Vol.8, No. 8 • Sept. 2000

• WEB WISE WAYS •  The Last Word on Web Monitoring and Clipping Services by Amelia Kassel, MarketingBase

CyberAlert
To recap from the previous article, CyberAlert [http://www.cyberalert.com] is an automated Internet monitoring and Web clipping service that provides clips about what is being said about a company or its products from a selection of Web publications, other Web sites, message boards, and Usenet news groups. Clients specify keywords or phrases, CyberAlert staff sets up the profile and search request, uses proprietary technology to automatically search the Internet every day, and finds, clips, and reports new mentions of the client’s keywords.

CyberAlert uses Boolean and set logic, which allows for nesting and proximity operators. Unlike most current Web search engines, the CyberAlert technology has much of the same functionality of traditional commercial vendors such as Dialog and LEXIS-NEXIS, but takes it even a step further for Web applicability. Proximity capabilities include, for example, finding terms within the same line, sentence, paragraph, or page — with page proximity referring to an HTML page. If the search string needs revision, the staff works with a client to refine it according to the client’s requirements. The CyberAlert software is proprietary.

CyberAlert delivers a daily report containing new citations from the previous 24 hours. At the present time, it searches 2,500 selected Web publications, with 50-100 new Web sites added each week. These cover the major news syndication services, newspapers, magazines, journals, and online publications on the Internet. Since some Web sites use “robot.txt” files to block outsiders, CyberAlert respects these wishes and won’t go to such sites. For example, The New York Times requires registration.

CyberAlert will add sites suggested by customers at no extra cost. (For basic charges, see my July/August Searcher column). However, the list of Web publications searched by Cyber-Alert is not for public consumption. CyberAlert considers it proprietary, having spent a good deal of time determining the utility and availability of sites. The names of publications, however, are available to customers and prospects who sign a confidentiality agreement. Publications are grouped into media categories including syndication services, national news services, newspapers by country and region, TV networks, advertising and marketing, aerospace and aviation, automotive, banking and finance, business, construction, chemicals, education, energy, entertainment, environment, games, government, legal, medicine and science, publishing, retail, sports, technology (with subsets), transportation, travel, Web-only publications, and more. CyberAlert also includes international publications, including those from Canada, the United Kingdom, and Europe. Clients can choose to monitor the entire list of publications, as I did in my test, or select categories. CyberAlert also uses multiple public search engines for searching more than 2.2 million Web sites.

I spoke with David Blumstein, director of National Sales. He explained that unlike the traditional print clip services such as Burrelle’s, Luce, or Bacon, CyberAlert does not provide clippings from print sources. Indeed, if a company uses CyberAlert in addition to a print clipping service, it will probably receive duplicates since many of the previously, in-print-only publications now appear on the Web on publisher Web sites where CyberAlert captures their content. Blumstein suggests, however, that CyberAlert brings value to clients who end up saving money by receiving timely information on a daily basis that they didn’t know about but should. For example, he commented on instances when CyberAlert clips alerted a client company about lawsuits filed against them of which they were not yet aware. Or, a client may learn that one of their dealers is functioning in some way that harms the reputation of the company. (CyberAlert, in fact, plans an upcoming seminar on Reputation Management.) One of the key reasons for using a Web clipping service is to track derogatory or harmful remarks.

Commentary conveyed over the Internet no longer goes unnoticed or unremarked. While writing this very piece, I received a daily alert from the Wall Street Journal Interactive [http://interactive.wsj.com/articles/HeardOnTheNet.htm, July 14, 2000] describing a lawsuit filed by Credit Suisse First Boston (CSFB) seeking $1 million in damages against a New Jersey man and 10 others who posted comments on a Yahoo! Finance message board criticizing one of CSFB’s investment research analysts. Blumstein points out that CyberAlert provides crucial information that would not readily turn up in a timely way. Finding the details, assuming you knew a story existed out there, could take two to three employees surfing the Web all day every day to get results. I’ve heard of some corporate libraries employing full-time staff for monitoring purposes. A service such as CyberAlert is a valuable addition since its technology gathers input automatically, freeing information professionals to spend their time evaluating usefulness and determining various applications for the information located, such as organizing and disseminating it to appropriate staff.

After subscribing to CyberAlert, you give your request(s) to a representative who sets up an account and search strings. Clients receive notification with results by e-mail and a choice of either a CyberAlert summary or abstract for each item, the number of hits for that day, and the direct URL for each item. Summaries contain the list of titles and URLs of search matches. Abstracts include titles, URLs, and several sentences occurring around the matched keywords. CyberAlert abstracts are not like the substantive abstracts or summaries known to information professionals from databases such as ABI Inform, ERIC, or Medline, but rather more like the keyword in context (kwic) format with several sentences surrounding the selected key words [See the sidebar samples.]

