Bringing New Visitors to Your Site
Most people take great pride in the Web sites they create and tend to view their site as a groovy place to come and surf around checking out the various new content. Web authors are likely to think of their site as, for instance, a food Web site with reviews, articles, searchable databases of recipes, restaurants, etc. and lots of articles about food. They imagine visitors come to their site, browse around the new lead articles, read the headlines, blurbs and editorials and pretty-much move through their site like a print magazine starting at the main entry page and moving through the site in a logical path. They probably don’t think of their site as a place where a bunch of food information is accumulated ready to be searched and used.
My own research and other studies show most people don’t decide to simply go to a particular Web site and cruise around it looking for things of interest. Although this behavior does happen, most people start their surfing sessions at search engines and directories like Yahoo and are looking for particular bits of information and arrive at sites as a result of a search – trying to find information about a particular wine or food ingredient for instance. Many people are not going to come to your main page and cruise through your site the way you would like them to. They are going to enter your site on some particular page that is not necessarily the entry point you would have chosen. The trick is to figure out what people come to your site looking for, where those entry points are, and provide the type of information they really want. Build more content of a similar nature so people will cruise around other pages on your site and increase your page views (and your bank account).
Here’s how you do it. First of all you have to figure out what is bringing people to your site. To do this you need to use a log file analysis program like WebTrends or IPro that provides detailed information about where your visitors are coming from, which of your pages they tend to come to directly, and what search words they use to get there. These reports can let you actually see the pages on other sites where there are links to your pages. You can then evaluate those links and figure out ways to improve them or get even better ones similar to the existing links. Find the pages visitors seems to like the most and improve them – split them up into separate pages, add new pages on the same subject, create links between these pages and others on your site on similar subjects. You need to create mini Web sites within your larger site focusing on subjects in even greater depth than before. In other words, if you can figure out what people are trying to find and give them more of it on more pages you’ll get more page views.
I publish WebDevelopersJournal.co.uk and a few months ago I was going through a WebTrends report and noticed the top search word that brought people to my site was ‘mmsystem006’. I had no idea what it was but soon found out that mmsystem006 is an obscure multimedia thing deep within Windows that often gets screwed up and needs attention if you want to view cool multimedia files. A lot of people seem to be having problems with it and apparently there is not much information on the Web about how to handle mmsystem006 problems. In one of my very old discussion groups several people had mentioned they were having these mmsystem006 problems. There were no answers suggested. Spiders had indexed these pages and when people were going to the major search engines and searching on mmsystem006 they found links to these old discussion group posts. They probably did a search on mmsystem006 in Excite or wherever, clicked on the results, saw we didn’t have the answer and immediately left my site for greener pastures. Which was OK – I got a few page views from them even if they did leave without having any joy on my site. But how about if I could not only give them the answer to the question so they can go away happy, but how about if I get a few more page views out of these people as long as they’re going to be there anyway?
So I hired a writer to research the mmsystem006 problem and write an article about what the problem is and how to fix it. I cut the article up into 8 separate pages with links between them. I basically built an eight-page mini-Web site that discusses the mmsystem006 problem and has pages for each of the possible solutions and problem scenarios. (You can see it at mmsystem006.com). Using WebTrends and the search tool built into Homesite, I did a search through my site and found all the pages in my old discussion groups and other pages where people had mentioned mmsystem006 and I added a prominent link to those pages worded "mmsystem006 problem solved!" leading to this new mini-site. So now I get multiple page views from all those mmsystem006 answer-seeking visitors instead of just the one page view I used to get from them before creating the new mmsystem006 site. Since we actually have help for them now they may think more favorably of our site and come back for other information.
But that’s not all I did. When I built the mini-site around solutions to the mmsystem006 problem I tried to optimize each page for search engine spiders carefully incorporating the common search terms that brought people to the old discussion group pages in the first place and adding a few carefully-selected ones of my own. To do this I first made a list of the search words and phrases WebTrends told me people were using to find us: mmsystem006 problem, Windows problems, CD-ROM drivers, download drivers, multimedia problems, installing drivers, midi player, mediaplayer, codec, decompressor, and Windows error (notice most of these are phrases – not single keywords). To those words I added words I thought people might use when looking for the answer to this problem.
My goal was to own all the mmsystem006 traffic from the major search engines so I wanted to own the search words as much as possible (go to any of the major search engines right now, type in mmsystem006 and see what happens). Each page of the eight-page site has different solutions to the problem. So I customized each page for search engine spiders knowing they are likely to rank my pages by search word relevance. We want to convince the spiders that these pages are the most relevant pages on the Web for these particular words. Each page on this mini-site therefore has the following unique characteristics:
- For each page the actual file name reflects the page contents like "download_mmsystem006_fix.html", etc.
- Each page title uses the key search words unique to that particular page. Like <title>mmsystem006 problem solved– The Web Developers Journal</title> or <title>missing mmsystem006 error – The Web Developer’s Journal</title>. If a search engine spider is going to look at the page titles to get a hint of what the page is about and list us accordingly, I would rather they list us with the important words for each page rather than the name of my site. People are unlikely to search for "Web Developers Journal" but we know they are searching for "mmsystem006". So put the keywords first.
- The keywords listed in the meta tags contain the words relevant to the particular page as well as the main words people have been using to find us when searching for the mmsystem006 problem answer. This should increase the relevance factor of the page for the particular keywords when the spiders come along. I do not have any keywords in my meta tags that are not used in the visible page text itself.
- The same keywords are used again in a complete sentence as the first clear text on the page. If you cruise through the HTML source for the page, the first actual text that appears is a complete sentence with the keywords in it. When spiders go through the page they will see that the keywords we claim the site is relevant to actually occur first thing on the page. I make sure they are used again further down in the page.
- The keywords most relevant to each particular page appear in the <H1> tag that contains the heading for the page. If the first words on the page include the keywords and the headline of the page contains them, the spiders should be convinced of their high relevancy.
I also have a Web site called WackyHTML.com and I noticed a very large percentage of my traffic was coming from a listing of HTML resources on Yahoo. I looked at the Yahoo page and saw the list was arranged alphabetically and, of course, WackyHTML.com is near the end of the rather long list – well below the fold. Since there is no way I can force the word ‘wacky’ to appear at the top of an alphabetical list, I cleverly started a new, site called AmazingHTML.com. This is not sneaky – this is legit. It's called using your brain.
Use your head when selecting keywords. First of all be sure the keywords you use in your titles, headlines and meta tags are actually used throughout your pages and are relevant to the subject. But stop and think a moment - lots of people tell you to carefully add meta tag keywords. But that’s not necessarily good advice. You should add the things you think people are likely to type in a search engine or directory search window when trying to find mmsystem006 information or whatever your page is about. Studies show people use search ‘phrases’ much more often than they search individual words. If I go to Excite and type in ‘drivers’ I’m going to get thousands of pages suggested for me to look at. This is useless. If I type in ‘Panasonic CD-ROM drivers’ I am much more likely to get a meaningful return. Most people have learned this and use search phrases. So my page on this subject uses that keyword phrase in the meta tags and elsewhere. Many Web developers are not aware of this and still use single search words in their meta tags and elsewhere. So be smart – list search phrases people are likely to actually use.
