Tom Belfort's AMA about SEO/Siloing/PBNs/Link Building 2019+

yesfuture said:

So when you post an article how long do you wait before you start building links to it?

Click to expand...

BlueBuzzy said:

@splishsplash wow thank you so much for the detailed answer! I just have a couple follow up questions - I hope that’s fine.

1) I’ve read you recommending either guest posts or PBNs (or both) as your way to go when it comes to link building. So getting 10 or so guest posts for $1000 a month will payoff after 6 months or so? I know there is no way to tell for sure, but is there a rule of thumb of some kind?

2) if you point good guest post links to one post, will other posts on the site get an indirect boost as the authority of the property increases?

3) have you ever noticed a difference between having you your hosted on a shared hosting / VPS etc. from the seo perspective?

And lastly,

4) I know it’s subjective, but how many articles (or thousands of words) do you need to post? I notice my site ranks for a few thousand keywords (quite a bit in the past couple months) but it doesn’t seem to have much correlation with traffic. Is it because they just enter the top 100 and need time to get closer to the top to get noticed? Just curious what you think.

Thank you so much, Tom!

Click to expand...

splishsplash said:

The most important thing here is. You do NOT rank with keyword density. AT ALL. 0.2% or 1.5% there is no positive correlation between ranking. What you ARE doing here is de-optimizing. So when I'm talking about keyword density, I am helping you to not over-optimize, and this is only ever an issue if you have an "SEO writer" or a low quality writer.

Ratios. Simple. Forget all the "calculate the perfect ratio from the averages of the top 10" stuff. 70% brand, 20% naked, 5% misc, 5% title/lsi. 0% exact. 0% partial.

Click to expand...

splishsplash said:

The most important thing here is. You do NOT rank with keyword density. AT ALL. 0.2% or 1.5% there is no positive correlation between ranking. What you ARE doing here is de-optimizing. So when I'm talking about keyword density, I am helping you to not over-optimize, and this is only ever an issue if you have an "SEO writer" or a low quality writer.

Click to expand...

splishsplash said:

Too broad of a question. This is like asking "How do I run a successful restaurant". Get in the top 5 by applying all the knowledge you get from this thread is the best answer I can give.

Your second question is too broad too. Yes, maybe and no is the answer. It's like saying "Will I get get hit by a truck if I go out today?"





Web 2's struggle to index. They have for a long time. Add more unique content to them to index them. I don't bother with web 2's because I start with aged domains.

Don't bother submitting web 2's to gsc. Too much work. Too much effort for no gain. Create a few, put a little extra content on them and move on.




Test your writers first with a custom piece piece. Never accept 'samples' other than an initial screening.

Always train your writers. Give them guidelines and examples. A well trained writer is like gold.



Yes, I do. I display the first 15-20, then use an ajax load more for loading the older ones.



Yes it matters, otherwise it's not siloing. Well, it's not physical siloing, it's just a form of virtual siloing. Physical are the most powerful.

Follow A1 and it'll be setup correctly.




Again, this one is too broad. It's unanswerable. Follow all the knowledge contained in this AMA and my other guides.



You shouldn't have immediately removed it. When you make changes it's common for a page to drop a few spots then rise up higher. Always give your on page changes 10 to 15 days before re-tweaking. But I don't recommend changes like adding a bolded word. In fact, I don't recommend changing ANYTHING on-page for pages that are on page 1/2. Page 3, sometimes, page 4+, yes. Here's some guidelines:-

Q5) What kind of on-page changes do you recommend I make to existing pages that are ranking?

A5) Multi-part answer..

1. For pages on your site that are on google page 1 and 2, don't CHANGE anything on-page.
2. Page 3. Depends on the keyword and page. It's a maybe.
3. Page 4+ you can change things.

HOWEVER. This doesn't mean you go changing all your page 4+'s. It depends on how competitive the keyword is. If it's low comp and you're on page 4-10, then something is probably wrong. If it's medium comp and you're page 4, you're probably ok, and if it's page 7+, then there may be something wrong. Understand? It's quite situational. You have to ask, and you can only do this with experience: "Should my page naturally fall here based on a) the difficulty of the keyword, b) the topical authority I have for the topic of this keyword and c) the overall strength of my site"

Hence the above are general rules. Especially 1. Never change anything for something on page 1/2.

