Web traffic: Encyclopedia - Web traffic

Web traffic

Web traffic is the amount of data sent and received by visitors to a web site. It is a large portion of Internet traffic. This is determined by the number of visitors and the number of pages they visit. Sites monitor the incoming and outgoing traffic to see which parts or pages of their site are popular and if there are any apparent trends, such as one specific page being viewed mostly by people in a particular country. There are many ways to monitor this traffic and the gathered data is used to help structure sites, highlight security problems or indicate a potential lack of bandwidth – not all web traffic is welcome.

Some companies offer advertising schemes that, in return for increased web traffic (visitors), pay for screen space on the site. Sites also often aim to increase their web traffic through inclusion on search engines.

Web traffic - Measuring web traffic

Web traffic is measured to see the popularity of web sites and individual pages or sections within a site.

Web traffic can be analysed by viewing the traffic statistics found in the web server log file, an automatically-generated list of all the pages served. A hit is generated when any file is served. The page itself is considered a file, but images are also files, thus a page with 5 images could generate 6 hits (the 5 images and the page itself). A page view is generated when a visitor requests any page within the web site – a visitor will always generate at least one page view (the main page) but could generate many more.

Tracking applications external to the web site can record traffic by inserting a small piece of HTML code in every page of the web site.

Web traffic is also sometimes measured by packet sniffing and thus gaining random samples of traffic data from which to extrapolate information about web traffic as a whole across total Internet usage.

The following types of information are often collated when monitoring web traffic:

• The number of visitors
• The average number of page views per visitor – a high number would indicate that the average visitors go deep inside the site, possibly because they like it or find it useful. Conversely, it could indicate an inability to find desired information easily.
• Average visit duration – the total length of a users visit
• Average page duration – how long a page is viewed for
• Domain classes – the top level domain of the ISP a visitor uses, useful for finding out geographical statistics
• Busy times – the most popular viewing time of the site would show when would be the best time to do promotional campaigns and when would be the most ideal to perform maintenance
• Most requested pages – the most popular pages
• Most requested entry pages – the entry page is the first page viewed by a visitor and shows which are the pages most attracting visitors
• Most requested exit pages – the most requested exit pages could help find bad pages, broken links or the exit pages may have a popular external link
• Top paths – a path is the sequence of pages viewed by visitors from entry to exit, with the top paths identifying the way most customers go through the site
• Referrers; The host can track the (apparent) source of the links and determine which sites are generating the most traffic for a particular page.

Web sites like Alexa Internet [1] produce traffic rankings and statistics based on those people who access the sites while using the Alexa toolbar. The difficulty with this is that it's not looking at the complete traffic picture for a site. Large sites usually hire the services of companies like Nielsen Netratings [2], but their reports are available only by subscription.

Web traffic - Controlling web traffic

The amount of traffic seen by a web site is a measure of its popularity. By analysing the statistics of visitors it is possible to see shortcomings of the site and look to improve those areas. It is also possible to increase (or, in some cases decrease) the popularity of a site and the number of people that visit it.

Web traffic - Limiting access

It is sometimes important to protect some parts of a site by password, allowing only authorised people to visit particular sections or pages.

Some site administrators have chosen to block their page to specific traffic, such as by geographic location. The re-election campaign site for U.S. President George W. Bush (GeorgeWBush.com) was blocked to all internet users outside of the U.S. on 25 October 2004 after a reported attack on the site [3].

It is also possible to limit access to a web server both based on the number of connections and by the bandwidth expended by each connection. On Apache HTTP servers, this is accomplished by the limitipconn module and others.

Web traffic - Increasing web traffic

Web traffic can be increased by placement of a site in search engines and purchase of advertising, including bulk e-mail, pop-up ads, and in-page advertisements. Web traffic can also be increased by purchasing non-internet based advertising.

If a web page is not listed in the first pages of any search, the odds of someone finding it diminishes greatly (especially if there is other competition on the first page). Very few people go past the first page, and the percentage that go to subsequent pages is substantially lower. Consequently, getting proper placement on search engines is as important as the web site itself.

Web traffic that comes from unpaid listing at search engines or directories is commonly known as "Organic" traffic. Organic Traffic can be generated/increased by including the web site in Directories (p.e. Yahoo, DMOZ), Search Engines (p.e. Google, Inktomi), Guides (p.e. Yellow Pages, Restaurant Guides) and Award Sites.

In most cases the best way to increase web traffic is to register it with the major search engines. Just registering does not guarantee traffic, as search engines work by "crawling" registered web sites. These crawling programs (crawlers) are also known as "spiders" or "robots". Crawlers start at the registered home page, and usually follow the hyperlinks it finds, to get to pages inside the web site (internal links). Crawlers start gathering information about those pages and storing it and indexing it in the search engine database. In every case, they index the page URL and the page title. In most cases they also index the Web page header (meta tag) and a certain amount of the text of the page. Then, when a search engine user looks for a particular word or phrase, the search engine looks into the database and produces the results, usually sorted by relevance according to the search engine algorithms.

