TUTOR BUSINESS MANAGEMENT A+ WORK, ON TIME, NO PLAGARIZING; ON TIME

Search Engine Optimization: Comparison of Link Building and Social Sharing Sonya Zhanga and Neal Cabageb

aCalifornia State Polytechnic University, Pomona, CA, USA; bProduct Managers Association of Los Angeles, Los Angeles, CA, USA

ABSTRACT The practice of Search Engine Optimization (SEO) can significantly increase a website’s search rankings, driving more traffic to the website, and thereby increasing revenue. Three primary SEO strategies are Content, Link Building, and Social Sharing. Our study compares the performance of three similar content-rich websites; each uses a different SEO strategy – one focused on Content only, the second uses Content and proactive Link Building, and the third uses Content and Social Media campaign designed to elicit social mentions and social sharing. We use Google Analytics to collect, analyze, and compare their site traffic, domain authority, search ranks, and ad revenue before and after SEO implementation, over the course of 18 months. Our study suggests that link building and social media both improve website traffic and revenue. Social media can be an effective method for quickly building traffic (for events); however, link building provides better return on investment in the long run.

KEYWORDS Google Analytics; link building; search engine optimization; social sharing; web analytics

Introduction

From fashion blogs to online technology magazines, today’s Web has never been so crowded with content-rich websites. As the cost and complexity of launching a website dropped significantly, many individuals and communities have created their own websites/blogs and published a great amount of content, hoping to make a profit from online advertisements and affiliate marketing. However, many of these websites only focus on content and design, while missing effective promo- tion of their content, thereby failing to monetize and sustain in the long run. Websites are often discovered while prospec- tive visitors search for what they need on popular search engines like Google, Yahoo, and Bing. As of September 2014, Google takes 67% of U.S. search market, followed by Microsoft Bing (19%), and Yahoo (10%) [4]. In this study we discuss how Search Engine Optimization (SEO) can help these content-rich websites to improve their search ranking and traffic, thereby increasing revenue. In particular, we compare two primary SEO strategies: Link Building and Social Sharing, in a multi-website experiment.

Organic SEO

There are two types of search results: organic (by search engine’s algorithm) and paid (by advertisement). The method of optimiz- ing a website to make it rank well in the organic search results is known as Search Engine Optimization (SEO). As the Web becomes increasingly crowded, new websites naturally rank lower on the search results, making it more and more difficult for prospective clients and customers to discover them; whereas established websites’ relevancy can also be affected overnight by

the constantly evolving search engine algorithms; therefore, SEO is critical for the success of all online business.

In the early days, SEO was a shortcut to gain more website traffic. One could easily rank well following gray or black-hat SEO practices such as (1) stolen, duplicate, spun, thin or no content; (2) keyword stuffing, cloaking, or spamming; (3) link farm, link spam, purchased links, or link manipulation; and (4) light affiliate, top heavy ad sites. Today, these tactics havemuch less of an impact on search rankings; in fact, as many have learned the hard way, sites using them can be penalized or even banned/blocked fromGoogle search results [14]. While fundamental SEO tactics like on-page tag optimization and site structure continue to help search engines to discover and understand the content of a site, this is no longer sufficient to rise to the top of search results in a crowded search landscape. Now search engines are also seeing user experience, responsive design, link profile and social presence as additional quality signals to determine the search rankings of the website.

Academic researchers [6, 12, 13, 15], industry experts [3, 5], and agencies [11, 14] have identified three primary SEO strategies: (1) Content, (2) Link building, and (3) Social sharing. Figure 1 illustrates how these SEO strategies can be implemented to a website, and how performance can be measured. In the following text, we explain each strategy and provide examples, which then leads to our research question of this study, which focuses on comparing link building with social-sharing strategy. In the “Results and analysis” section, we explain each metric, how we collected and analyzed the data in an experiment.

Content

Google’s Search Engine Optimization Starter Guide [11] pro- vides some best practices that both Google’s internal teams

CONTACT Sonya Zhang [email protected] Computer Information Systems Department, College of Business Administration, Cal Poly Pomona, 3801 West Temple Avenue, Pomona, CA 91768, USA. Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/ucis.

JOURNAL OF COMPUTER INFORMATION SYSTEMS 2017, VOL. 57, NO. 2, 148–159 http://dx.doi.org/10.1080/08874417.2016.1183447

© 2017 International Association for Computer Information Systems

and external webmasters follow to improve their sites’ crawl- ability and indexing. The very beginning of Google’s SEO Starter Guide states, “Interesting sites will increase their recognition on their own. Creating compelling and useful content will likely influence your website more than any of the other factors discussed here.” In other words, creating high-quality content and awareness can inspire viral sharing and natural linking. Unfortunately in reality, very little con- tent ever naturally goes viral or garners popularity in today’s highly competitive Internet market. Websites that rank well in search results depend upon not only “compelling and useful” content but also the right platform and delivery, and often a little push in the right direction, helping the website to build some momentum.

Both Google’s SEO Starter Guide [11] and Moz’s Beginner’s Guide to SEO [14] covered a dozen common areas for webmasters to optimize the platform and delivery of their content, including:

(1) Site structure: Using well-structured pages including sections and categories, self-documenting semantic URLs, well-selected internal linking and keywords, and effective navigation can get a user to any page of the site with fewer clicks, i.e., high click-through rates (CTRs). Using standard HTML and XML site- maps can further ensure that search engines quickly and accurately account for all of the website’s content.

