Hello,
In this blog post I’ll present my view on how to build a web site with Search Engine Optimization (SEO) in mind.
Note: the main resource for this article was the Search Engine Ranking Factors by SEOmoz. You will see below both the most important factors and a grade indicating the importance of that factor, according to SEOmoz.
First steps:
First of all, you should establish a list of keywords for which you wish to have good positioning; details, in Romanian, on how to do that, here;
A. On-page ranking factors regarding keywords
1. Keyword use in the title tag – 66 points out of 100 possible
The <title> tag should contain your required keyword; very important task;
2. Keyword used as the first words of the title – 63 points out of 100 possible
If possible, use the keywords at beginning of the title;
For example, let’s say my brandname is SEOmoz and I have about Link building;
In this example, the best title would be:
Link building – SEOmoz
or
Link building tips by SEOmoz team
or
Link building resources provided on SEOmoz.com
Either way, “Link building” is on the very first keywords;
Out of the three examples above, the first one:
Link building – SEOmoz
is best, since it’s shorter; it’s best to have a title with as shorter title tag as possible;
Avoid any title with a length larger than 70 characters;
3. Keyword use in the domain name (e.g. www.keyword.com) – 60 points out of 100 possible
This would imply naming your web site:
www.linkbuilding.com
or
www.link-building.com
or
www.link-building-tips-for-success.com
(actually, the latter version is the poorest, since it has way too many keywords in the URL);
4. Keyword use in the h1 tag – 49 points out of 100 possible
You should use your <h1> – <hn> tags to define the content of your page;
Put the title of the current page in <h1>, this is the most important advice; if you have very long pages (let’s say, a page with terms and conditions) and you have more than one title in the page, use <h2> for the subsequent headings; even longer pages? Use <h3>; <h4> and the following barely matter;
5. Keywords in anchor text on the page – 47 points out of 100 possible
This implies not naming your links:
Click here
or
Visit my page
and instead using keywords:
Link building on SEOmoz.org
Also, let’s say you have the same link 4 times on one page; what’s important to know is that only the first link anchor (the text used to describe the link) counts; read this here; (this “here” is a poor example of how you should use it)
More than this, don’t try to put keywords in the menu for “Home”; so, name your homepage in the menu as “Home” or “Homepage” and don’t use keyword stuffing to name your home anchor;
6. Keyword in external links anchor text – 46 points out of 100 possible
This implies a certain factor – you should link to outside web sites using keywords; so, again, if you link to another web site, avoid linking with “click here”, and instead use keywords that are related to your domain;
7. Keyword as the first words in h1 tag – 45 points out of 100 possible
If you use <h1> tag, use the keywords as the first words
8. Keywords in the first 50-100 words in HTML – 45 points out of 100 possible
Use CSS to accomplish this task;
9. Keywords in the page name URL (e.g. example.com/folder/keyword.html) – 38 points out of 100 possible
The URL of the current page should better contain the targeted keywords;
10. Keywords in lower importance headline tags (h2–h6) – 35 points out of 100 possible
As you see, <hn> tags other than <h1> are not as important, but still they do matter;
11. Keywords in ALT text for images – 33 points out of 100 possible
A very important factor; each image that you insert on your web page should have an ALT text;
12. Number of keyword repetitions in HTML Text – 33 points out of 100 possible
Yes, it’s fine to repeat the targetet words a few times on a page;
13. Keywords in image names (e.g. Keyword.jpg) – 33 points out of 100 possible
If you insert images on the page, instead of naming them DSC00002.jpg, use real keywords to describe what they contain;
For separators, it’s best to use “-“, instead of “%40” (space) or “_”;
Below are some factors which barely matter:
a. Keywords in <b> or <strong> tag – this is best done for usability purposes; so, yes, it is good to do a bit of bolding, but only if it helps your users);
b. Keyword density formula (keywords used / total words on the page) – I would take care that my texts are good for the human reader, instead of the bot; the keyword density counts, but it’s really not that important;
c. Keywords in list items <li> – if you use a menu with <li>, sure, that’s a fine solution; other than that, avoid having <li> items just for the sake of it;
d. Keywords in <i> or <em> tags – even less important than bolded text;
e. Keywords in the meta description – you should use a description for each page, but for different purposes than strict SEO; use the description for the two following reasons:
(i.) it’s best to have a description for the users to see what’s your page about; the meta description tag is only used on search engines to show what is your text about;
(ii.) the description is also useful for Facebook – whenever someone links via Facebook to your page, Facebook “grabs” the description as the description of the link; good to use it;
The description should not be longer than 155 characters;
B. On-page ranking factors with other criteria than keywords
1. Existence of a lot of unique content on the page – 65 points out of 100 possible
This is a very important factor; each page should have at least 4-500 words of unique content in it; if you have the same URL for two pages that are very similar (or identical) one to another, you should apply the canonical tag;
Avoid having a lot of pages on your web site that have the same content;
Also important is to avoid copying texts from other web sites; your site should have an unique content;
2. Freshness of page creation – 50 points out of 100 possible
It’s best to have a content that is fresh; so, if you have a blog, a blog post which is written today is more important for the engines than a blog post two years ago; “fresh is good” is a motto Google strives to follow, since a query deserves to receive fresh content (QDF);
