A longitudinal investigation of spam: pre-and post-can-spam legislation.
This study examined the impact of the Can-Spam Act of 2003. A review of tactics used by "spammers" to avoid filtering software and circumvent the legal requirements was examined. Tactics in this study include the use of counterfeit characters, gibberish in the subject line, hidden agendas, invalid return addresses and misleading subject lines. A content analysis was also used to assess the most common types of messages received. The research was comprised of three studies. The first study was conducted four months prior to Can-Spam legislation. The second study was conducted four months after the Can-Spam Act became law, and the final study was conducted one year later, January 2005. Findings were significant for legitimacy of valid email sources, and between the first two studies. Final results of the third follow-up study, one year after the Can-Spam legislation went into effect shows that the law has not been effective in reducing the amount of spam, nor have avoidance tactics been significantly reduced.
According to the Interactive Advertising Bureau (IAB) and Pricewaterhouse (PwC) advertising on the Internet reached approximately $2.43 billion by the third quarter of 2004 (Anonymous, 2004). This represents a 35.3% increase over the same quarter in 2003. Based on the $7.3 billion in revenue during the first nine months of 2004, it should be a record year. Although the industry is experiencing rapid growth in the use of online shopping, the number of spam messages is creating some doubt as to the effectiveness of this media for the future. AOL customers are found to be more likely to click the "report as spam" button than to use the opt-out links provided in the e-mails themselves. McGann (2005) reported that Osterman Research found that because of the increase in the number of spam messages consumers receive, 44 percent report they have decreased their use of email and the Internet in the last year. With this level of decreased usage, there will be a significant impact on legitimate online marketers. Miller (2004) found that as much as 17 percent of all legitimate e-mails are being blocked by ISPs through spam filters.
Spam has penetrated the online environment at rates that may be considered equivalent to an epidemic of catastrophic proportion. Individuals and businesses alike are forced to spend a significant amount of time removing the spam from their daily routines. Erv Shames, former president of Kraft USA, criticized the online industry for tolerating the presence of "spam". He indicated that about 45% of all e-mail is made up of commercial messages unrequested and unwanted by their increasingly angry targets (Elkin, 2003). According to some reports, the increase in spam is costing companies anywhere between $10 billion and $87 billion a year. Message Labs, a London-based Internet security firm has predicted spam to account for 70 percent of all e-mail by April 2004 (Landers, 2003). Jupiter Research reported that blocking legitimate e-mail cost marketers $203 million in 2003 and will jump to $419 million by 2008 (Jakobson, 2004). Henson Rogers, VP of Information Technology for Odyssey Health Care, Inc., reported that if one half of his company's 1000 employees spent five minutes daily dealing with spam e-mail, the cost to the company would be approximately $260,000 per year (Shein, 2004). His comments are based on the fact that every junk message that an employee deals with personally cuts into productivity for the company. The spam clogs company networks and reduces company revenue by slowing legitimate operations. McGann (2004) reported that spam consumed an average 10 8-hour work days per year, and costs employers approximately $2000 per employee. The number of messages delivered to members' spam folders fell 60 percent from a year earlier. A corporate e-mail security firm, Tumbleweed Communications Corp., reported overall spam volumes to still be increasing, making up about 80 percent of the company's inbound e-mail compared to 55 percent at the end of 2003. This company reported estimates of two trillion spam messages to be on the Internet in November of 2004 and at least 25 percent more in December as a result of the holiday (Richmond, 2004). Feig reports that currently the cost of spam to non-corporate users is about $225 million and the cost to all U.S. corporations is about $8.9 billion.
Recent research (Claburn, 2005) indicates spam filtering on average eliminated 68% of spam in 2004. In November of 2004, AOL received an average 2.2 million, daily reports of spam from its subscribers, down substantially from the 11 million daily reports a year earlier. A Nucleus Ferris Research of San Francisco, estimated that spam cost US business nearly $10 billion annual in lost productivity, anti-spam technology and technical support (Costanzo, 2004).
A Wall Street Journal report (Blackman, 2003) in August 2003 reported that Radicati Group, a market-research firm, predicted the number of daily spam messages to be more than 50 billion by 2007 and costs reaching almost $200 billion a year. Consumer response to the unwanted email includes a reduction in the use of e-mail. In a study by Pew Internet & American Life Project (Greenspan, 2003), 60 percent of those responding indicated they have reduced their e-mail usage because of spam and 73 percent avoid giving out their e-mail addresses. Results also indicate that while e-mail users receive slightly more messages in their work, the proportion of spam in personal accounts is higher. This study also reports that 86% of consumers delete the e-mail immediately without opening, 67% click remove me, 33% click to get more information, 21% report the spam to their ISP, and 4% report the unwanted email to a consumer or government agency. However, at least 7% report they order a product or service, 7% provide personal information, and 1% give money as a result of the unsolicited commercial e-mail (UCE).