(2) On-page SEO: Finding more precise and effective key- words/search phrases through keyword research, and using these keywords in title tag, headlines, alt attributes of images, and through-out text can further signal rele- vance and significance of content topics to the search engines thereby help reinforce search rankings.

Search engines have recently started providing better tools to help webmasters improve their search results. For example, Google Webmaster Tools (1) allow webmasters to set geo- graphic target, preferred domain, URL parameters, and crawl

rate so that Google can crawl the site more efficiently and rank the site better; (2) provide diagnostic tools that inform webmasters of malware or errors found while crawling their site; (3) provide HTML suggestions such as meta description issues, title tag issues, and non-indexable content issues; (4) provide statistics with unique insight to SEOs in particular, including keyword impressions, CTRs, top pages delivered in search results, and linking statistics; (5) allow webmasters to submit sitemaps, test robots.txt files, adjust sitelinks, and submit change of address requests when the website is moved from one domain to another. These on-page and site structure tactics play a critical role in boosting the discover- ability and accessibility of content, for human users and search engines alike. Quality and useful content delivered through well-designed site structure and on-page SEO is the most fundamental and agreed upon SEO practice for websites of all kinds [3, 5, 11, 14].

Link building

Google’s SEO Starter Guide [11] acclaims, “Link building is an art. It’s almost always the most challenging part of an SEO’s job, but also the one most critical to success.” What differentiated Google in the early years from other search engines was its focus on algorithmic use and analysis of inbound links (i.e., external citations of a website or page, also called “backlinks”) to deter- mine the relevancy of a website or page. Through links, Google can analyze not only the popularity of a website or page based on the number and popularity of pages linking to them, but also metrics like trust, spam, and authority. Similar to determining the prominence of an academic journal by the number of cita- tions they get in other quality journal articles, Google looks for links from other quality sites (i.e., sites with high authorization and/or popularity) to determine the significance of a website, thus to influence ranking propensity in their search results. Nowadays all search engines are doing this kind of relevancy search. Therefore, websites with a variety of high-quality inbound links can improve their relevancy to search terms/

Figure 1. SEO for website.

JOURNAL OF COMPUTER INFORMATION SYSTEMS 149

keywords, rank higher in the organic search results, and increase traffic through organic search.

Link building often works the best with a holistic strategy that incorporates all levels of link building, from low-value discussion type links to high-value directory type links, as building a vast and varied link profile brings the best Search Engine Results Page (SERP) ranks [11, 14]. An SERP is a page displayed by a search engine in response to a keyword search query, which usually contains a listing of the search results; each result includes a short description and a hyperlink to the original website or page.

There are many ways to build effective links [3, 5, 11, 14]:

(1) Manual “Outreach” Link Building. Outreach links can be manually built by e-mailing bloggers for links, submitting sites to directories (e.g., DMOZ, Yahoo Directory) and authoritative sites (e.g., .edu and .gov domains), or paying for listings of any kind. However, paid links are not encouraged as search engines like Google now discount the influ- ence of paid links in their organic search results; in fact websites caught buying links or participating in link schemes risk severe penalties/bans [14].

(2) Self-Created, Non-Editorial Link Building. This refers to creating links to the primary website through (external) blog comments, user profiles, guest book signings, or forum signatures on other popular blogs and websites. These links offer the lowest value, but can still have an impact for some sites in aggregate. In general though, search engines continue to devalue most of these types of links, and penalize sites that pursue these links aggressively [14].

(3) Editorial Link Building. This refers to creating rele- vant and valuable content on other (external) popular blogs or websites with a link back to the primary (internal) website. There are a couple of ways to build effective editorial links:

a. Guest posts: Volunteer to publish articles and blog posts in popular and authoritative online magazines and writer’s networks, such as Inc.com, BusinessInsider, Examiner, EzineArticles, and gather.com, and provide links in the content back to the primary website.

b. Link Wheels: Set up a series of micro blogs (such as in WordPress, Blogger) with content that flow from one website to another and eventually link them back to the primary website [3].

Finally, link research should be performed periodically to ensure the quality and effectiveness of the links built, includ- ing the number of new incoming links, overall inbound/back- link portfolio, quality and authority of the sites linked to the primary website (i.e., avoid spammy links), ranking and traffic change. Popular link analytic tools are Majestic SEO, Raven SEO Tools, Screaming Frog, SEOmoz, and Wordtracker. For example, Moz’s domain authority is often used in link research to predict a domain’s ranking potential in the search

engines based on an algorithmic combination of all link metrics.

Social sharing

Compared to traditional link building discussed in the last section, social-sharing strategy focuses on the art of creating interesting and viral content that attracts the attention of potential customers and leveraging social media (human approach) to build natural linking rather than manipulating search engines (machine approach).

Search engines have recently started integrating public social mentions and social sharing (e.g., Facebook Like or Share, Google Plus +1 or Share, Twitter Favorite or Re-tweet, LinkedIn News Feed Like or Share, Pinterest Like or Pin-it) in their SERP ranking algorithms [3, 5]. In addition, both Google and Bing have indi- cated that the relative authority of a user factors into how heavily these social signals are weighted. In acknowledgment of the sig- nificance of social sharing, Google has launched their own social network Google Plus, and used the activity on this network to influence Google search results. When a user is logged in to Google and performs a search, it will prioritize links that have been “liked” or “shared” by people within the user’s Google Plus network. Similarly, Microsoft enhances its search algorithms using Facebook’s social graph data.