3. Use of links point to other pages – 41 points out of 100 possible
This means that if you have a page that is related via its content with other pages, you should link the two; blogs have this with “similar articles” plugins; online stores show you a list of other recommended items you can buy along with the current one; you should have this practice on your blog, via automatically generated, or manually inserted (so, if you write today an article about a subject, and you reference topics in the past, you should link to those messages in the past; it would also be nice to update old posts and link to the new articles);
4. How often the site content has been updated – 39 points out of 100 possible
This implies updating the web site from time to time; instead of having a static web site, which almost never changes, you should have some fresh content from time to time (a blog, a news section, an articles section, etc.);
5. Use of links to other web sites – 37 points out of 100 possible
It’s best to link out to others, instead of keeping your web site as a closed entity;
6. URL with parameters or not – 33 points out of 100 possible
This means having URLs like this:
seomoz.org/this-is-a-great-url/
instead of this:
seomoz.org/article.php?id=30&num=20&start=12&look=search+term&browser=firefox
Below are some factors which barely matter:
a. Ratio of code to text – you should have more content than code, and you can do this with external CSS and JavaScripts;
b. Existence of a meta description – I told you before about this;
c. HTML validation – you don’t need to have an 100% valid code, but do look here and here for very serious problems, which might make your web site unaccessible by web spiders;
d. Flash elements – avoid having flash-only web sites; on the other hand, having a video or a banner in Flash is fine;
C. Site-wide ranking factors
1. Site architecture of site – 52 points out of 100 possible
The architecture of a web site is of small importance if you have a web site with less than 20 pages, all linked from the homepage; if you have a 20-100 pages, this is of a medium importance, while having a large web site (100+ pages) makes this a critical aspect;
The basic thing is to make any of your pages available from 3 clicks from the homepage; also important is to link to the most important of your pages right from the homepage;
Also important: every page should be reachable from at least one static text link;
For other tips and tricks see this and this;
2. External links to top sites – 37 points out of 100 possible
This means – if you link to other web sites, make sure they are trustworthy; don’t link to poor quality web sites (gambling, spamming, adult web sites);
3. How much time since domain registration? – 37 points out of 100 possible
It’s best to have your site set up on an old domain; so, if you want to launch a web site, it’s best to launch it on a domain which you own from a long time ago; not that good of an idea is buying from someone else an old domain with the sole purpose of using its age; this is a useless tactic from Google’s perspective;
4. Domain registration past – 36 points out of 100 possible
You should avoid buying a domain from someone who has bought it from some one else and this prcess repeated itself a lot of times;
5. Site uptime – 32 points out of 100 possible
It’s important to have a web site which is accessible most times;
6. How many sites are hosted on the same server – 31 points out of 100 possible
Avoid hosting your web site on a server which hosts potentially spammy web sites; use this tool to check current web sites;
7. Domain registration change – 31 points out of 100 possible
It’s best to have a domain which hasn’t changed it’s ownership; you own the domain and no one else has owned it;
8. Inclusion of feeds in Google news – 31 points out of 100 possible
If Google News gets news from you, this is a good sign;
9. XML sitemap? – 29 points out of 100 possible
While not a madatory requirement, the use of XML Sitemaps is a great method for “instructing” the crawlers how to navigate your web site;
Below are some factors which barely matter:
1. Domain ownership (who registered the domain) – it makes little difference who, in the past of your domain ownership, held the posession of the domain name;
2. Registration with Google local – while this is a good practice, it makes little difference to rankings;
3. Inclusion of feeds in Google blog search – this is a good practice, yet it matters just a little;
4. Use of feeds – feeds may be a good sign of proper use of technology and may help for crawling, but it does little for rankings;
5. Registration with Google Webmaster Tools – good practice, low results;
6. Use of security certificate – nice to have, not that relevant;
7. Use of a Google Search appliance – I like Google Search appliance, yet it can’t be that much of a factor;
D. Ranking factors which negatively influence your positioning (so don’t do them)
1. Cloaking with negative intentions – 68 points out of 100 possible
You should avoid cloaking (which implies showing the engines a different version of the site than you show to your users) altogether; but the manipulative intent makes it an even poorer practice;
2. Links from the page to spam sites – 51 points out of 100 possible
Make sure you don’t link out to bad neighbourhood;
3. Frequent site downtime – 51 points out of 100 possible
As it sounds, you should make sure the site is up & running most of the time;
4. Hiding text with same color text/background – 49 points out of 100 possible
Very old technique, very easy to spot by the engines;
5. Excessive repetition of the same anchor text – 46 points out of 100 possible
In order to prevent this, you might as well don’t bother about having a high percentage of keyword use on your web site altogether;
6. Hiding text with CSS – 44 points out of 100 possible
Yes, this is old also; very bad technique
7. Keyword stuffing in page content – 41 points out of 100 possible
