Web Site Optimizing
Kevin McElligott has owned a small business since 1993. He graduated from Sac State in 1989 with a marketing degree, and like any other Sac State graduate, he says, he went into a field different than his major.
He started working on computer networks, in 1993 opened his own business, and worked on these systems for several years. He explained that in 1996 the Internet was a hot item, and he started as an Internet service provider hosting Web sites for businesses. It was not a dial-up service; all his clients, he said, "wanted to go live and I wanted to make sure that they weren't paying somebody else to do that."
He said that the Internet was "new and great," and he got into designing Web pages, which turned out to be his true love, and marketing turned out to be a whole new prospect, to take his marketing out to the Internet.
At the SPCUG meeting, he talked about some marketing concepts in ways that will make sense to everyone, even those who don't run a business. He said he wanted to stay away from technical talk as much as possible and spend most of the time explaining how to get Web sites marketed properly out on search engines. He also talked about properly "meta tagging" sites as the means of putting a site on the Web and "talk to a search engine" to tell it exactly where to place it within the indexing.
He indicated that search engines are the resource most often used to find things on the web. Over 90 percent of all searches are conducted using 45 engines, the major ones; the other 10 percent of searches use the minor engines, sites for classified ads. Over half of the 90 percent is done on one site, Yahoo, and if you are not listed on Yahoo, you have lost the battle: if you are not listed on Yahoo, get listed, he stressed.
One of the primary reasons that small businesses fail on line is that they don't have a competitive listing; that is, they are not listed within the first few pages of search results. Most people do not want to go beyond the first couple of pages. The top ten listings, Kevin says, are the competitive ones. The big question is how to get listed among the top ones on the major search engines.
Later, he went on to show how a client was able to maintain a competitive listing on the major engines for about the last eight months. It took a long time, about seven or eight months, to get the client among the top ten and a couple more months to get them in to the top position, and now they "just ride there."
So that everyone would be talking the same language, Kevin took time to define a few terms.
A search engine is a program that searches for documents based on certain key words the user provides. Typing the word "insurance" will produce a list of sites that contain the word. Kevin did that on Alta Vista before coming to the meeting and his search produced 8.5 million possibilities, or hits.
But Kevin would want his insurance business to be first on the list. He said that there are far too many sites for insurance, and for churches, as well.
He explained further, however, that "search engine" is a misnomer, since it is the engine that actively searches, seeking out new documents or new Web sites to add to its index. Originally, it searched, not the user. It has since become a sort of generic term.
Technically, Yahoo is a web directory. You submit your site to Yahoo, and a person comes out to your site, visits it, and has three pages of criteria that you have to meet. If you want to get listed the same day, you pay a fee of $199; but if you are willing to wait a couple of months, you can get listed free and then you can climb up in the rankings.
Kevin also defined meta tags: descriptions of the site that announce and describe its purpose. These tags are hidden inside the pages and talk directly to the search engines to get it properly indexed according to the key words—pet insurance, for example—that will bring up your site.
Kevin showed in some detail the Web site of the City of Grace Church, a client of his, whose site was developed fifteen months ago. What the browser and the search engine see is the HTML code that has some meta tags imbedded along with a title, key words, and a description. The title itself also got a number of hits.
Key words, as we have seen so far, are words that people use to find the kind of site that they want. You build these into your pages as tags. Since we are talking about insurance in our example, insurance is a key word but, as Kevin pointed out, Alta Vista found too many sites, which means that we need to narrow the search: pet insurance, for example.
But since our primary business is insurance, that is our primary category and is listed first. Pet is secondary. A potential customer might type pet insurance and, Kevin pointed out, might get even more possible sites than just plain insurance since any site with the word pet in it will also come up.
Part of the trick is to guess what word or phrase somebody is likely to type in his search and provide that as a tag. The shorter, the better. A single word is best, a phrase no longer than two words is good, but phrases any longer than that are not very useful. In our meta tags, words should be listed separately: insurance, pet, quotation, healthcare, animal, dog, cat, bird, hamster, guinea pig, and then any phrases: pet care, pet insurance, animal care, animal insurance. Kevin said we shouldn't exceed twenty words as key words; and no word should be listed more than three times. We would be penalized if we do.
Kevin said he lists everything, including proper names, in lower case and has had good results. But it seems to make little difference whether they are in upper or lower case.
The form of the word you list is also important. For example, listing "pools" works better since if someone types "pool," the index will recognize it, but if the tag on your page says "pool," and someone types "pools," the search will bring up nothing, because that would be considered a different word. Other types of plurals, however, present other difficulties: the plural of "factory" is "factories," which computer programs see as different words, since "factory" cannot be extracted from "factories": the "y" makes all the difference.