In a recent article, Reda (2003) reports that spammers can send up to 650,000 messages every hour from an inexpensive e-mail server. Sephos, Inc., Richmond (2004) found that often spammers use computers compromised by viruses and hijacker programs to relay spam anonymously. Sephos estimated that 30 percent of all spam is generated by infected computers. By using the hijacked computers, spammers often avoid or evade the filters used to block e-mail using "blacklists" of known spammers and make it more difficult to trace the actual spammer. This study further reported that about 400,000 infected consumer PCs are being used as spam relay points. According to Susan Reda, Executive Editor of Stores (Reda, 2003), about 50 percent of all messages individuals and businesses receive is spam and at least 66% of that spam is fraudulent by hiding text in graphics so it can't be identified as spam, sending URL's as the body of a message or hijacking company servers. Multi-channel retailers are being caught in the crossfire of consumer demand for less spam and the level of filtering provided by many ISP's. The problems for these retailers include the steady decline in the number of e-mails being delivered or opened. According to this report, AOL blocks up to 780,000,000 spam e-mails from member mailboxes every day. It is expected that $653,000,000 in sales revenue will be generated from anti-spam and content filtering solutions in 2003 and reach $2.4billion by 2007.
Prior to the 2004 Can-Spam Act, the FTC published online information for consumers on how to stop unwanted email. (www.irs.gov) (FTC, 2002). The FTC recommended consumers reduce spam by:
1. not displaying email address in public
3. read and understand the entire form before transmitting personal information through a web site.
4. use two email addresses (one disposable)
5. use unique email address
6. use a filter
The FTC reported the most common spam scams to be chain letters, work-at-home schemes, weight loss chains, credit repair offers, advance fee loan scams, and adult entertainment (Teinowitz, 2003). In 2003 the FTC received more than 110,000 examples of spam on a daily basis and had a database of over 42 million spam messages.
The public and industry outcry for control of the unsolicited e-mail led to legislative action. On December 8, 2003 Congress sent President Bush a bill designed to curb the explosive growth of spam. This bill was called "Controlling the Assault of Non-Solicited Pornography and Marketing Act (http://www.spamlaws.com/federal/108s877.html). Many believed this bill would allow consumers to opt out of unwanted email ads, which accounted for more than half of e-mail traffic. Firms violating this law could be fined $250 per email for sending repeat messages to addresses that opt out of future ads. Companies may be fined up to $6 million.
The Can-Spam Act has five basic requirements: 1) emails may not contain false and misleading messages, 2) must have functioning return address and an opt-out mechanism that prohibits mailings after the 10-day opt-out period 3) must contain disclosure requirements (identify as advertisement, 4) provide notice of opt-out, and a warning label for sexually oriented material), and 5) prohibits aggravated violations, such as harvesting addresses (Manishin & Joyce, 2004). The requirement that consumers must contact each sender to opt-out of future email instead of forcing senders to have an opt-in permission has received criticism from the industry. Other limitations of the law include preempting some state laws, prohibiting private citizens from suing spammers, and allowing only state attorneys general or Internet service providers to sue (Gross, 2005). This legislation is a move in the right direction, however, legislation alone will not be able to stop the spam, since spammers often operate outside the law anyway. In order to control the proliferation of spam, it will take legislation, increased technology, and legal action against the violators. On January 1, 2004 the new federal anti-spam law went into affect. These requirements were listed and discussed above. This law permits tougher penalties for use of e-mail addresses from Web sites, use of automated tools to create multiple e-mail accounts, federal agencies to seek jail time, states to seek civil penalties up to $2 million and Internet service providers up to $1 million, and a federal study to create a do-not-spam list. However, Brightmail Inc. reports that 58 percent of incoming e-mail one week after the law took effect was spam, no change from the previous month (Jesdanum, 2004).
