A few things You need to know About Site Architecture.

Website architecture is one of the most important aspects of creating a search engine friendly website. Below are just a few questions I was asked recently on the topic of navigation, site structure, site maps and pages site.

If I have a relatively small [site] with a flat/linear file structure and each page has links on it to every other page that the spiders can follow does it benefit me at all to have a site map?

If you only have a five to ten page site where every page is of equal value to the rest then a site map is unnecessary. However once you get beyond that, or begin having pages that are sub-pages of a section of your site it's a good idea to create a site map. Even if all the pages are represented in the navigation the site map will help the user know what information your site contains without having to go look for it.

One primary benefit of the site map is that it gives the visitor a birds-eye view of the entire site. In most instances, the navigation itself won't accomplish this, though it can depending on how it's laid out. But even still, if the visitor looking for a site map does so because they perceive an inability to find something that your site should have. It provides them with a complete picture. They know if they cannot find it on the site map then there is no need to keep looking through the rest of the site.

If I have a relatively small site is there any benefit for ranking by having a silo or pyramid structure rather than a flat structure? Do the spiders always prefer silo structures or does it only make a difference if the site if it is over a certain number of pages and if so what is that number?

I think in this case you defer to the visitor over the search engine. Organize your pages in a way that makes sense and will help the visitor easily find the information they are looking for. There is nothing wrong with silo-ing your site structure, even for small sites. Just make sure you're not adding additional layers for the sake of adding layers. Create a structure that makes sense from the visitor perspective. If you do that then it will make sense from the search engine perspective as well.

Do the robots really produce a 404 error if they can't find a robot text file on my site and should I make one just for that reason even if its a small site that otherwise wouldn't need it?

Anytime a search engine comes to your site (at least those who honor the robots.txt protocol) will search for the robots.txt file. Generally, not having a robots.txt file won't hurt you in any way but I do recommend adding one anyway.

Without the file in place the search engines can interpret that any way they want. Mostly it will tell them that your site is either completely open or forbidden. If you have the file in place you are letting the engines know that you have made a definitive declaration of what they can and cannot do in regards to spidering the site.

I have heard that pages should not be more that 100 or 150 k. Is that true? Does it affect ranking besides obviously affecting download times? Is k the same as kb (I know I could probably figure that one out)? How can I tell how big my pages are?

Its always best to keep your code as streamlined as possible. I don't think there is a hard and fast rule anymore as to how big a page can be, however if the engines are finding pages that have significantly more code than content then it very well may have a negative effect. Large pages with lots of content generally won't be a problem, but if the content to code ration is skewed too far toward code, then that represents an issue that's worth fixing. If too many of these pages exist on your site then its highly likely that many of your pages won't get fully spidered and/or many pages of your site could be left out of the index.

Typically when referring to file size someone might say 150k when they mean 150 kb. However if they say it's 150k kb then that suggests its 150 thousand kb. Just about any web file manager such as dreamweaver can tell you how big the page is. If the file is saved on your computer you can also look at the file size through windows explorer. When considering page size, search engines generally are just looking at the code, not necessarily at any additional downloadable items on the page (images, flash files, etc.)

I use iweb to make my site and it is limited in some things that it can do. It uses iframes to inset html content (or you can mess with it yourself). The way I understand it the frames are actually an entirely separate page that is imposed on top of the page you put it on. I have heard frames are bad for search rankings are these frames the same as iframes.

I don't have any experience with iweb so I can't answer any question in regards to that. Traditionally frames are not a good way to go for a site, but there are many ways to have the benefit of frames, without actually having frames. For one, if you need separate scrolling areas, this can be accomplished with CSS. For another, if you want to have one file for navigation instead of putting the same navigation code on every page, the use of server side includes (SSI) is the way to go. Includes allow you to create one file then globally include that file throughout the site. Update that one file and every page that pulls the include file in shows the updated information.