Now, when you ARE changing, you generally want to de-optimize keyword density.(See A6) and especially deoptimize sub-titles(h2/h3/h4's). Don't do weird things. Just de-optimize when changing.

Now, notice I used the word "CHANGE". What you can always do, to any page on your site, whether it's number 1 or page 8 is ADD to it.

You can always add another paragraph, or add another section.

The exception to this is if you've got a list of something that needs to change like, "Shops you can buy X in", where you might be adding and removing.

After any change or adding, expect results in 3 to 20 days. Usually 7 to 12 days you'll see movement and it is common to DROP before you rise. A drop from #8 to #11 after 3 days for example, then it maybe dances between 8 and 13, then after 15-20 days it goes up to #6-7. This is common in google, with on page changes and adding new links. It's part of the algorithm to add confusion. Delays and dancing. If you made a change and 1 hour later you rose 2 spots, it would be VERY easy to figure out the algorithm. This is also why when you remove something, you won't just snap back to where you were. It makes it hard to pin-point specific changes to movements.


See this Google patent: http://patft1.uspto.gov/netacgi/nph...8,924,380.PN.&OS=PN/8,924,380&RS=PN/8,924,380

Q6) What keyword densities do you recommend?

A6) The quick answer is < 1.5% for trips and < 2% for pairs. I'd even say go for < 1% for trips and < 1.5% for pairs these days.

The most important thing here is. You do NOT rank with keyword density. AT ALL. 0.2% or 1.5% there is no positive correlation between ranking. What you ARE doing here is de-optimizing. So when I'm talking about keyword density, I am helping you to not over-optimize, and this is only ever an issue if you have an "SEO writer" or a low quality writer.

The longer answer..

You want to use as many "LSI" keywords as possible. You really don't need to mention a keyword more than once. If it shows up more than once naturally, that's fine. A good article written by a good writer will naturally be fine. Avoid "SEO writers" like the plague. Avoid writing "for SEO". A good, deep article will contain a ton of topical relevance and juicy keyword variations(LSIs).

The more variation you have in keywords the more topical relevance you build. So if you're writing about "drone reviews for teens", then you'd want to say, teens, teenagers, young adults, 13 year olds, 14 year olds, 15 year olds, reviews, top, best, great, amazing, our picks. Everything closely related. The days of ranking for a keyword are LONG LONG LONG LONG gone. Google is so advanced with content. I'll write about this in more depth another time, but for now just know you need natural, plenty of keyword variations and depth in your content.

Keyword density is an old, dated metric from before hummingbird and the knowledge graph, when google would rank based on keywords. It does NOT rank based on keywords now. It ranks entities when the following occurs:- A page matches a search intent. It's also worth mentioning here, if one page more closely matches searcher intent than another stronger page, it will outrank that page. If 2 pages closely match for searcher intent, the stronger, more trustworthy one will rank.




1. Don't over-link all over the place. Link mostly contextually, where is natural.

2. A silo's main purpose isn't to drive link juice. That's more like "PR sculpting". It's a misconception about silos. A silos purpose is not to artificially increase ranking by manipulating juice, which is why it'll never be blackhat. Its true purpose is to help google understand the topics of your site. To show google what your site is about, what its an authority in. This is why the linking isn't so important. Google is very good at understanding content. By linking from a dog silo to a cat silo, it won't confuse it. If anything it's a positive, because it understands how the 2 are related. If you as a human know something is related, so does google. So if you deem it a good link, it's good for google's understanding too.

3. You don't rank category, tag or archive pages. They have nothing to do with silos. An ecommerce site can rank its product pages, yes, and you should have them inside the correct silos. You would have the brand name in them. Nothing wrong with that. If I am selling a Samsung 22" Business Monitor S22E450BW, it wouldn't make sense to miss out "samsung" from the title. How will people know it's a samsung? People will search for samsung 22" monitor, or samsung s22e450bw.

Samsung.com using siloing.

https://www.samsung.com/uk/monitors/monitor-se450-s22e450bw/


So here we go. We don't need to mention samsung in the url. It is samsung.com. Likewise if we have site.com/uk/monitors/samsung/monitor-se450-s22e450bw, we don't need to have site.com/uk/monitors/samsung/samsung-monitor-se450-s22e450bw, BUT,

Look at the title.. "Samsung 22" Business Monitor S22E450BW | Samsung UK"

Then

The H1 is just 22" Business Monitor S22E450BW

Makes sense, right?