By mid-January 2004, it became apparent that the new law would not solve the spam e-mail problems. The Credit Union Executive Society (McDonald, 2004) reported that in an analysis of over 100 opt-in messages, 44 percent did not comply with one of the simplest requirements of the new law, a postal address in the body of each message. Mangalinden reported that Time Warners' AOL, Earthlink Inc., Microsoft Corp., and Yahoo, Inc. have filed six civil lawsuits against hundreds of alleged spammers. The problem is that all but seven of the 222 defendants are unnamed. The Internet providers cannot confirm true identities of the senders of the spam. Brightmail, maker of anti-spam technology, found further evidence of the difficulty in identifying and controlling spammers. In a study conducted by their firm, Brightmail found that in February 2004, 64 percent of all e-mail traffic was spam, up from 58 percent in December. In March of 2004, 68 percent of all email was spam. This was identified through a filtering process that filtered 2.93 billion fraudulent e-mails, up 25 percent from the previous month. MXLogic had similar results and reports only 3 percent of the e-mails received had met the FTC requirement for postal addresses by mid-February, 2004 (Marson, 2004). Although action is under way to reduce the amount of spam, the volume of spam during the first month the act was in effect actually increased and has climbed since. The Radicati Group predicted the number of spam messages worldwide to be about 35 billion in 2004, this is more than double the amount in 2003 (Garretson, 2004). According to a report in InformationWeek (Anonymous, 2004), Postini Inc, a company that processes 2.4 billion messages per week for 4000 business clients, 88% of all email in November of 2004 was spam, phishing, viruses, or directory-harvest attacks. This is further evidence that the Can-Spam Act did not stop spam, spammers are simply becoming more sophisticated and are looking for new ways around the law. Feig reports that of the 31 billion e-mail messages per day, 12.4 billion are spam. On the average, an email user will receive at least six spam messages per day.
The USA Today reported that there are 1,000 to 2,000 spammers worldwide with approximately 200 accounting for up to 90 percent of about 2 trillion junk e-mail messages each year (Swartz, 2004). Although Swartz reports that many treat the anti-spam law like jaywalking, many are starting to leave the business due to the convictions others have received. Damon DeCrescenzo, one of the world's biggest spammers, dropped out of the business during 2004 as a result of the new federal anti-spam law. Fielding (2004) reported that in September 2004, approximately 4 percent of unsolicited commercial e-mail complied with the Can-Spam Act.
In November, two North Carolina residents became the first to be convicted of felony spamming charges in Virginia (Feig, 2004). In December of 2004 a federal judge in Davenport, Iowa awarded an Internet service provider more than $1 billion; AMP Dollar Savings was ordered to pay $720 million, Cash Link Systems of Miami, Florida was ordered to pay $360 million, and Florida based TEI Marketing Group was ordered to pay $140,000. Although the judgment was received, the ISP provider does not expect to receive payment.
Manishin and Joyce (2004) identified three legal issues involving the Can-Spam Act. First there is question regarding the constitutionality of the Act. This is based solely on attempts to have the do-not-call list found to be unconstitutional. Second, there is concern that the federal law preempts the state laws, many of which were much more restrictive. The third major issue is the extraterritorial jurisdiction, especially between foreign statutes such as the European Union laws which are much more protective.
The Internet Crime Complaint Center, a joint effort by the FBI and the National White Collar Crime Center, has refined its databases, shares data, and provides education and training to federal and state agencies regarding techniques used by spammers, tactics to investigate spam schemes and the tools available as a result of the Can-Spam Act (Cox & Marson, 2004). In late October and early November of 2004, Congress passed several anti-spyware measures. The I-SPY Act calls for the FTC to oversee online software distribution and the SPYBLOCK Act regulates advertising delivered via interactive software and spyware designed to hijack end-user's computers. As of December 2004, the FTC had filed five suits under the CAN-SPAM Act. In addition Massachusetts and Washington have filed suit under the federal law and four major ISPs have gone after hundreds of spammers (Garretson, 2004).
PURPOSE OF STUDY
The overall purpose of this study was to investigate the preponderance of spam, and the tactics employed by spammers. Furthermore, the researchers wanted to examine the quantity of spam received prior to and following passage of the Can-Spam Act. Specifically the researchers were interested in determining what percentage of total email received was spam and how effectively the Can-Spam Act controlled spam.
According to Teinowitz (2002) the FTC has reported that about two-thirds of all e-mail consumers receive is misleading. They describe the use of deceptive sender addresses and subject lines as those that exceed the number of characters allowed and phony offers found within the spam mailer. Currently the FTC believes 40 percent of spam e-mails contain text that appears deceptive, 44 percent use a fraudulent subject or sender address, and approximately one third have phony "from" addresses. Many of the e-mails mislead the consumer by falsely suggesting the message is from a friend or business associate. Spammers use several different techniques to avoid e-mail filters. Some of the most common methods include counterfeit characters in words, gibberish (literally letters that don't comprise words) in the subject line, hidden agendas--use of codes or written in white on white background, and treacherous tracks--incorrect email addresses (Rapoport, 2003). Additionally, the researchers wanted to determine if the "Can-Spam Act" was effective in reducing the avoidance tactics used by spammers by comparing the total number of spam messages received prior to the implementation of the act to the total number of spam messages received after the Can-Spam Act became federal law and the effectiveness one year later.