These questions are just a drop in the bucket of what you need to know regarding site architecture and navigation. But I do hope you they provided you with some valuable insight. I have answered a number of other website architecture questions in the past that will give you even more information on this topic. You can also check out my architecture checklist.

A Few things You need to know about Keyword Usage

I recently received an email from someone looking for some advice and a variety of topics. I thought our readers here would benefit from my response. This is the first post of a series of questions and answers touching on keywords, links, architecture and more.

We'll start off with the questions and thoughts regarding keywords and how they should be used on the page vs. how the search engines interpret them. For many outside the SEO industry keyword usage can still be somewhat of a mystery. Hopefully this information will provide some new insights or reconfirm old suspicions.

Keywords are not case sensitive for search.

This isn't entirely true. Search engines try to determine the searchers intent to deliver results best targeted for what the engine believes they are looking for. For the most part, capitalization won't matter, however if the engines can determine that capitalization of a word changes the meaning or intent of the search they may produce different results. Take for example a search for "MASH" and "mash." This is both an acronym and a word and the results are slightly different in Google.

Another example is "fat," "FAT," and "Fat." The differences on the first two searches are noticeably different. The third is only slightly different from the first but still different. Capitalization doesn't effect all results, for example a search for "Business Strategy" vs. "business strategy" produces the same results, at least on the first page of Google.

Keywords are weighted heavier if they are in the title tag or a header tag.

Keywords in your title tag are definitely weighted heavier than keywords in the body copy or anywhere else on your site. The Title Tag of each page is probably the single most valuable piece of real estate that you can use. Its also what the search engines show as the clickable link in the search results.

There is also evidence to suggest that hx tags are weighted a bit heavier than the standard formatted body copy. But this is only the case if it's done within reason. Weighting is all about differences. If all your body copy is an h2 then your keywords won't have any more value than the rest of the content. Use your Hx tags sparingly and they will be weighted accordingly.

Keywords are weighted slightly heavier if they are bolded, underlined, italicized.

I think the optimal word here is "slightly." Bolding, underlining or italicizing words can give the search engine an indication that those words are important to the reader because they are being called out differently than the rest of the text. But if you only bold the keywords you are trying to rank for then I think it's pretty easy for the search engines to discount that because you're using it as a tactic rather than as a way to serve your visitors.

But the issue of weighting also comes into play. If all your text is bolded then there is no additional weighting going on at all. That just becomes the baseline. For something to be weighted more than something else it has to be different from the rest.

What if someone searches for the plural or another form of a word but I only have the original word? For example if I have kayak on my page will come up in searches for kayaks or kayaking? If I have kayaking will that come up for kayak?

The search engines are getting pretty good at associating words together. Technically you don't need to have all the various forms of a word on the page in order for the search engines to rank you for it. An imperfect example of this is to run a search for "kayak" on Google. Scroll down a bit and you see this set of results:

In these three results, somewhere halfway down the page or so, do not have the singular version of the word "kayak" in the visible title or description in the search results, but in all cases the words "kayaking" and "kayaks" are bolded. Granted the word "kayak" does appear in the page text but you can see by the bolding that Google is smart enough to figure out the relationship in the word variations.

In some cased you don't even need the actual word on the page at all for it to appear in the search results. If the search engines determine that the content of your page supports a certain word, even if it doesn't appear, then it can still rank. Though I'd think that pages with the word would tend to outrank those without pretty easily.

Assume someone searches for red trucks. I know that if I have the words red and truck anywhere on my page that my page will come up, but will my page be ranked higher if the two words are next to each other and in the same order on my page as opposed to being separated on the page? For example if my page has "buy red trucks" on it will that be ranked higher than having ...red... in one paragraph and ...trucks... in another? I realize that if they search for "red trucks" in quotes like that then it is better to have the words on your page just as searched for but I am not referring to that situation.

If someone searches using quotes, such as "red trucks," then the engines interpret that to indicate searcher only wants to see results that use that phrase as it was searched. Typically pages that do not use those words together as they appear within the quotes won't appear in the search results at all, unless there is a distinct lack of results to display.

