Estimating traffic generated by Internet marketing campaigns is difficult due to the dynamic nature of search engines, web sites, end-user behavior, etc. However, in this post we provide the foundations so we can develop relatively simple formulations that can approximate this behavior for highly targeted marketing campaigns.

In the following, we discuss how the amount of traffic generated to a website by Internet marketing due to SEO and PPC can be estimated. We develop two models that can estimate traffic due to SEO: 1) a model based on surfing behaviors; and, 2) a traffic model based upon the server’s weblogs. Other types of traffic – Email, SMO, etc., are addressed as well.

A. SEO Traffic Modeling Assumptions

One of the objectives of SEO is to increase volume and quality of (organic) traffic to a web site from the search engines. In general, the earlier a site is presented in the search engine results pages (SERP), the higher it ranks. In fact, one of the goals of SEO is to get organic rankings for the site in the top 10 or so listings of the SERP for identified key business phrases. The reason is that studies have shown that the higher the site is ranked in the SERP, then the likelihood of the user clicking the link associated with that site is higher [3].  Also, studies have shown that when a link is listed after the first 2 pages, it is rarely clicked [3].

Recalling what SEO involves, the following steps (very) briefly describe this activity:

1. Key Phrase Analysis – involves determining what phrases surfers utilize to visit the web site and  convert. Usually, a set of phrases are formulated that best describes the business and the products/services.
2. On-Page SEO – includes amending HTML code on existing web pages (i.e., modifying the title, meta tags, and content of pages to include key phrases); and, if necessary, creating additional pages with content relevant to the search phrases identified in Step 1. These actions increase the likelihood that users will visit the site using identified phrases.
3. Off-Page SEO – includes link building, which consists of submitting the website to major search engines and directories, and trying to obtain back links from “relevant” sites.

To develop a traffic model for SEO, it is important that the following assumptions are made:

Assumption 1. Reasonable estimates of key phrase rankings due to SEO efforts are possible.

Assumption 2. Key phrases identified in Step 2 are representative of organic search terms surfers utilize when visiting the site.

From our experience, Assumption 1 is actually relatively easy to do for an experienced SEO. The reason is that search engine ranking algorithms (i.e., PageRank, Yahoo!, MSN, etc.) tend to have ranking peculiarities and knowledgeable SEO’s understand and know how to take advantage of this.

Assumption 2 turns out to be much more difficult since user searching behavior tends to be very unpredictable. In fact, it is known that on any given day 20 – 25% of the searches in the Google query logs are new. This presents a problem, which we address in a later section where we develop Search Grammars that characterize phrases surfers are likely to search on.

In the next post, we’ll discuss an organic click-through model.

Internet marketing is the approach of applying advertising techniques to generate traffic to websites with the goal of consuming some service, product, information, etc. Generally speaking, common Internet marketing techniques include (but are not limited to):

• Search Engine Optimization (SEO)
• Pay-Per-Click (PPC) advertising
• Display advertising
• E-mail marketing
• Affiliate marketing
• Social media optimization (SMO)

In recent years, the business community has developed techniques that can drive traffic to web servers; however, little work has been done to develop models that can evaluate the performance of web servers resulting from increased traffic generated by Internet marketing campaigns.

From a capacity planning perspective, this is an important consideration given that Internet marketing is expected to grow at rates of at least 20% per year. For instance, some Internet marketing approaches employ PPC, which when implemented can generate traffic to a website in a matter of minutes. Others may take longer like SEO.

In these series of posts, we present traffic models that are able to estimate the traffic due to Internet marketing campaigns:

• A surfer click-through model
• A content building model
• A traditional capacity planning model that uses historical data to predict future traffi

The traffic models can then used as input to an M/M/1 queue where web server performance measures can be ascertained. Our empirical results suggest that Internet marketing can have a significant impact on web server performance – much more than expected. Thus, it’s important for system administrators, web hosting services, etc., to understand the relationship between Internet marketing and web server performance since user-perceived performance is often a major factor determining whether e-commerce transactions are completed.

In the next post, we’ll discuss how SEO traffic models can be developed and utilized to assess the performance of web servers.

After several months in the planning, Microsoft has finally taken the covers off its new web search engine, Bing. Set to go live in under a week, Bing is not just a re branding of Microsoft’s previous search site Live Search but is a major update introducing new features and a new interface, touted as a “decision engine“.

Having been in internal testing for a while under the codename of Kumo, Bing was unveiled today – as expected – at the All Things Digital D7 conference by Microsoft CEO Steve Ballmer.