The specific research questions for the initial study were:
1. What percentage of spammers used the following avoidance tactics?
a. Hidden agendas
b. Gibberish in subject line or body
c. Counterfeit characters
d. Invalid return address
e. Misleading subject line
2. What are the most common content "offenders?"
Methodology of Study 1
The two researchers monitored their individual email on the university server for a period of one week, during September 2003. Each email was evaluated to examine the most common tactics being used by "spammers." It should be noted that the university did not have any spam filtering software on the email server. However, information technology personnel manually block email from individual senders that are subjectively classified as serious offenders. Therefore, some messages were systematically blocked, but not technically filtered. For the purpose of this study, the tactics investigated include: agreement between the subject line and the body of message; legitimacy of an unsubscribe, opt-out provided by the sender; the use of counterfeit characters in the subject line; and gibberish in the subject line or body of the message. In addition, the researchers examined hidden agendas and whether or not the email address of the sender was a valid address.
The authors saved and printed emails identified as spam sent to their respective academic computing e-mail account for one week (7 days). A total of 326 spam messages were received, one account with approximately 113 messages, the other 211 messages. Each message was analyzed and the following data items were noted:
1. sender identification (legitimate email address?)
2. subject line of the message
3. content of the message
4. if the subject and body message content agreed
5. if there was the option to unsubscribe from a list
6. if the message contained counterfeit characters
7. gibberish in the subject line
8. hidden agenda
9. if there was a valid email or active link to unsubscribe
Results of Study 1
While it was presumed that many of the spam messages received were duplicated, in actuality this was not the case. There were several spam scams present in the dataset, as suggested by the FTC. These included: pornography, credit offers and low-rate loans, discount drugs, and money making schemes. See Table 1 for frequencies of the most prevalent content. Further analysis of the data reviewed the nine tactics listed above and usage rate for each. Results of this analysis found over 28% of the messages received did not have agreement between subject lines and body content. Over one-third contained gibberish, along with 17% that had counterfeit characters. Seventy-one percent did NOT provide an unsubscribe option, and almost 30% contained a hidden agenda, while only slightly more than one-fourth (28%) originated from a valid email address. Specific frequencies can be found in Table 2.
Since only slightly more than one-fourth of the messages were received from valid email addresses, a chi-square test of independence was calculated comparing the validity of the sender email/unsubscribe address and utilization of other tactic variables. Results were statistically significant for all tactic variables. Specific chi-square values and significance levels are presented in the Table 3, Chi-square Analysis.
Results of the cross-tab analysis found a greater number of messages with a valid email address had an unsubscribe option, slightly more exhibited subject message agreement, and fewer contained gibberish, counterfeit characters, or a hidden agenda. This would suggest that legitimate email marketers were attempting to put forth an "honest effort" in their email marketing tactics, while true "spammers" were definitely employing a high percentage of hidden agenda, gibberish, and not allowing receivers to unsubscribe.
Methodology of Study 2
During late spring of 2004, the researchers repeated the original data collection procedure for the purpose of comparison, between the pre- and post-Can-Spam Act. It seemed to each researcher that their individual email accounts contained a much larger percentage of spam, so this time each researcher counted the total number of emails received daily, and the number of these emails that constituted spam. Results, as seen in Table 4, show that three-fourths or more of the email received each day was actually spam. This finding supports the prediction of Message Labs that spam would account of 70 percent of all email by April 2004 (Landers, 2003). However, it should be noted that while the university did NOT have any spam filters in place during the study at any time, Information Technology department would block the worst offenders before email reached the ultimate receiver.
The data set was analyzed for the prevalence of body content. The most common types of content included: adult/pornography, low-rate loans, credit, discount drugs, and money-making schemes. The researchers noted a significant increase in the amount of pornographic messages. During the second series of data collection it was noted that the amount of pornography had increased almost ten-fold. Messages containing credit offers were more than slightly double those received in the fall. However the percentage of discount drug offers and money making schemes had decreased from 22.4% and 5.8% to 15.3% and 2.1%, respectively and low rate loan messages decreased slightly, 8.9% in the fall to 7.3% in the spring data collection. See Table 5 for specific frequencies and percentages.
Content analysis of the spam tactic variables for the spring 2004 showed that the number of messages exhibiting agreement between the subject and body decreased from almost three-fourths (2003) to slightly less than half, 48.6% in the spring data set. In addition, messages with a hidden agenda also decreased by almost 30%. The percentage of messages containing an unsubscribe option increased to slightly more than half, 52.6%, which was not quite twice the percentage of those present in the fall. The presence of counterfeit characters and gibberish increased, to 50.1% and 42.3% respectively. The most alarming result was the substantial decrease of messages containing an actual valid email or unsubscribe option. This decreased from 28% in the fall of 2003, to only 12% in the spring of 2004.