Titles are, of course, very important. Every page on your site should contain a title that is unique to that page, or each page should contain a copy of the same title; and every page should have the same title that has been optimized for the search engines.
Titles should contain the name of the company or product first. Then, you should write the description that you want the search engine to show along with your title.
Kevin recommended that all these be in as good grammatical form as possible, but should not exceed 250 characters, or you will be penalized. When the search engine pulls up your site, it lists the title and the brief description so that the user knows what you are about.
Sometimes, it is useful to list misspellings as well as all possible permutations among the keys that your word might take. For his example of the church, he lists the title of the Web site, www.cityofgrace.org, as one word.
Web sites can also be listed by category. On Yahoo's home page, the searcher can find things by category: Our site for pets and their insurance is listed under the general heading of insurance, not pet insurance and not animal pet stores because our site contains the proper meta tags including a tag for the category.
We need to be careful that our site is not called up in lots of different categories since that confuses the search engines, which will drop the site. The proper tags will, as Kevin said, "announce your site to the search engine spiders." The spiders, he explained, are the engines that are crawling the Web day and night searching for new sites. We need to be sure that all our pages contain the key word as the spider covers our entire site. When a person visits the site, he sees the whole thing, all the content, which is what the search engine brings up. Done properly, each page would have a menu across the top that would lead the visitor to other places on the site, including home.
Search engines love text on your home page, and Kevin suggested having at least 450 words on the home page. The key words should be repeated, but sparingly, near the top where the search engines can find them easily. Avoid too many graphics at the top of the page.
Some sites have text within a graphic images, which doesn't qualify as text in the traditional sense because search engines cannot read the text in an image. Text should be within the page itself.
Kevin warned many times about repeating key words more than three times. Search engines are on the lookout for spam and spammers. If a word appears too many times, they become suspicious and drop the site to near the bottom of any list.
One technique the engines look for is a word that is the same color as the background it appears on. The visitor can't see it, but the search engine picks it up, and if the word is repeated many times, the site is dropped. Kevin mentioned one such site. It was listed as Internal Revenue Service, IRS, over 1,000 times on backgrounds of the same color so that the viewer couldn't see them, but the search engine could. It occurred in 1966 in the run up to April 15 and was visited more than any other site at the time. The search engine decided that it must be the IRS, but, trouble was, it turned out to be a porn site. Rule: Don't spam. Kevin recommended submitting your site daily until it gets listed. You can submit one page each day to the major engines, but if you submit two on the same day, they don't like that and you will get penalized. If your site is built around a standard menu system, you need to submit only the home page. Once a true search engine hits your home page, it will spider your whole site. If you introduce new pages or new products all the time, then these new pages need to be submitted separately. Kevin said that if you do an intelligent job of presenting your site, posting your key words, and gain a high rank on the major search engines, then it costs almost nothing to maintain it. You need not pay somebody to market once you reach this status.
Kevin pointed out that a popular site will work its way to the top of a search engines list. The more people choose it, the higher on the list it gets to be. If you are really energetic, you can repeatedly access your own site and bring up its apparent popularity, but you need to do it very carefully. If do it without logging off each time, the engine's site picks up your IP address, and instead of making your site seem popular, your site will drop out of sight. This approach would be tedious and consume vast amounts of time.
You also need to track activity on the site and adjust things to make it better. There are tracking programs like WebStats to monitor hits and which search engine they come from, what key words were used to find you, and how deeply visitors traveled through the site. Also, you can go to any search engine and type "site tracking," which will bring a thousand options for tracking software. In relation to this subject, he talked about a program called PassingLane that helps to get high rankings on a search engine's pages. Toward the end of his presentation, Kevin pointed out that he and his organization can do for his clients all the things he had discussed.
Pretty Good Privacy
Ken Hopkins, replacing a speaker who cancelled, chose PGP as his topic of the evening for a number of reasons, he said. Since the SPCUG press release said we were going to discuss e-mail, this is an e-mail topic. Also, PGP has been in the news recently as the government is "stirring up the idea that we need to limit encryption for use by government agencies only, or they want a back door into the encryption" used by other people.
Encryption has been a controversial topic. Some people say that encryption should be illegal since bad people could use it to plan their illegal activities in secret, and some believe that encrypted messages were used in the planning of the events of 11 September. Ken said that he had seen no evidence of these activities.
