certain. 30:30

Why don't cheap bulk backlinks work?

Because a search engine has to assume anything available in bulk for pocket change is worthless, and builds a process specifically to discount it. Roughly a thousand links can be bought for the price of a sandwich, and those are exactly the links the system is designed to ignore. The whole good-versus-bad distinction exists because links are so easy to buy at scale, which is what forces the engine to separate the cheap ones from the earned ones. 15:56

Is SEO just a checklist I can follow?

No — it is framed as a competition, not a checklist, and wanting the list is named as the reason people fail. A checklist optimises each item in isolation, but ranking is always relative to competitors, so the same tactic can help or do nothing depending on the field. The durable move is to understand why each rule exists; understanding generalises to situations a list never anticipated, where a memorised list quietly breaks. 12:33

Sources

Good SEO Backlinks vs. Bad Backlinks - A Review - Episode 114

The sections above follow the episode's own order. Each timestamp is the point in the recording where the idea is discussed, so the post reads as a map of the conversation rather than a rearrangement of it.

Quotations come from the episode transcript with spoken filler ("okay", "right", "you know", "um", "actually", "kind of", "sort of") and false starts removed. No words are added: square brackets mark anything inserted for sense, and an ellipsis marks any cut. British spellings are restored, and speech-to-text errors were repaired against context only where the intended word is unambiguous. Where a line appears both in a cold-open teaser and in its place in the conversation, the timestamp cites the second — the point at which it is actually said.

Quotes are stamped with the moment they are said rather than with a speaker's name.