White SEO

White on Black. Advice on Ethical SEO.

Abandon your hope, nothing to be said about doorway pages here. This article addresses problems which a webmaster may encounter while trying to get a stable source of SE traffic. We will try to look into the essence of various search algorithms, exploring a search engine from the inside and concluding what exactly you need to do to your site to improve your rankings. This article is meant for webmasters who have at least basic SE experience. We will use some terms which are not explained in too much detail.

1.

To understand how you can occupy a decent SERP position and get SE traffic, first you need to understand what a search engine wants to see and what are its aims. Let's try to look through the eyes of an abstract search engine which is meant to produce relevant search results within certain time and resource limitations. How is the relevancy achieved? This is the most basic question which leads to understanding of ranking mechanisms.

So, let's imagine we're building a search engine from scratch. Let the knowledge base of the system be built basing on indexed public sites and other knowledge bases with well-defined links. The interaction with the user is carried out via a "search query". Ideally, this request should generate a relevant reply - certain data which the user needs (like, when the user asks how much is 2 plus 3, the answer is 5). But for most queries exact answers which satisfy all the users are impossible. We can provide more users with source(s) which are likely to contain the information which the users seek. Better relevancy can be achieved through knowing something about the user (personalizing the results basing on language, request history, preferences etc); feedback on system reply quality. Naturally, a Chinese physicist who is searching for "half-life" needs information on nuclear physics, not on computer gaming. Though, if we know nothing about the user, we have to produce results with both fields mentioned, allowing the user to enter a more exact query, or enter some personal information, or select an information source from the list. Thus we come to displaying the search engine reply as a list of resulting pages which contain the necessary information, selected using certain algorithms. In a general case the results should be sorted according to knowledge about users as a whole. This knowledge may be accumulated using adjacent and qualifying queries and other user feedback - something we will talk about later on.

Now let's look at the situation from a different angle. A page's popularity is a valuable parameter which lets us give the user a result with better relevancy. It's completely logical that uninteresting pages (i.e. irrelevant, unneeded) won't be of any interest. So, we are introducing a parameter of page importance, called weight, defined as a probability of a user to find and visit this page. But our knowledge base does not contain information about importance. Instead, we know about links, so, keeping in mind that linking is the web's basic form of navigation; we can evaluate the popularity in a way. Thus Google introduces the PageRank value based on a probability of a user to be directed using certain links inside the page (see http://en.wikipedia.org/wiki/PageRank). Now let's try to discourse on the queries themselves, trying to classify them. We need the classification to get more information and thus produce a more relevant result. To do this, a search engine may store and structure queries, build databases of related words, misprints etc. A search engine may use these databases for ranking. For example, "TV set". This query has "plasma TV", "LCD TV", "video", "screen", "inch", and other relevant words. It is natural that a site relevant to the "TV set" query will likely contain these relevant words, otherwise it would be hard to classify according to the user interest; thus a search engine can logically conclude that such a site is of no value. This is how we arrive at building ontology. Let us now take an outer look on this and try to understand what we can do to make a site relevant. A site will be relevant when it fully satisfies the information needs of the users making corresponding queries. You have to be the best, and you'll be ranked first. This is what an ideal search engine wants. Because real search engines encounter computing difficulties, we will try to help them with this choice further on.

2.

First, we need to have our site indexed, i.e. included into the search engine knowledge base. On the current stage we won't be saying that a key to a successful indexation is the 'addutl' button, or just a bunch of relevant links, which is, of course, important. Let's discuss the types of URLs. Often the same information is available through different URLs, for example, inside a PHP-based site working with sessions sent as GET requests, which makes the site URLs look like: domain.com/info.php?session_id=12346432246, with a unique ID for every user. This is why a search engine bot has problems indexing such pages, because during every visit it makes there are unique links leading to non-unique content. What are the ways out of this from a search engine's position? For example, it may decrease the site's indexing quota for pages which contain no information. A search engine might as well not index the links with "id=" in the URL at all, because in most cases these URLs contain no content. Or a search engine may not visit such URLs unless a page associated with these URLs updates several times which shows the URLs are permanent. The conclusion? It's better not to use "id=". After all, Google itself says it.