On the other hand, social networks influence Google both directly and indirectly. Social networking/media sites like Facebook, Twitter, Google Plus, Pinterest, YouTube, Flickr, and LinkedIn are crawled like any others and profiles containing links on these sites are a direct benefit to the recipient of those links. Indirectly, the promotion of a brand or a story through Twitter or Facebook also results in awareness and additional links by people who see these details. This, in turn, benefits the search rankings of the recipient. These effects are further weighted by the visibility or influence of people involved; some have much more social reach and effect than others [3]. Social sharing also goes well beyond SEO—significant direct traffic, branding value, and exposure to the target audience.

There are many ways to elicit viral social sharing:

(1) Link Bait. This involves creating share worthy con- tent (e.g., articles, video, Infographics, apps) on the primary (internal) websites, and then sharing the content via links, images, or quote on (external) social platforms such as Twitter and Facebook, giving credits to the content creator and the site where it originated. To create a viral effect, it is critical to create content that leverages aspects of usefulness, information dissemination, or humor. So that users/ audience who see it will want to share it with friends, websites, and social platforms. This high quality, edi- torially earned votes are invaluable to building trust, authority, and rankings potential [14].

(2) Social Presence. Most businesses have a Facebook page, where they post consistent and fresh news feed about the company or product. Some went further with Twitter, Instagram, Pinterest and Tumblr pages, and even apps. The critically acclaimed TV show Breaking Bad created an outstanding Facebook app that

150 S. ZHANG AND N. CABAGE

transformed a person’s name with element icons like the iconic logo, and allowed people to set it as their Facebook profile photo. The app was a huge success garnering over 4.1 million impressions and was used by over 800,000 people.

(3) Incentivized Social Sharing. Many consumer-based busi- nesses promote their products and websites with incen- tivized social-sharing campaigns. To celebrate its 30th anniversary, Kenneth Cole created a dynamic social hub, which aggregated social content fromTwitter, Instagram, and Vine using the #30yrsbold hashtag. To encourage customer participation and sharing, they offered a hash- tag sweepstakes where fans could win a $500 gift card as well as a customized prize like a street-style photo shoot or a night out in NYC with a special gallery tour. The campaign revived its website and created social impact across multiple platforms.

Finally, social analysis should be performed regularly to dis- cover what consumers are saying about the brand, the website content, and the competitors. Popular social analytic tools such as SalesForce’s Social Studio [16] help businesses to understand and use the most interesting topics and influential conversations to inform future marketing decisions.

Relevant SEO studies

While searching the keyword “SEO” and “Search Engine Optimization” returns thousands of book results on Amazon and tens of millions websites on Google, which shows vast interest in the subject, only a few SEO studies were published in recent academic journals and conference proceedings, and most were limited to: (1) the algorithm criteria, effectiveness and robustness of search engines [7, 9]; (2) the comparison between SEO and paid search marketing/advertising [2, 6, 10]; (3) the overview [12, 15]; or (4) the economic model [17, 18] of SEO with no or little empirical evidence.

Malaga [13] found that SEO is an effective method for improv- ing search engine rankings and site traffic. In addition, the costs and benefits of the SEO project proved more cost-effective than the pay-per-click (PPC) search marketing campaign. However Malaga’s study used only one eCommerce/retail website – SEO and PPCwere applied in phases to the samewebsite and data were collected for 6 months after. Without comparing to any control group it cannot be asserted that the ranking and traffic increase were due to the SEO efforts. In addition, the website in the study sold iPod accessories, which was a very competitive market, hence may have skewed the conclusions away from PPC advertising. Ghose and Yang [7] suggested that while the conversion rates, order values, and profits from paid search advertisements were much higher than those from natural search, organic search seemed to perform better at the specific keyword-level. However, Ghose and Yang’s data are also limited – from a single eCommerce/retail website. Evans [6] recognizes the importance of inbound links as well as directories such as DMOZ and social bookmarking sites to rank highly in Google. However, Evans’ study analyzed only the top 50 websites returned for one very specific and uncommon query/keyword. Agarwal et al. [1], Malaga [12], and O’Neill and Curran [15] provided overviews of

SEO strategies and pitfalls, including black hat, gray hat, andwhite hat on page (page title, meta tags, and URLs) and off page SEO (linking, content, and keyword research). Xing and Lin [19] proposed an analytical model for the sustainability of SEO firms and the impact of SEO on search engine/online advertising indus- try profit. However all these studies [1, 12, 15, 18] were not empirical, did not measure or compare the effectiveness of each specific technique/strategy, and did not include the most recently emerged social-sharing strategy. The limited number of academic studies in SEOmay result from researchers lacking ability to build or access to the backend data of multiple, comparable, and live websites that can reach substantial site traffic and decent ranking, as well as other resources (time, money, expertise, etc.) needed for experiments.

Research question and methodology

As discussed earlier, content is the most fundamental and agreed upon SEO practice; however the effectiveness of Link Building and Social Sharing is not as obvious. Our study intends to answer an interesting and valuable research ques- tion that has not been resolved or examined by any previous SEO study: which is a more effective SEO strategy for con- tent-rich websites—Link building or Social sharing? Assuming a content-rich website owner spends the same amount of time and resources in SEO that had similar content, site structure, on-page SEO, organic traffic, author- ity, search ranking, and ad revenue at the beginning, during the same period of time (in order to eliminate the factor that search engines alter their algorithms), we proposed following explorative hypotheses:

H1: Both link building and social sharing are effective SEO strategies. In other words, the websites that use link building or social sharing will realize greater increases in traffic and revenue, and higher domain authority and SERP rank, than those that did not use any.