Aside from that, you have a better chance of your page appearing in the results if you use the words as they are searched. The problem is you never really know how anything will be searched. Keyword research will give you the most common searches, but just as many searchers will perform a search differently as those who perform the search as it is most common. Therefore it is impossible to get every combination of the searched phrases on the page.

This then gives you the option of using the common phrase on the page but also to be sure to use the words independently of each other, so much as it makes sense to do so. Don't worry about always saying "red trucks" when you can also say "our trucks come in a variety of colors, including red, which is our best seller."

At what point are keywords considered spamming. For example I have a bullet point type of list on my home page that says that I am the only "blank" in "blank" area that provides service x. The words for the service I provide and the area I am located in end up getting repeated a lot. Will I face a penalty? It reads naturally and makes sense.

I think anytime that you use a keyword or a phrase over an over again you'll suffer some. How do you determine if you are using a word too much? Look at it on the page and tell me if a visitor would find it overly redundant. If it's used properly without any hint of overkill then you shouldn't have any problems. However, if it looks like keyword stuffing to the human eye, the search engines will likely consider it keyword stuffing as well.

How the search engines look at keywords really isn't magic. It's simply comes down to giving the visitor what they want. They don't want keyword stuffing or a bunch of keywords bolded on the page. They want information that relates strongly to the search they entered. If you provide that then you're pretty much set.

Something You Need To Know About Duplicate content And CSS

My last three posts covered a variety of questions regarding keyword usage, links and website architecture.. In this post I'll address the final question that has to do with the visual display of your pages, duplicate content and CSS.

While on the BBC website I noticed that they have an optional low graphics version for all of their pages. I am not sure how they do this but I decided I could do the same by making a low graphics imitation of each of my pages by having a button on each page that could allow people to switch back and forth. I have made all my low graphics pages not only with no or low graphics but with all web safe colors and web safe fonts. The text on the low graphics pages are identical to my regular pages.
Is this considered duplicate content? Will this hurt my search rankings? if it would hurt my ranking could I avoid that by using no follow tags on the links to the low graphics pages or would they still get indexed and subsequently hurt my rankings? Is there something I could inset in a robot text file so that the spider would not go to those pages at all?

They way this is done is by creating multiple Cascading Style Sheets (CSS). BBC.com allows its users to change display settings in a variety of ways, including six pre-set options. Each option, once selected, imports a different style sheet used to display the page. The page URL remains the same, but the way the content appears changes.

By using CSS to change the display BBC.com, or any site really, can have unlimited viewing options without creating any duplicate content issues. Based on the question above, the way the low-res version was implemented will produce multiple pages (URLs) that uses the same content that is also used on the "normal" version. This creates duplicate content that will potentially be a problem for the search engines. And yes, it could effect your rankings.

If you want to implement multiple layouts of your pages, or even a printer friendly version, CSS is the way to go. It's the easiest and cleanest way, and doesn't allow for any potential duplicate content issues. However, if you want to do things the hard way, there are a couple things you can do that will help prevent such duplicate content problems.

The first, as mentioned in the question, is to use the nofollow tag. All links pointing to these alternate versions should be nofollowed. I'd also back that up with a robots.txt exclude file.

Secondly, you could implement the canonical tag. In the <> section of your code place the following:

< rel="canonical" href="http://www.site.com/original-page.html">

This tag would need to be placed on each alternate version of the page with the link going back to the main version. This will tell the search engines that all other versions are not the real one and therefore should not be considered duplicate content.

These are band-aid solutions and I wouldn't recommend them. Creating unique CSS is simpler, cleaner, and ultimately the more perfect route to go.

Why You can't Trust Keyword Research?

I'm a big fan of keyword research but I know that the data provided from any keyword research tool has its limitations. For the most part, these tools can only provide you with information on what is being searched. What they can't do is tell you which searches were relevant, which results were quickly discarded in favor of a different or more refined search, or which searches actually provide the visitor with exactly what they are looking for.