Let us have a look at sending parameters through GET requests. A site can have links leading to it looking like info.php?n=333&g=4435690, where "g=" is of no importance. Perhaps it's a link on a message board with some URL parameters removed by the message board engine, which gave us a link looking like info.php?n=333&g=44, which duplicates the content of the former link. If it's a HTML page, the bot would have received a 404 error and leave without cloning existing pages in the knowledge base. Which brings us to the conclusion that the URLs should be static. There is no cheating about this; it's just a way to avoid these problems. Historically many sites have two URLs, the one with 'www' and the one without. It appears so that every page on one of these domains has a clone when seen by a search engine. A search engine may not find out about this, if, for example, there are only 'www.' links to this site, but it is unlikely for a major site. The 301 redirecting is an elegant solution to this problem, sending users and bots from pages of one domain to corresponding pages of the other one. We may finish with URL types here; I would only like to add that many people add keywords to the URL. This is of course an advantage for some of the search engines, though you have to keep this within reasonable limits. An URL looking like domain.com/how-i-can-make-very-big-money-a123.htm or how-i-can-make-very-big-money22.com makes people think this URL is not for convenience but for unethical SEO placement. And we don't want the engines to think that we are bad guys, do we? Domain.com/money/123.htm and money-22.com does look like a usability improvement.

There are always human beings behind a search engines. You've got to understand this, and this is why when you want to promote your site with ethical SEO, you have to win their trust - which is difficult with all the black SEO guys around. We start from choosing a domain name. As it's already been mentioned, long names with keywords are not out thing, we need a short, stylish, easy to remember name. You may register a domain and be ready to learn about sandbox filters (see http://en.wikipedia.org/wiki/Sandbox_Effect), or buy at auctions like afternic.com, sedo.com and others, making sure the domain you buy is not filtered. The algorithms of finding relevant information are not perfect, so there are plenty of ways to spam the search engines and flood the SERP with pages which are far from being interesting to the user. What we have as the result is that a beginner webmaster with a fresh domain has little chance to get noticeable amounts of SE traffic. But if we register or extend our domain name for 10 years, it means we're serious about investing in this site. A spammer won't be interested to do so, so we've gained some trust, having spent some money. When you register a domain, choose the right registrar, that is a registrar that registers lots of no-spam, 'white', legal domains. It's natural that cheap registrars popular with unethical spammers won't decrease the trust to your site, but won't increase it either. When you buy a domain, make sure the domain is SE-visible, has a positive reputation and is relevant to the field of your future site. I recommend buying such domains together with sites and keep as many of the domain details as you can (site content, NS servers, IPs etc). Why? Simply because search engines don't trust domains which change their whois and other info too often, as this is a popular spammer trick. Now, once you've got a domain name, you have to choose an appropriate hosting. If you host your site on the same IP with spammer sites, search engines will draw an easy conclusion that you use shady methods. Try to pick a hosting with legally promoted sites trusted by search engines. If you choose virtual hosting, make sure you have good neighbors - and try to make sure you don't have bad ones - as seen by search engines, of course.