Bing is expected to start rolling out in the coming days with worldwide availability expected by June 3.
According to Microsoft, Bing will initially have four main search verticals: shopping around for products (with reviews and price comparisons to help you make your decision), travel planning (using technology from their acquisition of Farecast), researching health conditions (possibly linked to Microsoft HealthVault), and local business search (with integrated reviews and maps).

You may be wondering where such a strange name came from. Yusuf Mehdi, the Senior Vice President of the Online Audience Business Group who demoed Bing on stage at D7 today, reveals “We needed a brand that was as fresh and new – A name that was memorable, short, easy to spell, and that would function well as a URL around the world – the name needed to clearly communicate that this is something new, to invite you to come back, to re-introduce you to our new and improved service and encourage you to give it a try.”

Along with the relaunch, the Virtual Earth mapping platform will now be known as Bing Maps for Enterprise, the advanced travel search features made available from Microsoft’s acquisition of Farecast will be fully integrated as Bing Travel, and the Live Search cashback program will also be branded to Bing Cashback.

Will Bing be able to save Microsoft’s ever-dwindling search share? The product will have to seriously impress consumers and become an excellent brand if it is going to reach anywhere near Google’s search market and many believe that Bing could be aimed more at finally removing Yahoo! search from the competition. An advertising campaign for Bing, costing Microsoft up to $100 million, is expected soon as they try – yet again – to make larger inroads into the world of internet search.

Introduction

Recently, we have observed that having as much relevant multimedia as possible on pages on our sites improves their rankings.

As an example, a year or so ago constructing a page with only text content related to, say, some phrase P with sufficient backlinks might be sufficient to rank #1 for optimized keyword searches, however, our analysis shows that this trend is rapidly changing.

Improved rankings for pages makes with a high amount of multimedia content sense intuitively if you understand the Google PageRank algorithm. Google wants to serve pages to users that are the most relevant for a given surfer’s query (which we will denote by Q). Thus, the more content the page has – be it  text, images, video, audio, whatever., then the more (possibly) relevant that page will be for some arbitrary Q.

In the following, we performed statistical analysis of 208 SEO pages that had multimedia content on them (i.e., text, images, and YouTube videos) on them vs. those that only had only text and image content. We assume that the number of  visitors to pages are proportional to the rank of the page for relevant queries.

The following gives summarization statistics of sample data of these pages:

Conclusion

This study suggests that having as much multimedia content on pages on your website increases traffic to your site from the search engines, which implies pages that have more relevant multimedia content, in general, get more traffic and rank higher than pages that do not.

This is the first in a series of posts that will discuss SEO and other Web marketing strategies. Before we dig into our first discussion of SEO, it is worth remembering that search engines are not the only way visitors get to your site. Google analytics for this blog (during the past month) showed that 49.32% of visitors came via searches, 41.24% came from referring sites, and 9.44% already knew the link and came here directly of their own accord. Optimizing your site for search engines is important, but it is only one part of your overall marketing strategy.

What is SEO?

According to Wikipedia:

“Search engine optimization (SEO) is the process of improving the volume and quality of traffic to a web site from search engines via “natural” (”organic” or “algorithmic”) search results. Usually, the earlier a site is presented in the search results, or the higher it “ranks”, the more searchers will visit that site. SEO can also target different kinds of search, including image search, local search, and industry-specific vertical search engines.”

Thus, to optimize our sites we need to know both how people use search engines, and how we can affect our site’s search results. This is important because we don’t seek high rankings for just any search. We seek high rankings on searches relevant to the content on our site.

Search engine usage habits

We all know, from our own experience, that we are more likely to visit the sites we find on the first page of a search result than we are on the 15th page. Sometimes we visit the first site on the list, sometimes the second, third or 10th, depending on what phrases show up with the results. Other times none of the results seems correct, so we use those results to modify the words or phrases used in our search and begin again.

While user habits may vary, all users are searching for something in particular. Some may search on one word, others a combination of words, others combinations of words and phrases, etc. More experienced users will apply boolean logic, while others may just look for simple word combinations. After typing in the terms, the searcher will review the listed sites, skim the phrases shown below the links, then visit the sites that seem most likely to provide the information they seek. If none of the sites deliver, the searcher will have to keep scrolling through the results pages or refine the search.

