Site Architecture: Cross Linking vs. Siloing
-
I'm curious to know what other mozzers think about silo's...
Can we first all agree that a flat site architecture is the best practice? Relevant pages should be grouped together. Shorter, broader and (usually) therefore higher volume keywords should be towards the top of each category. Navigation should flow from general to specific. Agreed?
As Google say's on page 10 of their SEO Starter Guide, "you should think about how visitors will go from a general page (your root page) to a page containing more specific content ." OK, we all agree so far, right? Great!
Enter my question: Bruce Clay (among others) seem to recommend siloing as a best practice. While Richard Baxter (and many others @ SEOmoz), seem to view silos as a problem.
Me? I've practiced (relevant) internal cross linking, and have intentionally avoided siloing in almost all cases.
What about you? Is there a time and place to use silos? If so, when and where? If not, how do we rectify the seemingly huge differences of opinions between expert folks such as Baxter and Clay?
-
I am confused. So lets that I have an ecommerce site that has 20 types(books, toys…) / 20 categories each / 20 subcategories each and thousands of products under each subcategory.
When we say go flat, is it ideal to go all the way like http://www.website.com/type (20 of these), and http://www.website.com/category (400 of these) and http://www.webiste.com/subcategory (8000 of these)and thousands of product pages. So there is no page more than 1 directory level down. Does this mean flat architecture?
-
the breadcrumb is one more signal about where you are in the process, on the site, in the section. Google also likes them and will often show the breadcrumb navigation links right in search results. They try to emulate breadcrumbs sometimes in search results if you don't have them, but if they do, since you're not feeding them an actual breadcrumb, theirs can sometimes guess wrong at the keywords they show in them.
-
Hi Alan...
Is there a case study -- where a silo is broken down and analyzed that I can use to understand this siloing concept !
My understanding of A Silo is -- you for example if you have a grocery store website - you create a dairy section then all related dairy products are found here and a deli dept then all cold cuts in this section etcc where all the pages are themed from the top then on downward, and trying to keep the silo 3 clicks from home
The breadcrumb -- not sure how this comes into play but if I keep the site 3 clicks from home at any time someone needs to get back to where they started from they are able to do it, so how does the breadcrumb help if I am already trying to keep the structure a 3 click structure for easy navigation and easy exit back to beginning.
-
Hey Todd, thanks.
While I definitely agree about having tightly themed categories, I'm not quite sure I am sold on using a silo. Correct me if I'm wrong here please, but isn't a silo when you don't cross link detail pages (within the same category) with each other? I think Alan feels the same way, or perhaps I've misunderstood.
Check this post about the importance of link architecture by Google. Specifically, the last Q&A.
-
I agree with Alan, and would like to add that I believe that using the silo method can increase the proximity of closely connected clusters of keywords better. In other words, by nature, in a silo structure, tightly knit keywords support each other and pass theme and relevance value to each other by default when a strong supportive breadcrumb is in place. Often with a flat site architecture extra programming needs to be done to establish those relationships as they relate to internal pages.
-
Anyone have anything else they'd like to toss into the discussion?
Any examples you'd like to share of detail page linking vs. silos?
[edit] Just found this (old) blog post by Google about the importance of (internal) link architecture... I quote:
Q: Let's say my website is about my favorite hobbies: biking and camping. Should I keep my internal linking architecture "themed" and not cross-link between the two?
A: We haven't found a case where a webmaster would benefit by intentionally "theming" their link architecture for search engines. And, keep-in-mind, if a visitor to one part of your site can't easily reach other parts of your site, that may be a problem for search engines as well.
-
exactly. "Tags" and "materials" are not exactly top level category stuff
-
I found a relatively "ghetto" approach to silo using wordpress, since I don't have the time or technical skill to implement it perfectly. Using a specific plugin, it will compare posts and reference a set number of related at the bottom, creating a link structure similar to a silo. It's not perfect but it is easy.
-
Yeah, your right. I would image those links aren't relevant when on store pages, and would definitely distract some people
On their product pages though, they use some cross linking to relevant topics. But I'm sure it's at the bottom- out of site- as to not distract people. So I would image those are mostly there for SEO. Would you agree?