Again, avoid repeating things (just like I do now);
8. Keyword stuffing in the <title> tag – 39 points out of 100 possible
Don’t put too many keywords in the <title> tag;
9. Keyword stuffing in URL – 37 points out of 100 possible
This tip is especially valuable for the main domain (www.keyword-keyword-keyword-keyword-keyword-keyword.com), but it also applies to pages (www.example.com/keyword-keyword-keyword-keyword-keyword-keyword-keyword.html);
10. Use of “poison” keywords in anchor text of external links – 32 points out of 100 possible
Clear and simple, avoid having “bad” words on your web site; examples:
Viagra, Casino, Gambling, Poker Online, etc.;
11. Excessively long URL – 30 points out of 100 possible
I tend to not apply this rule myself, but the web site should not have too long URLs, be them created with parameters or keywords;
12. Use of keyword-rich texts in footers – 27 points out of 100 possible
For the footer area, avoid having too many keywords with links;
13. Keyword stuffing in the meta description tag – 26 points out of 100 possible
This is actually a commons sense advice, since the keywords in the Meta Description tag don’t help at all, they are only important for usability purposes;
Below are some factors which barely matter:
a. Overuse of nofollow – you should not use nofollow at all, except for blog comments which may have spam or for other external links which you don’t control thoroughly; read more here and here;
b. Excessively long titles – as said previously, you shouldn’t have title tags longer than 70 characters;
c. Keyword stuffing in meta keywords – I would avoid having this meta tag altogether;
E. Factors for geo-targeting:
1. Country code top-level-domain – 69 points out of 100 possible
If you build a web site for Italy, you should have .it domain;
2. Language – 63 points out of 100 possible
Again, if you target Italian lanugage, the web site’s language should be Italian;
3. Geographic location of the host IP – 57 points out of 100 possible
The hosting of your web site should better be in Italy, for the same reason;
4. Geo-targeting in Google Webmaster Tools – 52 points out of 100 possible
Do set the preference in Google Webmasters Tools (GWT), but do note it’s of little importance if your site language doesn’t keep the promise (so if you set Italian in GWT and your site is in Spanish, it makes little difference what you do in GWT);
5. Address in text content – 41 points out of 100 possible
You should put the address on the contact page;
6. Address associated with the domain registration – 35 points out of 100 possible
When registering, you should specify an address in Italy, for example;
F. Other factors not listed above (from here, here and here):
1. You should have a HTML site map, in addition to a XML sitemap; details and an example;
2. Avoid having more than 100 links on one single page;
3. You should check your web site for broken links and avoid having these on your web site;
4. You should allow access to your web site without cookies and logins (whenever possible); also, without Google bots can’t create accounts and login to a web site; you should also allow access without JavaScript, Flash, CSS, forms;
5. Make sure your web server supports the If-Modified-Since HTTP header; this feature allows your web server to tell Google whether your content has changed since we last crawled your site; the feature is useful for bandwidth savings;
6. Test your site to make sure that it appears correctly in different browsers;
7. Monitor your site’s performance and optimize load times; Google’s goal is to provide users with the most relevant results and a great user experience; thus, it will favor web sites which load faster;
8. Don’t create pages with malicious behavior, such as phishing or installing viruses, trojans, or other badware;
9. Avoid having the same content on different URLs; for example:
http://www.site.com
http://site.com
http://www.site.com/index.html
http://site.com/index.html
should redirect the user to one page, and not have 4 different URLs for the homepage; same principle is valid for internal pages;
10. Your titles should be unique – so if you have a page about Italy and another one still about Italy, make sure you cover different topics and don’t put the same title for the two pages; also, your titles should also be descriptive (I should, very easily, know what’s inside a page just by looking at its title), should contain targeted keywords and also the brand name of the company; keep them short (70 character as a maximum);
11. It’s best to register your web site with Google Webmasters Tools;
12. The page file size (the HTML version of a page) should be no more than 150 Kilobytes (use CSS and JavaScript to reduce page size);
13. Avoid having more than two parameters in the URL (&num=10&lang=it);
14. You should avoid having more than 4 levels-deep URLs:
www.site.com/folder1/folder2/folder3/folder4/page.html
15. Avoid using frames to create your web site;
16. You should not use:
“noindex, nofollow, noarchive, noodp, noydir, nosnippet”
without a very good reasoning;
17. Below you will find the standard locations on which the search engines look for XML sitemaps:
http://www.examplewebsite.com/sitemap.xml
http://www.examplewebsite.com/sitemap.xml.gz
http://www.examplewebsite.com/sitemap.gz
18. Avoid having a web site with frames and splash page (intro page into the website);
19. Create a very good 404 page (details);
20. Your web site should only return 404 code in HTTP headers for a non-existent page; other than that, return a typical code, like 200 (for a typical file) and 302 (for a file that hasn’t changed since the last access of the web site);
21. Use robots.txt with care;
Things to ignore:
a. Don’t use meta keywords tag, it’s irrelevant;
b. Don’t create subdomains (e.g. subdomain.site.com) for your web site for Search Engine Optimization purposes; they’re pretty much useless;
Final note: Have a final look on this.
_____________
Bottom line: The tips on constructing a web site SEO-friendly can be further improved with new ideas, but these are the main aspects to know when building a web site.