Chi-square analysis and cross-tabs were again calculated comparing the validity of the sender email/unsubscribe address and utilization of other tactic variables. Results were statistically significant for all tactic variables. Specific chi-square values and significance levels are presented in the Table 6. Cross-tab analysis found that more of the messages containing a valid email address also exhibited agreement between subject line and body content, an unsubscribe option, while fewer exhibited counterfeit characters, gibberish, and a hidden agenda. These results would suggest that while fewer messages were coming from a valid email address, those individuals sending legitimate email marketing messages were now using fewer scam tactics, and it could be inferred, attempting to follow the new legislative requirements.
COMPARING STUDY ONE AND STUDY TWO
Chi-square analysis was performed to compare the data collected in the fall 2003, versus spring 2004. Results were statistically significant for five of the six spam tactics. The only variable not significant was the use of gibberish in the subject line to avoid filtering software. Specific results are presented in Table 8. Analysis using cross-tabs provided insight comparing the fall 2003 data frequencies of the variables compared to the spring 2004 occurrences of each agenda item. Although results found that the percentage of messages that provided an unsubscribe option almost doubled, the use of counterfeit characters more than doubled. The spring 2004 dataset contained fewer messages that had a hidden agenda, agreement of the subject and body, and most importantly, the number of messages sent from a valid email address decreased by more than half. These results indicate that while the Can-Spam Act has had an impact for legitimate email marketers, spammers have increased their use of scamming and avoidance tactics. It would appear that spammers have gone "underground" so that they can't be traced.
The purpose of the third study was to analyze the flow of spam to the researchers' email accounts, one year following implementation of the Can-Spam Act. The research again tracked the spam received in their respective individual email accounts, and compared the amount of spam received to legitimate emails. The university had implemented a spam filter in late fall of 2004. Therefore, during this study, the majority of spam was placed in a sub-folder on individual email accounts. Researchers could scan this folder and determine if the email was legitimate, and if so it then had to be released to the email account. However, the contents of the email body in the spam folder could not be examined unless it was released to the email account. Therefore, the researchers determined it would be advisable NOT to examine the actual body of the email for contents and agreement with the subject line, unsubscribe option, or for a valid email address. Rather, this study would examine the quantity of spam received, and the use of counterfeit characters and gibberish in the subject line, as well as the most frequent content in the subject line.
Each researcher recorded the information contained in the spam filter daily report, and checked individual email for additional spam messages that got through the filter. Results are presented in Table 9. Researcher A noted that the number of spam messages was approximately double the number received the previous spring term, while the total number of spam messages was consistent between the two studies for Researcher B.
The tactics examined in this study were the use of counterfeit characters and gibberish. Use of counterfeit characters decreased substantially overall, from 50% in 2004 to 8%, but the use of gibberish increased slightly, from 42% to 49%. See Table 10 for specific frequencies in the 2005 data. However, it should be noted that Researcher A's account contained such a significantly large number of messages containing gibberish in the subject line that it was impossible to determine the content of the messages in this account. Table 11 shows the frequency of gibberish and counterfeit characters, across the 2004 and 2005 data for the individual accounts.
Chi-square analysis and cross-tabs were calculated comparing the frequency of gibberish and counterfeit characters for years 2004 and 2005. Results were statistically significant. Specific chi-square and significance levels are presented in Table 12. Cross-tab analysis found that while messages containing gibberish were less than expected in 2004, they were greater than expected in 2005. The inverse was found for the use of counterfeit characters in the two years. The individual researchers noted what appeared to be a significant difference between the use of these two tactics in the individual accounts. It should be noted that the use of gibberish increased from 50% to 55% for researcher A, and increased from 5% to 9% for researcher B. The use of counterfeit characters significantly decreased for Research A, going from 57% of spam received in 2004 to 5% of the spam received in 2005; however counterfeit characters decreased from 14% in 2004 to 9% in 2005 for researcher B. Therefore, another chi-square analysis and cross-tabs were calculated comparing the frequency of these two tactics for Researcher A and Researcher B. Results were statistically significant. Specific chi-square and significance levels are presented in Table 13. Cross-tab analysis found that Researcher A's account had significantly more gibberish while Researcher B's account had much less. Cross-tab analysis for the use of counterfeit characters found that Researcher A's account had fewer message than expected employing this tactic, while Researcher B's account show expected levels.
The most frequent content of spam messages included adult/pornography and discount drugs. The frequency of credit offers, discount loans and money making offers decreased to only a handful or less, see Table 14 for frequencies of content.