-
Etsy's got a good structure with their category and sub-category sidebar that balances SEO and user experience. note though that when you get deep into the individual Etsy stores, that's gone, because it would dilute the individual store owner's account focus and distract users.
-
I found a relatively "ghetto" approach to silo using wordpress, since I don't have the time or technical skill to implement it perfectly. Using a specific plugin, it will compare posts and reference a set number of related at the bottom, creating a link structure similar to a silo. It's not perfect but it is easy.
-
I think you're right Alan, that makes great sense. Thanks. Do you think Etsy's sidebar is a good compromise between the two? I'm sure testing each site is the best way to figure out what users prefer on that specific site. But in general, do you think that's a good balance to use in order to keep too many links off the page, yet still keep detail pages within a category linking to each other?
-
Having all listed and linked is ideal for SEO, however you rapidly cross into usability problems if there are more than a handful. (Would you want 50 or hundred links in a sidebar nav? ) When a site is so big that there are more than a handful that could be linked from that sidebar, it's actually best practice to NOT have any others linked from the sidebar, else you confuse users even more (listing only some, but not all). User Experience is paramount when making these decisions. Even at the expense of SEO in some cases. And if that happens, other tactics need to be employed. Like having a separate, dedicated funnel for "featured properties". Which requires even more unique content in that funnel. But it at least boosts the ranking value for those properties included.
-
Agreed.
I spent some time working on a hybrid silo structure in my blog, and proper cross linking on the main area of the site thanks to the discussion here.
-
Sorry for the confusion Alan, and thank you very much for the discussion.
To help clarify for others reading this discussion (and for myself), are we both agreeing that: in the attached image it is an 'SEO AND usability best practice' for the hotel detail pages inside the Tallahassee category/directory to link to each other?
*Of course, there are always caveats, such the maximum outbound link limit recommended by Google, etc.
But as a general practice, would you have "Hotel 1", "Hotel 2" and "Hotel 3" (inside the "Tallahassee" category) link to one another?
-
This is a great question and an even better discussion.
Special thanks to Alan for sharing all of the details.
-
Indeed Alan, that's good advice we all should follow. Thanks. I'll follow suit from here-on
-
You're dead on accurate in the need for and importance of how "consensus" can help new people get started. The trick is helping them find enough truly experienced people who have done that testing on a wide enough variety of sites, as well as lots of disclaimers being plainly stated on all such discussions. It's why I strive to always refer to "in my experience"...
-
If I have a category California Hotels, sub-category San Francisco Hotels, then having links in a sub-navigation bar to each (if there's only a handful), each of those links reinforces the strength of the top level Hotels, 2nd level California, and third level San Francisco related phrases. They all support each other.
If, on the other hand, I have a link to "nearby hotels", that implies I'm going from a single hotel details page to a uniquely filtered "geo" category page that shows hotels based on some criteria - it might be all San Francisco, or all within a distance radius, or all within a zip code radius.
Even if it's all other hotels in San Francisco, it's not a link pointing to another (or several) same-level page(s). It's pointing one layer higher.
That's a filter more than a properly constructed category drill-down. And it implies that the page I'm on will NOT be listed on that target of the "nearby" link.
-
Also agreed. However, when new SEOs enter the sphere, they must start somewhere. And, clearly there's value in studying other's work to help clarify, expand or even challenge one's own hypothesis and practices. I also avoid implementing a tactic/ strategy on a paid client project, if reputable SEO's and/or the community as a whole, recommend against it. I may try it on my personal site, but not a customer's. Thanks for all your help Alan.
-
just to clarify regarding my input - my perspective is based on my experience with client sites on all scales, small, medium, large and mega sites.
To me it's more important to see how things work on our own sites and evolve them over time as compared to purely looking for what others do or say as it's own reason for taking action.
-
Respectfully, what's the difference between the nearby hotels example and the cars example?
More specifically, If these 'nearby hotels' links might dilute that articles topical focus, why wouldn't a link to 'mercedes' from a 'BMW' page?
Thanks Alan. -
I actually don't, because I've always thought it was a bad idea. But it seems other folks don't think it's so bad under the right circumstances.
I'd be interested in seeing a good example of an effective silo as well....
anyone? -
Thanks Dave. This is exactly why I posed this discussion.... it seems as if a lot of us are getting something different from these architecture type posts.