H2: The social-sharing effects are realized more quickly than the link-building effects. In other words, the websites that use social sharing will increase traffic and revenue more immedi- ately than those using link building.

H3: The link-building effects are longer lasting than the social-sharing effects. In other words, the websites that use link building will realize higher traffic and revenue than those using social sharing, for a sustained period of time (the 18 months of this study).

H4: Link building is more effective than social sharing. In other words, websites that use link building realize greater increases in traffic and revenue, and higher domain authority and SERP rank, than those using social sharing.

In order to test the hypotheses, we selected a group of three content-rich websites for the study: Site A, Site B, and Site C (alias of the websites are used throughout this article to protect commercial confidentiality):

JOURNAL OF COMPUTER INFORMATION SYSTEMS 151

● Each website is standalone and has a unique domain name and URL—an improvement reducing noises com- pared to other previous SEO studies that used subdo- mains or subdirectories of the same website, such as Malaga’s [13] study in 2007.

● All three websites have similar quantity of content including text, multimedia, and ad units (1000+ pages each website, 500+ words each page).

● All three websites use similar quality and useful content to drive traffic thereby generating ad revenue—Each site pre- sents information about a city on hundreds of web pages, and in six categories: Neighborhoods, Entertainment Events, Professionals (doctors, lawyers, accountants, etc.), Restaurants, Hotels, and Factorials.

● The three cities featured are medium U.S. cities similar in geography size, population, culture, and tourist attractiveness.

● All three websites have similar target audience groups— people who plan to travel or relocate or have recently relocated to that city, as well as current residents.

● All three websites are built at the same time (September– October 2010), used the same SEO-friendly WordPress fra- mework for site structure and on-page SEO, and had similar traffic, domain authority, and the number of links at the beginning of our study (details in “Results and analysis” section). In particular, all three websites were optimized following best on-page SEO practices suggested by Google [11] and Moz [14], considering audience, competitors, key- words selection and display, and content writers (matched

with Killoran’s [8] study in 2013 on key SEO lessons for professional communicators).

● All three websites make revenue through Google AdSense and affiliates (local.com and hotels.com), and had simi- lar ad placement, traffic, domain authority, the number of links, and ad revenue at the beginning of our study.

Figures 2–4 show the screenshots of the websites’ home pages (screenshots partly blurred to protect commercial confidenti- ality). During October–November 2011, link building was implemented to Site A and social sharing to Site B, while Site C was used as a control group (i.e., content only, no link, or social SEO implemented). Table 1 shows SEO strate- gies, actions, and expense incurred for the three websites.

Results and analysis

Four Key Performance Indicators (KPIs) suggested by Google [11] and Moz [14] were used to measure site performance before and after SEO implementation: traffic, domain author- ity, SERP rank, and revenue (indirect). Each KPI and their results are discussed in the following sections.

Site traffic

Google Analytics (GA) was used to collect and analyze the site traffic data of all three websites before and after SEO for 18 months (August 2011–January 2013). No paid advertising was

Figure 2. Site A (link building) home page.

152 S. ZHANG AND N. CABAGE

used to any of the websites, so the site traffic was either (1) Search Engine Organic (visitors referred by an unpaid search engine listing, e.g., Google.com), (2) Referral (visitors referred by other websites, e.g., Facebook.com), or (3) Direct—visitors who visited the site by typing the URL directly into their browser. “Direct” can also refer to the visitors who clicked on the links from their bookmarks/favorites, untagged links within e-mails, or links from documents that do not include tracking variables (such as PDFs or Word documents). SEO focuses on obtaining a natural placement on organic SERPs and bringing in organic traffic.

Table 2 shows a breakdown of total site traffic for all three websites over 18 months. Note that GA uses “session” to measure traffic. A session is a period of time (default 30 min) a user is actively engaged with the website, app, etc. All usage data (Screen Views, Events, Ecommerce, etc.) are associated with a session. All three websites began with a similar amount of traffic. Site A gained the most total traffic (10,553) as well as most search engine organic traffic among all three (8,523); Site B was the second in total traffic and gained the most significant referral traffic (3,262). The traffic results were consistent with the SEO implementation – Site A implemented Link Building and gained most of its traffic (80% of total traffic) from search engines/organic traffic, and its top sources/medium is Google; Site B implemented Social Sharing and gained a significant part of traffic (35% of total traffic) from social referrals, and its top sources/medium is StumbleUpon.com—one of the top social media traffic gen- erators in the United States, attracting over 25 million regis- tered users.

Figure 5 compares the monthly search engine organic traffic of the three websites over 18 months. All three websites started at about the same level, arrived at its high- est organic traffic volume 8 months after SEO implementa- tion before taking a steep decline, most likely due to the Google Digital Millennium Copyright Act (DMCA) updated on August 10, 2012, which affected most content- rich websites online, and then followed by a gradual recov- ery for Site A (link building) and Site B (social sharing); however Site C (control) never recovered from that point. For the data sets in Figure 5, t-test results also show that the three websites are significantly different in organic traffic.