COMPARISON AND ANALYSIS ACROSS THREE DATA SETS
Comparing the three datasets shows the following trends:
1 The number of spam messages increased substantially with each year, more than doubling from May of 2004 to January of 2005.
2 The use of counterfeit characters increased substantially in the spring of 2004, but had significantly decreased in the spring of 2005.
3 The use of gibberish has continued to increase at approximately 5% with each year's data collection.
Specific frequencies for the number of spam, and percentage of spam containing counterfeit characters and gibberish can be found in Table 15. Chi-square analysis comparing the three data sets found the differences in use across the three years was statistically significant, and results are presented in Table 16.
The questions this study attempted to answer were: what percentage of spammers use avoidance tactics, what content was most common, and how effective has the Can-Spam legislation been in controlling spam? Results found that prior to the Can-Spam Act, the majority of spammers used hidden agendas, did NOT provide an unsubscribe option and came from an invalid email address. Post-Can-Spam Act data show that more of the messages received provided an unsubscribe option, now slightly more than half; half contained counterfeit characters, and slightly less than half contained gibberish. Most startling was that fact that the percentage of messages coming from a valid email address decreased from 28% in the fall to only 12% in spring 2004. However, validity of email address could not be verified in the third study. Furthermore, the number of spam messages received had almost doubled. The answer to the question posed at the outset of this study is: Has the Can-Spam Act been effective in controlling spam? These results clearly show "spammers" are avoiding the requirements set forth in the legislation.
Both the private sector and business organizations continue to call for control of the spam email crisis. It has been predicted by some that email will no longer be recognized or utilized as a major communication tool unless the spam epidemic can be controlled. Some individuals suggest charging for email, the development of "no email" lists, additional government legislation and involvement, prosecution of violators, and an outright closure of the email process. In an effort to respond to these and other issues in dealing with the problems created by spam, TRUSTe, a leading provider of privacy certification and seal programs, testified before a Senate hearing that more than half of consumer complaints are a result of unwanted spam. As a result of these complaints, ISP's are creating filters, which block about 40% of all e-mail as spam. This sometimes creates a "false positive" resulting in about 15% of legitimate e-mail not getting delivered (Hodge & Mattox, 2003). In an effort to provide consumers with a method to screen unwanted email and still receive legitimate messages, IronPort Systems and TRUSTe launched the Bonded Sender Program. This Program allows legitimate senders of mail to avoid being blocked by overly aggressive spam filters by allowing senders to identify themselves, adhering to standards and posting a financial bond (Landis, Matick, Hodge, & Sullivan, 2003). In October 2003, the IAB (Interactive Advertising Bureau) with TRUSTe released the "Email Marketing Pledge"--a set of email marketing guidelines (IAB, 2003). These guidelines require informed consent before sending email. The Pledge is expected to increase industry accountability by more clearly differentiating between legitimate mail and spam.
Until consumers can easily differentiate between spam and legitimate email, they will employ tactics to make their inbox manageable. In many instances these actions mean the marketing messages sent, even those sent by legitimate senders will not reach the receiver. Consumers reported that during the Holiday 2004 shopping season, they simply deleted additional email they received (McGann, 2005). Furthermore, 27% unsubscribed from email lists that sent them messages too frequently; 23% regularly used the ISP's mail program "this is spam" button. These numbers, and the increasing number of spam filters being employed by individual consumers, as well as ISP's and corporate mail servers, should serve notice to legitimate email marketers. Marketers should move from mass marketing to targeted marketing, as well as be sure that the receiving consumer is a legitimate customer who wishes to receive email marketing messages.
It appears that the Can-Spam law, in effect for a year, has not been successful in squelching unsolicited e-mail (Hulme, 2005). Even worse, it is estimated that about 75% of email is spam (Snyder, 2004); the volume of spam is so high, that it has dominated Internet message flow. Christopher Conkey (2005) fears the legislation will not make much of a dent in the steady flow of illegal spam, since industry analysts report the phenomenon worsened in 2004, and most estimates indicate spam account for 70-80% of total email traffic. Results of this study support these findings. The Can-Spam Law has had little if any effect on the number or type of spam emails being generated. There is clear evidence that "spammers" are becoming savvier in the types of avoidance tactics utilized. Companies must develop means to filter and control the amount of spam email messages received before the public becomes so distrustful that they will no longer open any commercial email message. The results of this study should be a great concern to legitimate email marketers, as the current state of the email marketing environment shows that spam is a serious two-fold threat: first to email marketers ability to get their messages through the clutter of spam, and to consumers' ability to trust the message senders.