On the serps, people are looking at the title, so you want to say samsung monitor.

The url doesn't have to repeat, because it already is implied it's samsung.

The h1 doesn't need to say samsung, because, you've just clicked, you are on samsung.com or in site.com/uk/monitors/samsung and you can see the title.

Have a look at how samsung.com are siloing. It's an EXCELLENT example. Their architecture is spot-on.






I haven't tested this. It's believed that changing the niche devalues it. It won't be a 100 or 0 situation. It'll be a sliding scale, like the further away from the niche it is, the more the links are devalued. I wouldn't ever expect them to be devalued completely.

This is a great time to test though. Please let me know what happens after you place a link.



Q7) Which keyword tools do you recommend?

A7) semrush and ahrefs. For keyword research I google some base keywords, then pop the sites in ahrefs or semrush, and go from there. You'll get a ton of keyword ideas, and you can further search google and see even more sites. It is by far the best way to do kw research.





1. Don't spend money on kw research other than an ahrefs or semrush subscription. It depends how low is low. SEO is hard with a low budget. I don't see many people succeed in SEO with budgets of less than $1k/mo unless they personally have the ability to write a ton of content and do their own outreach. You need content and links. The cheapest way to do it is, like I said, write your own content and do your own outreach. You can get free guest posts, but you need to be able to write well.

2. $1k/mo is good. $1k on content month 1/2/3, then $700 content, $300 links month 4/5/6, then 50/50. It's easier to rank authority sites than micro niche sites. I almost never see small sites rank these days.

3. It matters. But it's not about the domain age. It's about the age of the content and the links. It can take a couple of months for links to kick in properly, and sites do definitely get better with age. Add a new article to a 6 month only site vs a 3 year old site and you'll see the importance of age.

4. You don't want expired domains. You want non-dropped, aged domains. And for a money site, $300 to $5000. Generally $700-$1200 you'll find something good.

5. There's not a "biggest". If your on-page isn't right, links won't do anything. On-page is the direction you are pointed, links are the push. If you are pointed south, and you want to go north, the biggest push in the world won't help. LIkewise, if you are pointed north, and don't push, you won't move.

6. Yes, it does. Google likes authority sites. Also, which, as a human would you consider an authority on animals. A site with 5 pages and 10k words of content, or a site with 750 pages and 5 million words of content? Trying to rank for keywords is silly. 1 article, ranking for 1 keyword is 2012 stuff. An article ranks for a tight topic, and a site ranks for a broad topic. The bigger your site, the more keywords you'll rank for. The bigger the article, the more keywords you'll rank for. More longtails.




It varies. Sometimes 2 weeks, sometimes 2 months. The algorithm has a lot of randomness to confuse and make it difficult to assess what works and what doesn't. Usually never less than 2 weeks, and usually never more than 2 months, so if you have no results after 2 months, what you did, did not work.



Nope.

https://www.blackhatworld.com/seo/the-ultimate-guide-to-seo-siloing-dominate-google-in-2018.985967/



75-200 words in silos, and 500-1000 on the homepage. For informational sites. Ecommerce is different. I'll write a siloing for ecommerce guide soon.



yes correct.

$50-$100/mo is not enough unless you are able to write at least 2k-3k words a day yourself and do your own outreach, and writing the guest posts. There's really nothing you can do with $50-$100. $50 isn't even enough for a semrush or ahrefs subscription. That kind of money will cover your hosting and 2-3 articles a month, or 5-10 pbn links and nothing else. Ideally you want at least $250/mo. $500/mo is good and $1k/mo is great. More than that is only when you really know what you're doing and what your ROIs are.





I've just had a few clients recently who have had some issues around them so I'm recommend people avoid just now. They will still work, but because of that I'm recommending people avoid unless they already know what they're doing. Niche edits can be expensive too, so it's a big risk to drop several hundred $ or a grand and have it hurt your rankings or do nothing.

If you're going to do them start with only 1-2 and see how it works for you and keep them relevant.






I just googled dogs and best dogs. They are very different. Can you show me a review page ranking for 'dogs'? Google understands intent, and when you type in dogs, you don't want to know about the best dogs. You want information on dogs, maybe dog breeds etc.