I'm not a big searcher, but I find that when I do I start with the basic concept of what I'm looking for. I'll take a few seconds to scan through the results, and then I'll go back and perform a more specific search. This pattern could happen two, maybe three times, until I feel that I've gotten the information I need.

I rarely expect my first search to produce results that brings me to my final destination. Instead, I'll use the search results to give me clues about how I can search better the next time. With each search, I refine my request, based on what I found in the previous set of results. This pattern continues until I create a search query that gives me a good handful of sites that fit the content I'm looking for.

But it might take me three or four searches to get there.

Recent studies show that searchers are using more words in their search queries (pdf) more often. Still, the lion's share of searches are performed with one, two or three words.

For marketing purposes, the two or three searches I performed prior to performing my final query were of little value. They were valuable to me as a searcher, but only because they helped me refine my search. To the companies listed in the results, I was just another wasted search. But the keyword research tools we use don't know this. They see only queries being performed, not the intent or deliverables of those queries.

Does this mean that keyword research tools are worthless?

No, but you do need to take the data with a grain of salt. Or at least, use that information only until it can be confirmed to provide quality visitors to your site.

There is no perfect way to know if the keywords you uncover with the research tools are going to work for you until you try them.

First, you can run a limited PPC campaign for your words. You're not looking for traffic amounts so much as traffic conversions. You have to run good ads and get at least a few hundred clicks in order to get some accurate results. If your ads are poorly crafted then you might deliver poorly qualified clicks. If your ads are good then you improve your chances of getting conversions, which is what you're looking to do.

With each keyword run this limited campaign and compare the results. Find the keywords that produce the highest conversion rates and use those as your starting point for optimization. But you're not done there.

Trends change over time and keyword usage ebbs and flows. Once you optimize for your keywords, you need to keep monitoring your conversion rates. A high-converting keyword today could be a poor converting word in a month. One that didn't convert so well six months ago may start converting better tomorrow.

Don't assume your keyword research tools are providing you with the best data. They are limited in what they can do. By having your own conversion data, in addition to the keyword research, you are better able to target the keywords that matter most at any given time. Instead of targeting keywords just because they get more traffic, you're targeting keywords that earn you more money.

BY one of the expert in search engine marketing.

Read more...

A few things you need to know about Links

Because of persistent manipulation of on-page "optimization" search engines had to look to other ways of measuring site quality. Enter link algorithms. But links can be manipulated too so it became not just a battle of numbers but a battle of quality. Quality is much more difficult to achieve and requires a lot more work. Some of the best link building strategies you can employ are those involved in building quality content and improving your site for your visitors. These things alone can do wonders in getting people to link to you.

Below are some questions and statements presented to me regarding links. I have provided my thoughts and input that will hopefully give you additional input or confirmation of what you already believed to be correct.

The number of links to a page is important in determining search ranking.

The number of links is a factor, but certainly not the only, or even the most important factor. The PageRank algorithm was essentially built for calculating the number of links to a page and included measurements for valuing links based on the value of the site (as determined by the number of incoming links) doing the linking. But once that started being manipulated other factors had to come into play as well.

It's easy to go out and spam-generate thousands of links, but those are worthless to the visitor and are not a reflection of the quality of the site. So Google and other engines had to start putting quality metrics into the link algorithms to ensure that the quality of the link was more important than just a basic link count.

Links from authoritative sites are the best.

Yes. If you can secure a link from an "authoritative" site, that'll work more in your favor than a link from any other average site. But the location of the link, what page it is on, how it's used are also a factor. If you get a link from an authority site on a page that is deemed to be of little or no value that won't play much in your favor. If you get a site-wide footer link, or advertising link, those can be devalued completely by the search engines. If you get a textual link on a page that gets good traffic and the page itself has a good amount of incoming links, then you have a very valuable link pointing to your site.

Links form sites that are relevant to your site topic are better than irrelevant links.