Figure 6 shows that Site A (link building) gained the most organic traffic over time, Site B (social sharing) gained less than half of Site A’s organic traffic, and Site C (control) was the least.

Figure 7 compares the monthly total traffic of the three websites over 18 months: all three websites started at about the same traffic (ANOVA tests find no significant difference); Sites A and B had grown significantly over time while Site C remained at the same level. Noticeably Site B had a significant spike during November–December 2011, right after social sharing was implemented to the website. This traffic increase was mostly created by social website referral traffic (not search engines/organic traffic) during a short period of time (November 16–20, 2011). GA Traffic source/medium and traffic referrals show detailed activities of these five days (Table 3).

Figure 3. Site B (social sharing) home page.

JOURNAL OF COMPUTER INFORMATION SYSTEMS 153

A total of 1917 visits during these five days were from StumbleUpon.com. GA Site Content shows that the most viewed page during these five days was a blog article pub- lished on November 16, 2011, about a unique event in that city (2,025 total page views, 65% of the total page views during these five days). The article went viral mostly due to that an active member of the StumbleUpon community added this blog page to StumbleUpon’s index of high-quality, human- curated content, which was viewed and shared by thousands

of StumbleUpon’s frequent users. The same blog article also received the most Facebook likes and Twitter tweets during these days among all blog articles on Site B. Traffic reached highest on the day the article was published (960 on November 16), calmed down in the following two days, and then went high again on Saturday (916 on November 19). Traffic came down quickly after the weekend (266 visits on Monday, November 21), and remained steady 10% growth for the rest of the year. The blog articles were part of the planned

Figure 4. Site C (Control) home page.

Table 1. SEO implemented (Timeline: October–November 2011).

Site SEO

strategy SEO actions Type Expense

Site A Link Building

Link requests to other bloggers Manual outreach $2,500 paid for research, writers, SEO agency, and paid directoriesDirectories submission Manual outreach

Comments, reviews with signature and primary site link on other blogs, forums, and websites

Self-created, Non editorial

Guest posts on other blogs and websites with author bio and primary site link

Editorial

Link wheels Editorial Link Research, Keyword research Audit

Site B Social Sharing

Articles on social news websites (Digg, StumbleUpon, Reddit etc.)

Link bait $2,500 paid for research, writers, SEO agency, and contest prizes

Infographics on social media Link bait Account, content, and engagement on social networks (Facebook, Twitter, Tumblr etc.)

Social presence

Incentivized social sharing contest Incentivized social sharing

Social analysis Audit Site C -Control None None None $0

154 S. ZHANG AND N. CABAGE

social-sharing strategy and at least one became viral, created the momentum of the referral traffic spike, and gradually helped building up links and organic traffic thereafter.

Figure 8 compares the total site traffic of the three websites over the course of 18 months: Site A (link build- ing) generated 46% of the overall traffic, Site B (social sharing) generated 41%, and Site C (control) generated 13%. There were 9,238 total unique visits to Site B from August 1, 2011, to January 30, 2013, among which 3,565

(38%) were via social referral, including 2,290 (64% of total social referral) from StumbleUpon, 365 (10% of total social referral) from Facebook, and 36 (1% of total social referral) from Twitter. In comparison, there were 10,553 total unique visits to Site A during the same period of time, among which only 83 visits (0.7%) were via social referral, and 8,472 visits (80%) were search engine (Google, Bing, Yahoo, etc.) organic. These numbers confirmed the traffic effects of link-building and social-sharing SEO.

Table 2. Site traffic breakdown analysis: August 2011–January 2013.

Website Type of traffic

# of sessions

% of total traffic

Top source/medium

# of sessions of top source

Site A Organic 8,523 80.8% Google/organic 7,565 (88.8% of organic traffic)

Referral 562 5.3% Google/referral 242 Content + Link Building

Direct 1,468 13.9% Total Traffic

10,553 100.0%

Site B Organic 3,758 40.7% Google/organic 3,455 Referral 3,262 35.3% StumbleUpon.

com/referral 2,290 (70.2% of referral traffic)

Content + Social Sharing Direct 2,218 24.0%

Total Traffic

9,238 100.0%

Site C Organic 2,121 73.7% Google/organic 1,720 (81.1%) Referral 199 6.9% Google/referral 101

Content Only (Control)

Direct 559 19.4% Total Traffic

2,879 100.0%

Figure 5. Search engine organic traffic: August 2011–January 2013.

Figure 6. Total search engine organic site traffic.

Figure 7. Total site traffic: August 2011–January 2013.

Table 3. Site B traffic breakdown analysis: November 16–20, 2011.

Date Social (# of sessions) Direct Organic Total daily site traffic

Nov 16, 2011

712 (74% of daily site traffic, including 705 from StumbleUpon)

241 7 960

Wednesday Nov 17, 2011

150 (57% of daily site traffic, including 145 from StumbleUpon)

110 4 264

Thursday Nov 18, 2011

166 (58% of daily site traffic, including 163 from StumbleUpon)

119 0 285

Friday Nov 19, 2011

673 (73% of daily site traffic, including 665 from StumbleUpon)

238 5 916

Saturday Nov 20, 2011

243 (66% of daily site traffic, including 239 from StumbleUpon)

124 0 367

Sunday Total 5 days 1944 (70% of total daily site

traffic, 1917 from StumbleUpon) 832 16 2792

JOURNAL OF COMPUTER INFORMATION SYSTEMS 155

Domain authority and link research

The change of domain authority and links was monitored for all three sites using Open Site Explorer (OSE), a link research tool provided by Moz, a well-known American Internet marketing software company. OSE is powered by Mozscape, which is Moz’s index of the links on the Internet. Beyond the standard link data, such as linking URLs, linking domains, and anchor text, Mozscape includes unique metrics that provide insight into the authority and trust of pages and domains. Major search engines (Google, Bing, and Yahoo) built similar indexes to help evaluate the importance and relevance of content on the web. OSE provides the following metrics:

(1) Domain metrics ● Domain authority—predicts this domain’s ranking

potential in the search engines based on an algorith- mic combination of all link metrics. Root domains (e.g., TripAdviser.com), without the sub-domain or sub-directory, are used.