Usually, the top organic result get most of the clicks from internet users. According to some studies the top result gets between 5% and 10% of the clicks. Each subsequent result gets between 30% and 60% of the clicks of the previous one. So it is definitely important to appear in the top results. There are some companies that specialize in search engine marketing. However, it is becoming common for webmasters to get approached by "boiler-room" companies with no real knowledge of how to get results. As opposed to Pay per Clicks, search engine marketing is usually paid monthly or yearly, and most search engine companies cannot promise specific results for what is paid to them.

Because of the huge amount of information available on the internet, crawlers might take days, weeks or months to complete review and index all the pages they find. Google, for example, as of the end of 2004 had indexed over 8 billion pages. Even having hundreds or thousands of servers working on the spidering of pages, a complete reindex takes its time. That is why some pages recently updated in certain web sites are not immediately found when doing searches on search engines.

In return for a small payment many larger companies choose to advertise their sites on other popular sites. This e-marketing usually takes the form of:

• Banner advertising: Banner impressions are sold by the thousands, and referred to as Cost Per Impression (CPM). As of 2004, prices range from $1/CPM for a run-of-network to about $50/CPM or more for specialized targeted runs. Most popular web sites sell banner advertising space, with the notable exception of Google.
• Pay per clicks: Advertisers "buy" keywords or keyphrases by bidding on them against other advertisers. The so called Pay-per-click engines sell their premium spaces showing in the searches the highest paying advertisers. Google sells paid advertisement through its AdWords and AdSense systems, which place sponsored links on search pages. Overture, now owned by Yahoo!, is one of the most popular pay-per-click advertising venues.

As users got used to seeing banners, some companies chose to make the advertisements more intrusive – pop-up ads became particularly popular to attract attention. However, most people consider pop-ups a nuisance and several software companies offer free pop-up blockers. Even Microsoft included a pop-up blocker in Service Pack 2 of Windows XP.

Web traffic - Traffic overload

Too much web traffic can dramatically slow down or even prevent all access to a web site. This is caused by more file requests going to the server than it can handle and may be an intentional attack on the site or simply caused by over-popularity. Large scale web sites with numerous servers can often cope with the traffic required and it is more likely that smaller services are affected by traffic overload.

Web traffic - Denial of service attacks

Denial-of-service attacks (DoS attacks) have forced web sites to close after a malicious attack, flooding the site with more requests than it could cope with. Viruses have also been used to co-ordinate large scale distributed denial-of-service attacks.

Web traffic - Sudden popularity

A sudden burst of publicity may accidentally cause a web traffic overload. A news item in the media, a quickly-propogating email, or a link from a popular site may cause such a boost in visitors (sometimes called a flash crowd) that overwhelms the site.

Web sites have been forced to close after an unexpected mass increase of traffic, particularly those run by an individual leasing the bandwidth from an ISP or hosting site. Some sites backed by large companies running their own servers have also been caught out by the problems of overpopularity. When first announced, the Vision of Britain Through Time site, containing information taken from the 1901 UK census, was advertised on numerous television programmes and causing such interest that the site had to be taken offline until different arrangements were made to cope with the traffic. The site was hosted by a project at the University of Edinburgh and they had not foreseen the amount of bandwidth and the server load that would be required. Ironically, by the time the site was able to cope with the traffic both the interest and the free advertisements of the site had greatly slowed, giving them excess capacity.

There are some particular web sites that are so popular that any links to external sites can cause problems for the destination host. These include:

• Boing Boing — being "BoingBoinged"
• Fark.com — being "Farked"
• Heinz Heise — the "Heise effect"
• Instapundit — an "instavalanche"
• Kuro5hin — being "Kuroded / Corroded". Doesn't happen often.
• Memepool
• Metafilter
• Slashdot — the "Slashdot effect"
• Something Awful
• Penny Arcade — being "Wanged"
• Sensible Erection — getting "SE'd" or "Sensibly Shafted"
• Digg — being "Digged" or "Dugg"

Web traffic - Top web sites

As of September 2005, the top English language web sites in terms of traffic ranking as listed by Alexa [4] were:

1. Yahoo
2. MSN
3. Google
4. Passport.net
5. eBay
6. Microsoft
7. Amazon
8. MySpace
9. Google UK
10. AOL
11. BBC Online
12. CNN
13. Go
14. Fastclick
15. Blogger
16. Alibaba
17. Xanga
18. Casale Media
19. eBay UK
20. craigslist

See also

• Search engine optimization
• Traffic exchange

Adapted from the Wikipedia article "Web traffic", under the G.N U Free Docmentation License. Please also see http://en.wikipedia.org/wiki