I think it has to do with making same-level detail pages link to each other. Especially if you link to them using the anchor text they're trying to rank for.
For example, what I get out of an article like Richard Baxter's post on SEO Architecture, is that detail pages should link to each other, and that Silos should not be used. And the more architecture posts I read on SEOmoz, SEOgadget and Distilled... the more I think it's a 'best practice'.
That said, it seems from these comments that some folks read those articles differently. I think this is a serious discrepancy that we SEOs should address
-
The slides will be going up at some point in the next few days. And I'll have a follow-up post that includes the notes for each slide. In the mean time, I did an article on Search Marketing Wisdom yesterday directly related to the last slide in that deck.
-
The "nearby hotels to consider" feature is a user thing. It may or may not pass quality page rank.
In some cases, that extra link could dilute the topical focus / strength of the page it's on.
So if I get to resort X's page, and there's a link to "nearby hotels", there's an implied relationship. Good for users. But for SEO, sure it's related stuff, yet maybe not laser focus related.
Another example is blog posts that end with a following box "related articles" and that box contains three or five links to other articles. Maybe they're highly related, maybe loosely. If they're loosely related, sure it MIGHT be good to help users. Yet it probably dilutes this article's topical focus.
-
Agreed, absolutely agreed! Thank you very much Alan!
PS. Could you share the slides from your presentation at SMX Advanced please? If not, how about a link to a post of yours?
-
Well it depends. Is there only one BMW or are there several? If there is only one, then yes - cross link all the luxury detail pages. If there are several, then that's the level for cross linking detail pages, even though it's so deep. If that's the case though, you'd better get inbound links pointing to the parent luxury category page.
And in any regard, don't just have a bunch of links on those category pages - have descriptive paragraph content focused on that category's primary topical focus.
-
Great find on that post. It lays it all out. As long as the silos are thin (not more than 2 layers beneath the home page) it can bring a benefit to adding extra ranking pages with minimal work comparatively. Rand talks about eliminating the bottom layer of the hierarchy to push the content up a level and make the resulting pages extremely stout. The major problem is always going to be the end of the chain. He calls them PageRank sinks.
-
Ultimately, the silo process just takes a bit of time for each new post making sure it links to another category?
I know there are plugins for wordpress that will do automatic linking based on any word you input, and it will link a set or random number of times throughout your site.
It could be worth setting up for me and just include some keyword phrases in the correct articles just to get the link process going properly.
-
Take for example, a resort detail page on oyster.com. They have a section called "nearby hotels to consider", which I believe serves two purposes...
#1) it's likely helpful to users, as most people don't restrict themselves to staying at just one specific resort, and
#2) it helps search engines flow PageRank, crawl and index other pages in the 'Aruba' category.What I can't figure out is, what benefit would it have to not include these links to nearby hotels? (Except perhaps, on checkout process pages of course.)
What if the 'Raddison' Resort for example, got a ton of inbound links and the 'Westin Resort' had only a few? Well, you could cross link them and help the Westin Resort page rank... and simultaneously show your users more relevant options.
-
Yeah, I agree Alan. I don't usually think it's a good return on invested time to practice PageRank sculpting either. One could for example, being building links or generating content with the time/ resources instead
I just re-read what you said, "Individual services details pages should cross-link to each other within that service section at that level though, for usability." To be sure I understand what you're saying.... if your architecture is for example
vehicles -> cars -> luxury -> bmw
vehicles -> cars -> luxury -> mercedes
vehicles -> cars -> luxury -> jaguar
then bmw, mercedes and jaguar would link to each other... correct? -
If anything, sometimes silo structure is not the best for user experience, or the drill down too deep, into ever more thinner content to the point where it's so thin as to have a negative impact on SEO and user experience.
-
I agree with Rand's '09 article in general, however there are some things I think take it a bit too far (such as redirecting PDF documents for link juice). If a PDF is truly the most relevant content on a topic, I believe it should be indexed.
The biggest factor is that if we get completely bogged down in this process just for SEO sake, we lose focus on user experience.