I don't see ANY review/best pages on page 1 for 'dogs'. I see dog news articles, rich snippets with info about dogs, list of dog breeds, complete guide to caring for dogs, dog - wikipedia, dog health center, a dogs page on humanesociety.org, dogs available for adoption, dog breeds again, and dog supplies.

If you want to rank for dogs, then you need to re-assess your seo strategy. Trying to rank for single keywords in 2019 is completely the wrong strategy. It's even worse trying to rank for broad keywords. You should be trying to rank for 250k keywords in a broad topic, like 250k keywords in the musical instruments niche. THAT is how you make money with SEO and build a monster site.

Click to expand...

BBulldog said:

Hey man, thanks for doing this. Since you already answered a bunch of questions about SILOing, I thought I'd ask a bit about PBNs and domains.

PBNs
1. How much content does your PBN start with?
2. After building it, how long do you wait before adding a link to any of your websites?
3. What is the ratio of non-link posts/link posts or do you link to a site in every post?
4. Do you feel your PBN service was underpriced?

Domains
1. When choosing a domain for a PBN site, what do you value more: quality links (DA > 80) or more referring domains (as long as they are not spam)?
2. Do you only buy expired domains from sellers or do you buy from auctions/scrape them yourself?

Click to expand...

davids355 said:

You've answered the two questions I was going to ask. I've been doing something similar to this for a while, very low or zero exact/partial match anchors and lowering on-page optimization. I does seem to vary depending on the niche for me, but definitely its good to see you putting this in writing.

I have a few follow-up questions as well;

Do you think the optimum figures for anchor text and keyword density vary dependent upon niche? Does Google take some sort of average/mean figure across the board for each keyword/niche for example?
Also, do you think Google takes those two factors into joint consideration? For example could you afford to have a higher on-page density if exact match anchors were lower?

And a new question I have as well;

How many keywords or phrases do you suggest targeting per article? For example, would you target several closely related keywords?

Click to expand...

splishsplash said:

I take it you're asking about the wolf.. I can't really talk about that here. It's against the rules to promote services outside of the marketplace thread.

I'll respond in PM to these questions for you.






I don't think they vary across niches. I think that whole calculating the anchor ratios for each keyword is just something SEOs decided was potentially cool. The fact that some top 10's for a keyword have higher than normal exact matches doesn't necessarily mean google is looking for that. There are SO many confounding variables at play. Ratio to the page, ratio to the overall site, age of the site, when they got the links, their trust/authority. A site with higher trust can get away with a lot more. So what you could be seeing is a niche with older, higher trust sites, who just happen to have had more aggressive link building done, but if we try to replicate that, we will not make it up there.

From what I know about Google, I just cannot ever imagine them doing something as crude as taking averages.

Also, if you think about it.. How would that work. Chicken and egg. How do they first decide who to rank, if to decide who to rank, they need to take the average of the top 10.

It's like looking at the top 10 runners in each university and taking an average of their calories per day, and saying "This is what I need to be an elite runner". It's meaningless.

It would be a weird algorithm if they somehow calculated the ideal anchors for each keyword and ranked based on that. They rank don't rank at all based on anchor ratios.

In the past exact anchors worked because they pushed up your relevancy.

Now, anchors, like kw density isn't about ranking, it's about avoiding a penalty. The key is to stay under the radar, not to push rankings. You push rankings with strong contextual links and you gain relevancy from your page content, on-page factors, total page content, number of articles on the site in the same topic, number in close topics, overall site architecture, internal links, then a bit from relevant links, ie, the article linking being relevant to your page, and lastly, a little relevance from anchors, but if you try to gain more relevance from anchors you just over-optimize.

Ranking for a keyword mostly just comes down to content/topical relevance. It's so so important these days. Google are ultimately trying to move towards a place where they won't even need backlinks. Imagine a human reviewer. They can just pick the absolute best 10 pages without needing to look at backlinks.

Google is great at understanding content now, but still far of that. What this means is they do still rely on backlinks, but if they REALLY like your content/site/topical relevance, they will rank you with less backlink power than others.

Let's look at some examples here for the keyword "best hoverboard"

#1 :-

DR90 - A monster domain. Digital products
106 backlinks to page
Heavy on exact and partials. About 30% exact/partial. Aggressive.