(2) Page metrics ● Page authority—predicts this page’s ranking poten-

tial in the search engines based on an algorithmic combination of all link metrics.

● Linking root domains—Number of unique root domains containing at least one linking page to this URL.

● Total links—All links to this page including inter- nal, external, followed, and no-followed.

OSE metrics scores relevantly reflect the authority and trust of the website, and mostly correlate with the website’s traffic and SERP rank. Since OSE updated its data every 4 weeks, the domain authorities of the websites were collected monthly over the course of 18 months. Figure 9 shows that all three websites started at about the same domain authority (12/100), and all of them had significantly and gradually improved their domain authority since October–November 2011, during which the link building and social sharing were implemented to the websites. All three website had a big upward spike during April/May 2012, which is likely due to the Google Penguin update on April 24, 2012. Penguin update favored websites that follow Google’s publishers guidelines, while punished those spammed its search results using paid link

or link network. Similarly, Google DMCA update (August 10, 2012) and other minor Panda and Penguin updates timeline coordinated with domain authority ranking increases (small spikes). Among the three websites, Site A gained the steepest increase (158%) in domain authority (31/100), Site C gained the most modest increase (75%, 21/100), and Site B sat in the middle (136%, 26/100). Even though no link or social SEO was implemented to Site C, it is common for quality new websites to gain domain authority over time. For the data sets in Figure 9, t-test results also show that the three websites are significantly different in domain authority. The domain authority trends shown in Table 4 and Figure 9 are overall consistent with site traffic trends in Figures 7 and 8.

Table 5 shows the trends of inbound links, including linking root domains and total links, of all three websites before and after the SEO implementation. Site A (link building) achieved the highest number of both linking root domains and total links— 26 unique websites provided link to Site A, which generated a total of 598 (average 23 links per site) links to Site A by the end of the experiment in January 2013. In comparison, Site B and Site C also grew their links over time. Site C (social sharing) obtained 9 more linking root domains, most of which were social sites including Digg, StumbleUpon, Reddit, Facebook, and Twitter, and addi- tional 199 links, most of which were from the blog article shared on StumbleUpon, as discussed in Site Traffic.

SERP ranking

Google AdWords Keyword Planner was used to perform keyword research that finds the primary keywords that havemore potential to attract traffic to each of the three websites. Google AdWords Keyword Planner is one of the most popular tools in the industry to get keyword ideas and traffic forecasts not only for paid advertising campaigns but also for SEO purpose. It helps to identify a keyword’s growing (or shrinking) popularity, filter out low-volume searches, and create hundreds of long-tail keyword combinations that target specific product or service (in our case, content as a product) and landing page, while avoiding high- competitive head keywords.

These target primary keywords were then applied to the WordPress framework (e.g., in title, meta description, and content) that was used by all three websites to improve their on-page SEO. Google Analytics was then used to identify the

Figure 8. Total site traffic. Figure 9. OSE domain authority.

156 S. ZHANG AND N. CABAGE

top performed primary keywords, which usually matched with the target primary keywords. Then Google SERP ranks were then tracked using these top performed primary key- words manually or using third party tools. Figure 10 com- pares the average Google SERP ranks of the three websites using their top performed primary keywords over 18 months. All three websites improved their Google SERP ranks, ranging from the most modest 38% for “(city C) wedding guide” (No. 71 in July 2012) to highest 540% for “(city A) area guide” (No. 5). For the data sets in Figure 10, t-test results also show that the three websites are significantly different in SERP ranks.

Revenue

Likemost content-rich websites, the commercial value and goal of all three websites in our study were to make revenue through advertising (i.e., Google AdSense) and affiliate marketing (e.g., hotel booking). As the biggest online ad network, Google AdSense helps website publishers to achieve the advertising revenue goal by displaying targeted Google ads and leveraging the proprietary Google search and page-ranking technologies. Figure 11 com- pares Google AdSense revenue before and after SEO across the three websites. The specific amount of ad revenue is omitted to protect commercial confidentiality. The ad revenue graph displays a similar pattern as site traffic shown in Figure 7 and domain authority in Figure 9. All three sites began with similar revenue (ANOVA tests find no significant difference). The revenue of Site C (control) stayed stable. Site B experienced a quick upward spike in revenue during the short period of time the social- sharing SEO was implemented, and dove quickly and far behind Site A over time. Whereas Site A (link building) experienced more frequent spikes and continued to grow to the highest revenue, long after the links were built. Interestingly, with only 14% more traffic, Site A was able to realize 69%more in revenue compared to Site B, suggesting that traffic coming from out- bound links is of better quality or higher conversion rate than

traffic from social referrals. For the data sets in Figure 11, t-test results also show that the three websites are significantly differ- ent in ad revenue. Figure 12 compares the total ad revenue gained from the three websites over the course of 18 months: Site A (link building) generated 54% of the overall ad revenue, Site B (social sharing) generated 32%, and Site C (control) generated 13%.