It's right up there with page and link sculpting - to me, it's a waste of time and harms user experience. And the time spent going that far is, in my opinion, in 2011 much better spent on other SEO tactics. Not just because Google has changed how they deal with nofollow links.
-
What are the perceived negative effects, if any, of doing a silo structure?
-
I spent some time using a silo plugin for a wordpress site early on, and also spent some time with a theme that had a silo format, but ended up switching over to a flat site architecture, I just did like the theme for wordpress that used silo, and the plugin seemed like junk.
I'd love to take a look at a well run silo site if you know of one.
-
Not at all, thanks Alan. I think we're in agreeance.
As long as one is not exceeding Google's approx. outbound links per page... and as long as the the navigation make sense to the users.... specific detail pages within the same category should be linked to each other. Is that what you're saying as well?
Here's one example of why I think this is best for indexation reasons. I've attached an image of the page where I circled some stuff
What do you think Alan?
-
Oh - wait I just re-read your question as to not wanting detail pages to link to one another...
If I'm at a sub-category, I would not want, nor need, every individual product/event page in that group to link to each other. Individual services details pages should cross-link to each other within that service section at that level though, for usability.
Does that make sense? Or did I just confuse you?
-
Yes - if there are X number of pages within a section, it becomes too many to reasonably link from a sub-nav. X being a subjective value that needs to be determined case by case.
Ideally, it might lead to yet one more sub-level (such as in sub-sub categories), or in pagination (not blocked from search. That itself is challenging to do in the right manner so as to avoid going too deep or too thin.
There's no other reason I can think of though, and no other method I'd consider a best practice.
-
Thank you dignan99. What's your opinion of silos? Do you like to cross link detail pages within a category to each other, or even category pages to each other?
-
I definitely agree EGOL. We like to meticulously plan out sites and SEO/PPC campaigns prior to launch, but over time a site's architecture definitely needs to be revisited. Usually at that time, we try to also implement any more advanced programming knowledge we might have accumulated to help ease the pain as well
Thanks EGOL!
-
Thank you Dave.
I guess it comes down to flowing PageRank within a category vs. restricting PageRank to the pages that have more links. Any idea why would someone prefer the latter?
-
Thanks Alan. You mentioned, "where all the pages in that section have a link to all the other pages in that section".........
Can you think of any reason why you would not want detail pages within a category to link to one another?
-
I really enjoy topics like this, thanks for asking such a great question.
-
The problem that a lot of people have is that their site grows in unexpected directions. So the problem is not so much deciding upon the structure but more a problem of making the most of the expanding beast!
-
Great question. While everyone has their schools of thought; both methods have their benefits. I tend to favor flat architecture with targeted cross linking. I guess you could call it a hybrid strategy. I begin with a totally flat architecture and silo where it makes sense for the rankings and the navigation for the user. It's all about logical grouping and don't forget the pages must all be link-worthy on their own. If the pages are all strong enough to generate links the problem tends to take care of itself.
-
There's never one perfect solution, however here's the bigger issue. Some people hear "flat" and they take it to the extreme. Which is a terrible concept in 2011.
If you go too flat, you muddy up the proper group relationships. This is where Siloing comes in.
In my presentation at SMX Advanced this week, one of the many methods I recommend for "sustainable SEO" is to group your content, and reinforce that group relationship in URL structure, then with breadcrumbs, and finally with section-level navigation, where all the pages in that section have a link to all the other pages in that section, but where that specific sub-navigation is replaced or disappears as appropriate when you leave that section.
If you've got more than a handful of pages in a section, you should definitely go deeper.
The trick is knowing how wide, how deep to go. It's an art as much as a process studying site data over time.
Another factor is the competitive landscape for a particular niche market. The more competitive, the more important this concept becomes.
Got a burning SEO question?
Subscribe to Moz Pro to gain full access to Q&A, answer questions, and ask your own.
Browse Questions
Explore more categories
-
Moz Tools
Chat with the community about the Moz tools.
-
SEO Tactics
Discuss the SEO process with fellow marketers
-
Community
Discuss industry events, jobs, and news!
-
Digital Marketing
Chat about tactics outside of SEO
-
Research & Trends
Dive into research and trends in the search industry.
-
Support
Connect on product support and feature requests.