#2 :-

DR48 - Very strong site. Review site.
53 links to page
almost no exacts. I see 1 partial, 1 single word keyword and 1 exact. A ton of url, some misc, some brand, some title

#3 :-

DR16 - Weak site. Very weak. New. Started getting backlinks in march this year. Probably has hidden pbns, but still, it's not a strong site.
32 links to the page. 1 exact, 2-3 partial, 15 naked, a few mixed brand, some misc, couple title.

#4 :-

DR73 - Monster review site
222 links to page. boat load of exacts. repeated exacts. "best hoverboard" + other aggressive anchors.

So, a monster domain, a ton of exacts, a LOT of links to the page, and a targeted article, and it's losing to a new site. So none of that stuff helps. There's a missing piece of the puzzle. It's not over-optimized, because it wouldn't be #4 if it was, it would be at least page 3. The missing piece is to do with the content/topical relevance. At first glance, it does look a bit thin. For whatever reason, despite all the links, google thinks the #3 is a better match for people searching.

No point going on, but if you look at a lot of keywords you'll see a lot of variation like this, so taking mean averages just doesn't make sense imo.

The algorithm is too complex to try and work out the details of a keyword, like why is X above Y, but what we can see is that it's not just simply about anchors, or word count, or page links, or domain strength. But the one thing that does stand out for the most part is better, more indepth content usually always does better, with less effort.

And in answer to your last question about how many keywords to target.. Zero.

Don't target keywords, don't think about keywords.

You have your "main keyword", let's say its "best hoverboards", BUT, in reality, you do not want to try and rank for that or even think about that beyond the title.

The #1 for it is ranking for 5.8k keywords, #2 6.1k keywords, #3 2.3k keywords and #4 1.5k keywords.

According to ahrefs, the #1 has 10k traffic total from those 5.8k with 766 from "best hoverboard". That's about 7.5%. Not much is it? That's why we don't care about individual keywords or stress about whether they go up or down a little.

You can't possibly try to rank for 6k keywords. What you do is just focus on the topic. The intent. Your goal is to help people learn about the different types of hoverboards, what are the best ones, why, what are important considerations. You want to write an article for people looking for for help with buying a new hoverboard. The more you match with that intent, the higher google will rank you, because that's what they are doing. Matching user intent with solutions, not connecting pages to keywords like pre hummingbird.

Click to expand...

davids355 said:

Thanks for the quick answers. Totally makes sense what you are saying. I think average was the wrong word used by me as well; I was thinking more along the lines of TF-IDF. I just wondered if a similar principle could be used by the search engines for anchor text - making assessments based on the entire data-set.

Click to expand...

BR92 said:

Hello congratulations on your initiative to help you better understand how silos work
I have some questions

1 - About internal links, can an article in one silo link to an article in another silo?

for example: mysite.com/seduce/woman have a link to mysite.com/review/ebook-perfect-seduction

2 - buying link packages sold on bhw can damage my site?

Thank you, sorry my english

Click to expand...

splishsplash said:

You're welcome, David.

I don't think they will. It doesn't seem to me to be a good way to assess quality. The only way to do that would be to have manual reviewers pick a seed of 10-20 sites for each keyword and then assess the rest based on those anchors.

It makes more sense that they take a seed of the web's authority sites and see what their anchor profile is to determine a natural anchor profile, then base an algorithmic change(penguin) off that to try and "pick out" spammy/manipulative sites.

Since the goal of penguin, kw density and anchors is not for google to determine who should rank, but to identify unnatural link building. They are using other ways to rank our sites. Content/topical relevance/link juice. It's just that, those things won't count if you trigger one of the algorithmic penalties.

Click to expand...

zimba said:

What is your way to go method for getting seo clients, inbound, cold email or calling, guest posting, referrals, direct mail, ads etc or if you do them all which one have you had most success with?

Click to expand...

SirTop said:

Thanks for the AMA Tom .

I would like to ask if you’ve got experience working with LOCAL SEO clients. Since there are lots of factors also in ranking local client. What are your top checks to successfully rank a local website in first page within 3 months.

Click to expand...

mikey29 said:

Best spot to build links on own site
Home
Silo
Or post you trying to rank

Do you break the silo for highly competitive kw. And make it a page

Great guide btw

Click to expand...

Chiefbobo said:

Thanks for this. I don't think this has or will be asked so:

What are your views on Sape? I have seen real business websites ranking for years with just Sape links. I also used it on sites before and had mixed results.

Click to expand...