Hypotheses results

Table 6 shows the hypotheses results.

Discussion

We selected three websites that had similar content, site struc- ture, page structure, organic traffic, domain authority, number of links, SERP rank, and revenue at the beginning, and applied different SEO strategies to each one of them (content + link building, content + social sharing, and content only for control purpose) with the same amount of expense and in the same timeline. We tracked and analyzed their site traffic, domain authority, SERP rank, and ad revenue over the course of 18 months. The results show that the website that applied link building generated the most substantial traffic, domain author- ity, SERP rank, and ad revenue increases, and such increases occurred gradually and exponentially after the links were built. The website that applied social sharing generated the quickest traffic and ad revenue increase during the short period of time that the social-sharing SEO was implemented; however most of such increase slowed down almost immediately. The control website with only content SEO and without a link or social SEO

Table 4. Domain authority before and after August 2011–January 2013.

Website SEO strategy

Domain authority before

Domain authority after

Domain authority increase

Site A Content + link building

12 31 +158%

Site B Content + social sharing

11 26 +136%

Site C Content only (Control)

12 21 +75%

Table 5. Links before and after August 2011–January 2013.

Website SEO

strategy

Linking root

domains before

Linking root

domains after

Linking root

domains increase

Total links before

Total links after

Total links

increase

Site A Content + link building

2 26 1,200% 10 598 5,880%

Site B Content + social sharing

2 11 450% 10 209 1,990%

Site C Content only (Control)

3 5 67% 13 37 185%

Figure 10. Keywords and Google SERP ranks.

Figure 11. Ad revenue.

JOURNAL OF COMPUTER INFORMATION SYSTEMS 157

applied also grew in traffic, domain authority, SERP rank, and ad revenue but most moderately among three websites. During the period of data collection (August 2011–January 2013) Google has rolled out 40+ minor to major updates of their search engine algorithm. All three websites had a domain authority upward spike during April/May 2012, which is likely due to a major Google Penguin update (April 24, 2012). Overall, the traffic of all three websites grew continuously throughout the Google updates, suggesting that the websites were developed with sustaining white hat SEO strategies, tech- niques, and tactics.

Implications for practice

For Internet startups, site traffic is critical to their business success or failure. Many have spent a great amount of their financial resources on paid advertisement to attract targeted traffic. Depending on the competition for the keywords they are bidding on, they may not make enough to cover the ad cost. Compared to paid advertisement, organic SEO takes longer time to show effectiveness; however, it costs much less and can have a long-term effect if well executed. While most research studies, white papers, and online articles talked about the primary SEO strategies (content, link building, and social sharing), few studied which one(s) are more effective. The answer depends on many variables such as how the SEO is executed, the size of sources, and the type(s) of the websites. Our study suggests that websites should attempt to build a diverse SEO profile that includes all; and when the resources (e.g., time and money) are limited, focus more on link build- ing and less on social sharing, as link building is more likely

to generate high-quality traffic and create long-term positive effects with the same amount of efforts and resources.

Implications for academia

Although algorithms of search engines remain business secrets, academic researchers worldwide continue to experiment, observe, and triangulate, hoping to find the criteria search engines may use, the quality and cost of organic and sponsored search results, and the effective SEO and search marketing strategies for online businesses to increase traffic and revenue. Common limitations in previous SEO studies have been using single-website data, low traffic/search ranking, and limited KPIs. These limitations may have affected the quality (reliability and generalizability) and quantity of empirical evidence in order to find valuable results. Our study overcame such limitations by setting up a group of three comparable (similar content, site structure, traffic, ranks and revenue to begin with) live websites, applying different SEO stra- tegies to each, collecting and analyzing their KPI data over the course of 18 months. The results answered an inter- esting research question, whether link building or social sharing is a more effective SEO strategy for content-rich websites, which had not been investigated or resolved in any previous academic SEO study.

Limitation

One limitation of our study is using only three websites, and each applied a different SEO strategy (control only, content +

Figure 12. Total ad revenue.

Table 6. Hypotheses results.

Hypotheses KPI results Supported?

H1: Both link building or social sharing are effective SEO strategies.

Site A (Link) increased 267% more in total site traffic, 304% more in revenue, 63% higher in domain authority, and 46% higher in SERP rank than Site C (Control).

Yes

Site B (Social) increased 221% more in total site traffic, 138% more in revenue, 37% higher in domain authority, and 18% higher in SERP rank than Site C (Control).

H2: The effects generated from social sharing occur quicker than the ones from link building.

Site B (Social) increased 503% more traffic and 628% revenue than Site A (Link) in the first two months after SEO implemented, and such increase quickly diminished in following month (−12% in traffic and −44% in revenue than Site A).

Yes

H3: The effects generated from link building last longer than the ones from social sharing.

Site A (Link) increased 14% more in traffic and 69% more in revenue than Site B (Social) over 18 months.

Yes

H4: Link building is more effective than social sharing.

Site A (Link) increased 14% more in total site traffic, 69% more in revenue, 19% higher in domain authority, and 34% higher in SERP rank than Site B (Social).

Yes

158 S. ZHANG AND N. CABAGE

link building, or content + social sharing). Another limitation is that all three websites have a local appeal because the content is geographically based. In such case variables related to location (i.e., differences in demographics in these geogra- phically localized markets) might influence the results.