You can use your own NS servers for your domain, or the NS servers provided by your hosting company. I don't recommend using public NS servers - they might be unstable and cause problems with undesired neighbors. Why take risks? Let's address the content of our site. A search engine wants unique content, so let's do our best to place heavy design elements into separate files with Java elements (a problematic solution), or use simple design instead. When the design is heavy, the search engine may see the design as most of your page information, which will make the engine think you are not worthy enough to be re-indexed often. Ideally a search engine has to separate the design from the content, but alas, these algorithms are far from being perfect. Okay, let's imagine we've written some adult paysite reviews, for example, and the reviews have links to these sites. A typical mistake of many webmasters is that they try to conceal links using redirects which send to "/go.cgi?url=porn.com" instead of "porn.com". This script is also mentioned in robots.txt as something search engines should not index. This trick is mainly used for preserving the PageRank of the page, but webmasters forget about a site looking natural. These things might be reasonable, just like Nofollow for message boards, guestbooks and other sites without moderation. Still, a site linking to popular sites or new sites relevant in their content deserves more attention than a faceless pile of content without any links. You have to remember that the more traffic your site gets, the more likely your site is to be reviewed by a guy from a search engine team - so make sure you leave some information about yourself, terms, FAQs, a nice logo and all other attributes of a nice legal site. We have to remember that in many cases a user is guided by a snippet of the search engine, so you have to provide a catchy description of most important pages of your site in the 'description' meta tag (which is especially important for Google). Also, remember to add 'alt' tags to images, saying what the images contain. This may give you some extra traffic, from general and specific search engines as well, like Google Images. If you write articles focusing on a specific topic, try to use set phrases for this field, this will make your content look natural. Make your 'title' unique and rich in information, placing not only the site's name, but valuable information like a relevant keyword etc. Don't use sophisticated JavaScript and Visual Basic scripts without absolute necessity. This may look suspicious, as scripts may conceal redirecting from search engines. If you still decide to use them, make sure they're simple and logical. Never use redirect scripts - use meta tags instead, or 301 and 302 server replies. Always develop a plain link structure. It is completely logical to make most important pages closer to most important pages (like the main page). Use less general navigation texts; try to make your links more site-specific. For example, use your site's logo with keywords in the 'alt' tag, like 'adult white project', instead of the 'Home' link. I'd also like to point out that search engines may use information about visitors which return from your site to the search engine. Frequent returns may show your site contains little valuable information. Time of stay on the site is also important and is in many cases defined by the query itself. For instance, when a user wants a weather forecast for tomorrow, his or her time of stay may be several seconds, but when it comes to adult content, it may be minutes or even hours before the user returns to the search engine to find more.

Now let's address the outer factors. Having a nice and content-packed site made in accordance with these tips, we may be brave enough to say that links are of utmost importance now.

We may assume that sites of relevant topics have more links between each other than completely different sites. So, it brings us to the necessity to place links to relevant sites - and get links from them. The more a search engine trusts a site, the more valuable a link from this site becomes. Of course, important sites may be of no particular topic, e.g. catalogs like dmoz.org. This is why you make sure you exchange as many links as possible with relevant sites, even if they are not ranked too high. So that a search engine trusts your site, learn how the number of links to other sites grows. In most cases a site which links or is linked to every three months looks less worthy than a site with steady growth. Try to place natural links to your site. Don't repeat the mistakes of the first Russian spammers who just placed a bunch of irrelevant links separated with a comma. The link should look natural and may even not contain any keywords, or contain them among other words. Now let's try to imagine what the future has in store for us.

3.

As I already mentioned in the beginning, personalized search listings may help improve search engine efficiency a lot. For example, when we have a complete picture of what a visitor did on this or that site, we can offer completely different, more relevant search results. The search engine will know how users surf the web, so methods of approximate popularity evaluation, like PageRank, cease being necessary. At the same time a more clickable link will become more important. This will make placing non-user oriented links useless and enable the search engine to eliminate possible unethical techniques. Unpopular links may be regarded as SE spam, too. Indexing sites will be also a lot easier. Currently toolbars (like Google toolbar) in theory give the search engines some general information, but these toolbars are not too popular. Search engines may encourage users rate the sites they visit. It is not yet clear what may be the incentive, but the Google toolbar already has the voting buttons; 'for' and 'against' a site, while some search engines incorporate voting into their SERP. We are most likely to see usability professionals in great demand tomorrow, so you better think whether your site is usable right now. There may be less search engines in the future, though the amounts of SE traffic may increase and botnets and proxy server cheats may become more widespread. Using increased amounts of processing power for data analysis will enable the search engines to evaluate documents on the fly. When you have enough quality, understandable content and relevant links, your site will pass the black SEO tests with ease and success.

Let us finish with this optimistic note. Good luck with your ethical SEO, and don't forget that building a 'white', legally promoted site requires a lot of effort which starts paying back a lot later.

AreaFix