Some people argue that because encryption is available, the terrorists must have used it to prevent our finding out what they were up to. And since we did not know in advance of their plans, they obviously encrypted their messages.
Ken was not convinced. He intended to show that encrypting messages does not necessarily mean that secrets are being kept.
PGP was developed by Phil Zimmerman, and stands for Pretty Good Privacy. He didn;t know much about encryption when he started out. When he did, a controversial issue was whether users should be able to encrypt at all. Phil freely gave away the program and its source code. At present, you can download it from Massachusettes Institute of Technology in the United States, and there are a dozen or so foreign sites that are not restricted.
Ken said that it is a nice program and has had a pretty good following for a number of years. His presentation was to demonstrate its various parts, and having written some encryption code himself, he has a good idea of what is involved.
He explained its workings in the least possible technical terms and said that we didn't necessarily need to understand the mathematical aspects of the process, even though encryption is basically a mathematical process that Ken says is weird. "It is just weird. You're shifting this and you're shifting that and you're re-masking it and doing all kinds of weird things."
The result is that you write messages that the recipient can read, but anyone intercepting it cannot because they don't have the key. Only the sender and the receiver have the key. The encryption algorithms produce a key that is long enough and large enough that no one is ever likely to break it. There just isn't enough time in the universe, Ken said; the codes are just too big.
Breaking a code made up of one bit is easy: it takes two tries; two bits takes four tries; the progression is exponential. Some of the simple protocols have a 128 bits, a really big number. Some keys have been broken, but they were short and some really fast computers were available; the government DES standard, a 56-bit key, was broken, but it took 28 days to do it with a specially-built computer that cost a million dollars to build. Adding one more bit would increase the time necessary to break it to about forty days with a certain amount of luck needed as well.
The current version of PGP is designed to drop into various mail readers. Ken tried working with Tom Anderson and his mail reader, Calypso, but they were not entirely successful. Working with Larry Clark, Ken had no trouble.
In the demonstration, Ken succeeded in getting PGP to insert itself. With PGP installed, there is a little icon in the system tray. Click it and additional buttons came up: encrypt, sign, launch. He then sent a message, a test, to another of his e-mail addresses.
In setting up the message, he had a couple of choices. He chose to do a signed message that uses one of the algorithms referred to as a hash, a one-way encryption. You can create a hash, but you can't get the original data back; it goes in only one direction; you can't undo it. Signing the message attaches your ID, looks at the entire message, and makes a hash of it, a compression of everything. The result is some string of numbers of a selected length.
He sent the message to himself, received it and then used the button that says decrypt. In this case, however, it wasn't encrypted, just signed, but Ken said it amounted to the same thing and is signed with Kenneth J. Hopkins. If someone else tried to sign it with a different name, their address would show up; and if it was someone not in your list, then it wouldn't recognize the source and would not verify who it is. Ken then encrypted a test message, succeeded in sending it, and then decrypting it.
For encryption to work, the recipient needs to know how to decrypt the message. Somewhere along the line keys have to be exchanged with anybody that you wish communicate with. Ken thinks that very soon everyone will want those who send e-mail to have these signatures to verify who the senders are. The problem, then, is how to exchange keys.
Ken said there are whole industries being created that are trying to develop trusted keys. You would then subscribe to a service that has trusted keys. If you were to get a letter from Senator Foghorn, you could go to the service to verify that the message was indeed from the good Senator. Ken thinks that this is the most likely kind of system that people will want, and soon.
In trying to send an encrypted message, PGP might say that it can't. It has to find a possible key at the other end.
Two kinds of keys exist: public and private. You have a public key that senders have access to for sending you messages; you have your own private key that allows you to decrypt them. In order to send someone an encrypted file, you must have their public key; when they receive the message they have a private key that allows them to decrypt the file. Even you now can't decrypt the file that you encrypted, only their private key can decrypt. The whole process, at a certain level, is automatic.
Public keys can be listed in some database on www.pgp.com, for instance, on your Web page, as part of your signature at the end of your e-mails, or generally published on the Web. If you receive encrypted messages this way, then you can verify that the sender is indeed who he claims to be.
It is possible to encrypt anything, and to make everything hidden. Ken said that when he has gone to shows on encryption, the people there think that we should encrypt everything. That way nobody would know what to look for as they try to chase down the bad guys.
In the early days of encrypting e-mail, theoretically, everything was diverted to the NSA in case they thought they needed to look at what was so secret that you had to encrypt it. If everything is encrypted, there is no way they can figure out what might be important. Encrypting only what is important would give the secret away.