This is absolutely right. If you can get an on-topic link from an on-topic sight or blog post this will provide much more value than a link that is on a completely irrelevant site. Links from unrelated sites generally won't serve much purpose to the site's visitors nor will they generate much traffic. You're always better off getting links from places where the link will generate traffic.

Anchor text should have your keywords in them.

Yes, but don't go out and get hundreds or thousands of links that all have the same anchor text. That's simply not natural. If you have control over the link text then mix it up a bit. You can say "red trucks," "trucks come in red," "red and yellow trucks," etc. If you were to get 100 links from 100 different sites via natural means, what are the odds that all of those links would read exactly the same?

Links on a page with many outgoing links are rated lower than links on a page with few.

The value of each link on a page diminishes based on the number of other outgoing links on that page. For the sake of simplicity, lets say that a page has a total of 100 points of link value it can pass. If there is one link on that page then all 100 points are transferred to the page its linking to. If two links are on the page then each link passes 50 points. If a page has 100 links, each link passes one point.

Links from older pages are better than links form new pages. (Does older mean the amount of time the page has been hosted or indexed?)

I'm not sure how much the age of the page has to do with the value of the link. It seems that in standard page content brand new links have to age (like wine) in order to reach full maturity in value. However, the opposite seems to be true with blog posts, where links in new blog posts seem to have much more value than older blog posts.

But what also factors in is the value of the page based on historical trends. If the linking page's value increases over time because people keep linking to it then that will increase the value of all links pointing off the page. However if that page remains stagnant with no new incoming links and relatively no traffic then the value of the link juice being passes will probably not improve.

The age of a page would start from the time the search engines found the page, not when it was first hosted.

Reciprocal links are ranked slightly lower than one way links.

Define a reciprocal link. Many sites (usually blogs) link back and forth not out of some quid pro quo but because of the value of the content being posted. Does this make these links reciprocal? Why would a search engine devalue those links just because they have linked to each other naturally like that?

On the other hand if you are building reciprocal link pages then yes, those will be devalued. Mostly because those pages provide little value to the visitors anyway and if the search engines can spot them they will take that into account.

Does it hurt or help my page to have outgoing links? Does it matter what those links are to?

Who you link to has a profound impact on your site. If you link to other sites of authority then you are telling the search engines that you know where the valuable content is related to your industry, and you want to provide your visitors access to that content. You're essentially associating yourself with that site.

That can also work against you. If you link to garbage sites, you're associating yourself with garbage. That will serve to decrease your site's value, especially if you continue to hold and/or build those junk associations.

Is any link better than no link? What about a link from a low ranking page whose topic is not relevant to mine and who liked to my page with bad anchor text? Is it better to not have it at all? Does it depend on anchor text? For example if it was a link from a low ranking irrelevant page but had good anchor text would that make the difference between it being worth having or not?

In some cases I would say that you're better off with no link. As much as you want to be careful about the sites you associate yourself with through linking, you also don't want to be associated with junk sites by them linking to you. In most circumstances, those sites linking to you will not hurt you. But if you are reciprocating then it most certainly will. If you have an opportunity to get a non-relevant link on a non-relevant page in a non-relevant site with very low rankings and bad anchor text, I'd pass. The time it takes to say "please link to me" is costing you ROI. Now if you get that link through no effort of your own, then just don't worry about it.

Is it ever worth having a link with bad anchor text?

Any link will pass value . The anchor text uses is an added bonus that will let the search engines see what people are saying about your site. If you get a thousand links that say "your site sucks" you just gained 1000 links that will in all likelihood work in your favor. At least in terms of the value measured by the search engines. They'll work against you with the visitors.

All links are not created equal. Some can work against you (bad outgoing links) and others will work for you a little or a lot. You can't always control what happens outside of your site or manage who links to you and how, but you can manage your own site. This makes linking out extremely crucial to your link building efforts. Don't engage in pointless reciprocal linking or link to sites that you wouldn't want to visit yourself. If you make it attractive and compelling then you're more than halfway to your goal on creating an effective link building strategy that will pay off.