Anonymous. (2004, November 15). 2004 interactive advertising revenues total over $2.4 billion: Fourth record-setting quarter. Retrieved January 20, 2005, from http://www.iab.com/news/pr_2004_11_15.asp
Anonymous. (2004, December 6). Ok, spam, enough already. InformationWeek, 12.
Blackman, A. (2003). An 'easy target' for e-mail spam. The Wall Street Journal.
Claburn, T. (2005, January 17). Spam: Are we winning the fight? Information Week, 20-21.
Conkey, C. (2005, January 12). Ftc wins order to shut down spam from adult web sites. Wall Street Journal, 2.
Costanzo, C. (2004, December 1). You've got spam! Community Banker, 13, 24-28.
Cox, J., & Marson, C. (2004, November 20). Is the law's arm long enough? Network World, 21, 50-51.
Elkin, T. (2003, May 7). What online media sellers aren't doing well. Retrieved January 15, 2004, from http://www.adage.com/news.cms.?newsId=37781
Feig, N. (2004, December). Spam it! Community Banker, 13, 68.
Fielding, M. (2004, December 15). Can spam compliance spotty. Marketing News, 38, 20-21.
FTC. (2002). You've got spam: Federal Trade Commission.
Garretson, C. (2004, December 13). Federal anti-spam law gets mixed results. Network World, 21, 9.
Greenspan, R. (2003, October 23). Spam: Always annoying, often offensive. Retrieved January 30, 2004, from http://cyberatlas.internet.com/big...ions/print/ 0,1301_3097351,00.html
Gross, G. (2005). U.S. Can spam act struggles to make a difference. Networld World, 27, 16.
Hodge, C., & Mattox, K. (2003, May 21). TRUSTe weighs in at senate hearing on spam. Retrieved December 1, 2003, from www.truste.org/about/ Senate_Testimony_Alert.html
Hulme, G. V. (2005). Another fight to wage: Companies buried by spam focus their attention--and resources--on the battle against spyware and adward. Information Week, 60-63.
IAB. (2003, October 13). A new front in the war on spam. IAB Informer October 13. Retrieved December 1, 2003, from http://www.iab.ent/informer/ informer_10_03.asp
Jakobson, L. (2004). First spam, now spim. Incentive, 178(4), 14.
Jesdanum, A. (2004, January 31). Spam is winning the war despite new laws and technology, inboxes still flooded. Lexington Herald Leader.
Landers, J. (2003, December 9). Skeptics wonder if anti-spam bill sent to bush will really work. Lexington Herald Leader.
Landis, L., Matick, S., Hodge, C., & Sullivan, M. (2003, October 13). Ironport systemstm launches next generation of bonded sender programtm program. Retrieved December 15, 2003, from http://www.truste.org/about/ BondedSenderLaunch.htm
Manishin, G. B., & Joyce, S. A. (2004). Current spam law & policy: An overview and update. Computer and Internet Lawyer, 21(9), 1-7.
Marson, C. (2004). Ietf to lead anti-spam crusade. Network World, 21(15), 1-2.
McDonald, L. (2004). E-marketing can(t) spam. Credit Union Management, 27, 58-59.
McGann, R. (2005, January 26). Holiday e-mail campaigns receive mixed report. Retrieved February 5, 2005, from http://www.clickz.com/stats/sectors/email/ article.php/3464351
Rapoport, R. (2003, August 19). Spam senders change tactics. Lexington Herald Leader, C1-C2.
Reda, S. (2003). You've got spam. Stores(September), 26-30.
Richmond, R. (2004, December 28). AOL notes decline in junk e-mail, as trend reverses. Wall Street Journal.
Shein, E. (2004). Get rich quick fighting spam. CFO.com, 1.
Snyder, J. (2004, December 20). Top spam fighters offer feature diversity. Network World, 21, 34-38.
Swartz, J. (2004, May 6). New software laws push some spammers to log out. USA Today, B.01.