Related Questions
-
How Many Links to Disavow at Once When Link Profile is Very Spammy?
We are using link detox (Link Research Tools) to evaluate our domain for bad links. We ran a Domain-wide Link Detox Risk report. The reports showed a "High Domain DETOX RISK" with the following results: -42% (292) of backlinks with a high or above average detox risk
Intermediate & Advanced SEO | | Kingalan1
-8% (52) of backlinks with an average of below above average detox risk
-12% (81) of backlinks with a low or very low detox risk
-38% (264) of backlinks were reported as disavowed. This look like a pretty bad link profile. Additionally, more than 500 of the 689 backlinks are "404 Not Found", "403 Forbidden", "410 Gone", "503 Service Unavailable". Is it safe to disavow these? Could Google be penalizing us for them> I would like to disavow the bad links, however my concern is that there are so few good links that removing bad links will kill link juice and really damage our ranking and traffic. The site still ranks for terms that are not very competitive. We receive about 230 organic visits a week. Assuming we need to disavow about 292 links, would it be safer to disavow 25 per month while we are building new links so we do not radically shift the link profile all at once? Also, many of the bad links are 404 errors or page not found errors. Would it be OK to run a disavow of these all at once? Any risk to that? Would we be better just to build links and leave the bad links ups? Alternatively, would disavowing the bad links potentially help our traffic? It just seems risky because the overwhelming majority of links are bad.0 -
Should I just redirect all my sites to my main site.
Hi, Over the last few years I have built many sites and own a lot of domain names. Some have high page rank some have high domain authority and some have many back links. I'm finding it very difficult to keep up with all the links and being able to provide quality content for everything. Should I just redirect everything to my one site that make the most money as all sites are for the same industry, but in different categories of that industry. So I could 301 redirect all the sites to the relevant page on my money site. Would it be a problem is 1000's if not 10,000's of links all of a sudden pointed in to one site?
Intermediate & Advanced SEO | | cibble030 -
Micro sites?
Hi, I have been speaking to seo firms regarding strategies and they mentioned setting up micro sites under domains that are relevant. i.e setting up armanidoamin.co.uk and we use it as a blog type site to update all info, product reviews, news relating to armani. Whats peoples thoughts on this? Does it work? Is it worth the effort? Im not so sure but obviously looking for ideas. Cheers
Intermediate & Advanced SEO | | YNWA0 -
Can links indexed by google "link:" be bad? or this is like a good example by google
Can links indexed by google "link:" be bad? Or this is like a good example shown by google. We are cleaning our links from Penguin and dont know what to do with these ones. Some of them does not look quality.
Intermediate & Advanced SEO | | bele0 -
New Domain name vs Low Ranked Existing Site
I am going to build a new site. I could hang it off an existing site with similar content or buy a new keyword rich domain and start over. The existing site does not have much trust or authority beyond the domain being registered for 5 plus years. I would prefer to start over and build linksfrom scratch but I realize we are starting at the bottom. The keywords we will be competing against are not super competetive so I think we can get ranking within 6 months or so. These post Panda days I am rethinking everything so any input is appreciated. I did a similar niche site a few years ago and found the site ranked well fairly quickly for its little nice. Today though it may be different. I have no experience in buying domains and would have no idea where to start there. New or existing? Thanks for any input.
Intermediate & Advanced SEO | | Reportcard0 -
2 sites or one sites: 2 locations
Hello, I have a dog training client who is offering services in 2 separate locations. We're looking to be first in the non-local search results and also rank well in google places. Would it be better to go for 2 separate sites or one site and try to rank for 2 different locations with one site? There's both local and standard search results when we type in our keywords. Thanks!
Intermediate & Advanced SEO | | BobGW0 -
I currently have a client that has multiple domains for multiple brands that share the same IP Address. Will link juice be passed along to the different sites when they link to one another or will it simply be considered internal linking?
I have 7 brands that are owned by the same company, each with their own domain. The brands work together to form products that are then sold to the consumer although there is not a e-commerce aspect to any of the sites. I am looking to create a modified link wheel between the sites, but didn't know if my efforts would pay off due to the same IP Address for all the sites. Any insight on this would be greatly appreciated.
Intermediate & Advanced SEO | | HughesDigital0