Future research can attempt to overcome such limitations, and test the same hypothesis using more websites involving a larger variety of business models (eCommerce, SaaS, etc.). However, researchers must be aware that it takes time, money, expertise, and other resources to build or gain access to the backend data of a large group of comparable websites, apply various SEO strategies, collect and analyze the data, for this kind of studies. The websites cannot be simply mock websites or student projects because as discussed throughout this article, the websites must reach substantial site traffic and decent search rank within the top 100 search results for reasonable keywords in order to generate valuable data. Our study focuses on content- rich websites that profit from ads and affiliate sales. It would be interesting to conduct a similar study to different type(s) of websites such as eCommerce and SaaS in future.

Conclusion

Good SEO practice can increase a website’s natural search engine ranking, and drive more traffic to a website—more traffic means more potential revenue from advertisement and/or affiliate for content-rich websites. This study looked at the efficacy of proac- tive link building and social media promotion, compared to creating a content-rich website alone. We found that the website that implemented the link-building strategy generated the highest traffic, search rankings, domain authority, and ad revenue. And the website that implemented social-sharing strategy realized a quick but temporary burst of traffic that slowed down shortly after the sharing activities. Both outperformed the control website that focused only on Content. Our study suggests link building to be more effective in long term. With the same time and budget, link building generated higher traffic, higher SERP rank, and domain authority. These positive results subsequently lead to an increase in ad revenue. These increases occur steadily over time. Social sharing also had a positive effect by creating a quick temporary burst of traffic that slows down shortly after the sharing activities.

Ideally, websites should employ a combination of all three of the techniques studied here. Rich content provides a solid base for natural linking and social sharing. Promoting that content through proactive link building and social media campaigns will help that content to find an audience that otherwise may never discover it, and jumpstart to a fortuitous cycle of natural linking and social sharing, if the content is good and it connects with the audience. In this way, all three approaches can be complimentary, and should be part of a holistic online content marketing strategy.

References

[1] Agarwal S, Singh T, Vipat R. 2011. Latest search engine optimiza- tion techniques to increase website’s ranking in search results. Proceedings of the International Conference & Workshop on Emerging Trends in Technology, Mumbai, Maharashtra, India, February 25–26, pp. 1361–1361.

[2] Berman R, Katona Z. 2013. The role of search engine optimization in search marketing. Marketing Sci. 32:644–651.

[3] Cabage N. Five Pillars of SEO in 2012. http://nealcabage.com/five- pillars-of-seo-in-2012/, November 11, 2011.

[4] ComScore. ComScore Releases September 2014 U.S. Search Engine Rankings. http://www.comscore.com/Insights/Market-Rankings/ comScore-Releases-September-2014-US-Search-Engine-Rankings, October 14, 2014.

[5] DeMers J. 2013. The three pillars of SEO In 2013: Content, links, and social media. http://www.forbes.com/sites/jaysondemers/ 2013/05/23/the-3-pillars-of-seo-in-2013-content-links-and-social- media/, May 23.

[6] Evans MP. 2007. Analysing Google rankings through search engine optimization data. Internet Res. 17:21–37.

[7] Ghose A, Yang S. 2008. Comparing performance metrics in organic search with sponsored search advertising. Proceedings of the 2nd International Workshop on Data Mining and Audience Intelligence for Advertising, pp. 18–26.

[8] Google. 2010. Search Engine Optimization Starter Guide, http:// www.google.com/webmasters/docs/search-engine-optimization- starter-guide.pdf.

[9] Jansen BJ. 2007. The comparative effectiveness of sponsored and nonsponsored links for Web e-commerce queries. ACM Trans. Web. 1:Article 3.

[10] Killoran JB. 2013. How to use search engine optimization techni- ques to increase website visibility. IEEE Trans Prof Commun. 56:50–66.

[11] Li K, Lin M, Lin Z, Xing B. 2014. Running and chasing: The competition between paid search marketing and search engine optimization. Proceedings of the 47th Hawaii International Conference on System Sciences, Big Island, Hawaii, Jan 6–9, pp. 3110–3119.

[12] Malaga RA. 2007. The value of search engine optimization: An action research project at a new e-commerce site. J Electron Commerce Organizations. 5:68–82.

[13] Malaga RA. 2008. Worst practices in search engine optimization. Commun ACM. 51:147–150.

[14] Moz. 2014. “Beginner’s Guide to SEO”, http://moz.com/begin ners-guide-to-seo/.

[15] O’Neill S, Curran K. 2011. The core aspects of search engine optimisation necessary to move up the ranking. Int J Ambient Comput Intell. 3:62–70.

[16] SalesForce. 2015. Social Media Marketing. http://www.exacttarget. com/products/social-media-marketing.

[17] Sen R. 2005. Optimal search engine marketing strategy. Int J Electron Commerce. 10:9–25.

[18] Su AJ, Hu YC, Kuzmanovic A, Koh CK. 2014. How to improve your search engine ranking: Myths and reality. ACM Transactions on the Web 8:Article 8.

[19] Xing B, Lin Z. 2006. The impact of search engine optimization on online advertising market. Proceedings of the 8th International Conference on Electronic Commerce, Fredericton, New Brunswick, Canada, pp. 519–529.

JOURNAL OF COMPUTER INFORMATION SYSTEMS 159

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.