Teinowitz, I. (2002, September 4). Consumer groups seek tougher spam crackdown. Retrieved December 20, 2003, from www.adage.com/news.cms?newsId=35958
Teinowitz, I. (2003, April 29). Ftc attacks spam as threat to e-mail. Retrieved December 1, 2003, from www.adage.com/news.cms?newId=37726
Peggy Osborne, Morehead State University
Michelle B. Kunz, Morehead State University
Table 1 Most Prevalent Content Categories Fall 2003 Spam Content N = 325 Adult/Pornography 4 1% Credit 5 1.5% Discount Drugs 73 22.4% Low rate loans 29 8.9% Money making 19 5.8% Table 2 Frequency of Spam Scam Tactics Fall 2003 Variable Freq Percent Subject Body do not agree 87 28.7 No Unsubscribe option 228 71.7 Use Counterfeit characters 55 17.3 Rambling/gibberish 121 37.9 Hidden Agenda 93 29.7 Invalid email 226 72.0 Table 3 Chi-square analysis Valid Email Fall 2003 Variable [X.sup.2] df p Unsubscribe option 306.23 1 .000 Gibberish 18.52 1 .000 Counterfeit characters 6.86 1 .009 Subject agreement 11.22 1 .001 Hidden agenda 7.79 1 .005 Table 4 Frequency of Spam Received--Spring 2004 Account A Account B Day Total #spam %spam Total #spam %spam Sunday 79 69 97.34 13 11 84.62 Monday 109 89 81.65 21 18 85.71 Tuesday 94 79 84.04 19 17 89.47 Wednesday 108 72 66.67 27 23 85.18 Thursday 92 69 75.00 24 21 87.50 Friday 93 73 78.49 26 17 65.38 Saturday 65 56 86.15 25 19 76.00 TOTAL 640 507 79.21 avg. 155 126 81.29 avg. Table 5 Most Prevalent Content Spring 2004 Spam Content N = 626 Adult/Pornography 62 9.9% Credit 23 3.6% Discount Drugs 96 15.3% Low rate loans 46 7.3% Money making 13 2.1% Table 6 Frequency of Spam Scam Tactics Spring 2004 Variable Freq % Subject Body do not agree 322 51.4 No unsubscribe option 297 47.4 Counterfeit characters 314 50.1 Rambling/gibberish 265 42.3 Hidden agenda 62 9.9 No valid email 552 88.0 Table 7 Chi-square analysis Valid Email Spring 2004 Variable [X.sup.2] df p Unsubscribe option 46.03 1 .000 Gibberish 54.74 1 .000 Counterfeit characters 64.22 1 .000 Subject agreement 23.09 1 .000 Hidden agenda 2.18 1 .000 Table 8 Chi-square comparison 2003 versus 2004 Variable [X.sup.2] df p Tau p Unsubscribe option 54.82 1 .000 .058 .000 Gibberish 2.42 1 .120 .003 .075 Counterfeit characters 100.08 1 .000 .105 .000 Subject agreement 26.61 1 .000 .028 .000 Hidden agenda 54.55 1 .000 .057 .016 Valid Email 34.05 1 .000 .036 .000 Table 9 Frequency of Spam Receipts Spring 2005 Account A Account B Day Total #spam %spam Total #spam %spam Monday 191 179 92 21 18 85.71 Tuesday 208 167 80 19 17 89.47 Wednesday 228 213 93 27 23 85.18 Thursday 226 191 85 24 21 87.50 Fri-Sun 500 463 93 26 17 65.38 TOTAL 1353 1210 89% avg. 1823 156 86% avg. Table 10 Frequency of Spam Tactics January 2005 Variable Freq Percent Counterfeit characters 114 92 Rambling/gibberish 675 49 Table 11 Frequencies 2004 vs 2005 Gibberish between 2 accounts 2004 2005 A B A B Freq % Freq % Freq % Freq % Counterfeit 297 57 17 14 65 5 49 9 Gibberish 254 50 11 9 671 55 4 0.7 Table 12 Chi-square comparison 2004 versus 2005 Variable [X.sup.2] df p Tau p Gibberish 7.18 1 .000 .004 .000 Counterfeit char. 452.03 1 .000 .225 .000 Table 13 Chi-square comparison Research A vs B for 2005 Variable [X.sup.2] df p Tau p Gibberish 167.19 2 .000 .072 .000 Counterfeit char 379.6 2 .000 .161 .000 Table 14 Most Prevalent Content Spring 2005 N = 1384 Spam Content Acct A Acct B Total % Adult/Pornography 36 25 71 5 Credit 0 0 0 0 Discount drugs 90 39 129 9 Low rate loans 1 2 3 .002 Money making 0 0 0 0 International source 346 44 390 29 Table 15 Frequency of Spam Receipts for all 3 data sets Year Frequency %Counterfeit % Gibberish 2003 326 16.9 37.1 2004 626 50.2 42.3 2005 1384 8.2 48.8 Table 16 Chi-square Comparison 2003, 2004, 2005 Variable [X.sup.2] df p Tau p Gibberish 19.41 3 .000 .008 .000 Counterfeit char 466.05 3 .000 .119 .000
|Printer friendly Cite/link Email Feedback|
|Author:||Osborne, Peggy; Kunz, Michelle|
|Publication:||Journal of Strategic E-Commerce|
|Date:||Jan 1, 2005|
|Previous Article:||The role of internet-based transactions in external collaboration between healthcare trading partners.|
|Next Article:||.Edu dilema: the web accessibility challenge facing public and private universities.|