I don’t want to bore you with the obvious, but for today I want us all to be thinking about the basics. In theory, searchability is simple. If we want searchers to find us, all we need to do is to match our content to their search terms. Of course there is more to it than that, but if we start with a solid foundation we can implement additional strategies later.

Maximizing searchability through content and format

First and foremost, match your content to your goals and the goals of your anticipated readership. If your site is about Pygmy Hippos, then tell us about those hippos. Show us pictures, describe their habits, habitat and physiology. If you think your readers will want more information than you can provide, create a page with links to other sites that you find to be reliable. If your content is good, visitors will read your content, and possibly link to your site in the future. I’ll discuss how links factor into searchability in a later post. For now, let’s look at the structure of your pages. We’ll work our way down from the top, by reviewing some of the coding you should include in your page

Doctype Declaration

The Doctype let’s browsers know how to properly render your page. Below the doctype, in the <html> tag you should also include information such as the language the text is written in. This is important for users who are searching for sites written in a specific language. This information is followed by <head> Examples of this code can be found on the W3C’s Recommended DTDs to use in your Web document page.

Document Title

The document title is what users see at the top of their browser (to show them where they are). It is also the title shown in search results. As such it should be descriptive of your site and reflect its content. Here on the Web Development Blog I use: <title>Web Development Blog: Creative Services: Marketing and Communications: Case Western Reserve University</title> for the title. It’s simple and also shows this blog’s connection to its parent department and the university. I use the same title on my blog pages because the pages are generated by my templates. On a regular Web site you can make the title specific to each page.

Keywords Meta Tag

The keywords tag is not a magic bullet-most search engines don’t even use it. Clients will often request this, so it doesn’t hurt to include it as long as you use it wisely (overuse of keywords in either a metatag or elsewhere, can penalize you). For the search engines that do index this tag, you will have better luck if you keep the list short (less than a dozen keywords/phrases) and only use keywords that are actually found in the text of your page or site. There are those who recommend using currently popular keywords to get more traffic, but this is not a good idea. Even if this worked, why would you want to? Traffic for the sake of traffic doesn’t serve your end goal. If you are recruiting students to your Ph.D. program, including “Harry Potter” as a keyword isn’t going to bring you students preparing for graduate school, it will only bring Harry Potter fans-only a small number of which plan on studying polymers. The few search engines that support this tag give it very little weight, so don’t spend much time on this. Just pick a few words that are directly related to your content, then move on.

Meta Description Tag

This tag is important. The search engines not only index the words used in your description, they also display this text on the search results page. Users rely on this when deciding whether to visit your site or someone else’s. Your description should reflect the nature of your site, and include relevant keywords or company name as appropriate. For this blog I’ve used: <meta name=”description” content=”Case’s Web Development Blog offers tips, tricks and a place to exchange ideas with web maintainers on campus.” /> It’s short and to the point. And while I’m getting pretty good search results (see below), perhaps I could refine it. The words “tips” and “tricks” aren’t used very often, except in the category listing. Now that I’ve had this blog for a few years it might be a good time to review my content and adjust this description accordingly.

Using Keywords in Headlines & Text

It’s well known that search engines and browsers know that words in headlines such as <h1>’s are more important or descriptive than words in the main text, but there seems to be some debate regarding the effectiveness of using keywords in these areas. I would focus more on making these headings meaningful to readers. If the headlines reflect the topic of discussion many of them will naturally be keywords anyway. As with the Case Web templates, put the name of your site in the <h1> tag, a descriptive topical headline in the <h2> or <h3> and descriptive sub headings in your <h4>’s, <h5>’s and <h6>’s. Be specific rather than vague. If you are making a page about your research in elastomers, make your heading “Elastomer Research” rather than just “Research.” Using specific headlines helps readers navigate your content more easily. Choosing key words for your main text should also happen naturally. Users are searching on terms they expect would be relevant to the topic. As such you’ll probably be using these words as a matter of course. If you think visitors will search on term X, but you’ve not used it in your site, don’t try to force it somewhere it doesn’t fit. Instead find a logical context for it. This may entail writing another page or paragraph, but if you think this word is relevant then you should use it in a meaningful way.

In addition to the above, there are other meta-tags you can consider, but overall if you focus on your content you should be in pretty good shape. I’ll continue this discussion of search engine optimization strategies in future posts.

Additional Resources on Search Engine Optimization

• Google’s Blog for Webmasters
• Google Webmaster Help Center
• Search Engine Ranking Factors V2
• Search Engine Roundtable
• Search Engine Watch