On the first day, you’ll get more hits than subsequently since the output comes from an initial scan of everything that is covered. From day two and on, CyberAlert only delivers new clips based on what’s in the preselected sources from the previous 24 hours. Depending on the request, you may get no clips at all on any one day. In my request to CyberAlert, I received results almost every day through the test period of about 2 weeks. On those days with zero results, I received notification by e-mail of zero results, confirming that the system functioned every day. The e-mail alert contains hotlinked URLs, and you can scan quickly for items of most interest and click straight through to the site —  depending on your e-mail software.

Rather than using the links from e-mail, however, I recommend logging in to your CyberAlert Web account, where you’ll find some wonderful features that turn CyberAlert into more than just a clipping service. A Welcome Screen [see Figure 1] contains an Inbox with your clips by folder name (assigned by CyberAlert) and several housekeeping folders, including Help. When you click on your clips folder, the first 15 results display [see Figure 2], and a database of every clip from the time you began your subscription is accumulated for you. Each clip is saved during the entire life of your subscription unless you decide to throw it away by placing it in a trash folder. As an aside, if your usage exceeds 3,500 citations, a surcharge for space applies. Depending on the request, you could reach that number fairly quickly. Within a 2-week period, my topic results totaled about 650 clips.

I found this “clip database” (my phrase) particularly handy. For one thing, I didn’t have time to review my clips on a daily basis and enjoyed finding all the results waiting for me when I finally had a chance to log on and use it at my leisure. On the Web site, CyberAlert offers a Boolean search capability for isolating particular terms or companies within your folder. In this way, you can look at a limited number of hits and group topics in a variety of ways, similar to the FOCUS capability in LEXIS-NEXIS. Unfortunately, CyberAlert offered no explanation about how to conduct a search, i.e., no search tips. After discussing this problem with a CyberAlert representative, I was told that an explanation would be added. This type of help is particularly important since every system I’ve seen, from commercial to Web search engine, has some variations on a theme when it comes to searching.

Here are some other useful features on the Web site:

• A pull-down mention option allows you to sort results by date order (the default), or by Title, URL, List, or Media Group.
• You can view 5, 15, 50, or 100 hits at a time (the default is set for 15).
• There are three format (mode) choices — compact, normal, or expanded — with the default set at normal containing the header, URL, and one or more of your keywords in context as an abstract [see Figure 2]. Compact mode [see Figure 3] displays only the folder item number, the citation number, the title, the source publication, the date found, and the URL. Extended mode displays the item number, the citation number, the title, the source publication, the date found, the URL, the abstract, and a Notes field for comments entered by the user. This mode also permits sending an individual citation to a printer for hard-copy output. With a little adjusting you can see results in one mode and store them in another.
• A capability for sorting citations into different folders that you create for easier use.
• A function that allows you to forward a particular citation or even the whole folder to another person via e-mail.
• An option to review a clip as it appears at its own Web site or in text only, without graphics. If you opt for the CyberAlert full-text format, your keywords are highlighted in boldface and underlined, making scanning fast and easy. Clicking on a highlighted keyword will jump the screen to the next keyword, if any. The fact that the full-text format eliminates graphics has pros and cons. Sometimes looking at graphics is important, but CyberAlert offers both options.

CyberAlert Adds Even More Value
Apart from the usefulness of using a clip database on an “as needed” basis, another crucial reason to use CyberAlert involves the practice by which many Web news sites that update their content at least once daily employ: The same URL is often reused for current news stories. For that matter, some news sites change the news at their site every 2-3 hours. So if you don’t look at a clip as soon as alerted to it by e-mail, often impractical, the content may change by the time you get there. Since CyberAlert saves the full text of every item clipped at its Web site in your folder, you have access to any item at any time, even after the Web site has changed.
 

What of Archiving Web Clips — And Copyright?
I asked Blumstein about copyright concerns in relation to archiving Web information. For now, CyberAlert will continue to operate as it has been, since its attorneys take the view that the copyright issue on the Web is a gray area and that CyberAlert is working within legal parameters. Blumstein also reiterated that the company does not pull information into its own site from subscriber- or fee-based services or those with “robot.txt” files — and that its full-text clips do not copy whole sites or graphics, just textual information. According to Blumstein, when one sees graphics-laden pages on the service, the system has moved to the original Web site, though the “framing” makes it feel like you’ve never left CyberAlert. However, how much protection all this gives from copyright strictures, we still don’t know.
 

CyberAlert Is for Regular and Ongoing Use
Blumstein emphasized that companies should use CyberAlert as an ongoing business intelligence tool. Nevertheless, although some competing services only offer annual contracts, CyberAlert offers more flexibility, allowing a 1-month minimum and month-to-month accounts. Interestingly, while I was in the process of testing CyberAlert, one of my clients, a PR firm, wanted a media search conducted for any mention of a high-tech event sponsored by one of its clients, including a search of some of the Web. I used the usual commercial database vendors for current business, trade, and news sources, Newsbot [http://www.newsbot.com], Excite [http://www. excite.com], and a handful of specialized high-tech Web sites such as CNET.com [http://www.cnet. com], which can pipe a search through to other Tech Sites all at once — including PC World, TechWeb, ZDNet, and IDG.net.

I must admit I wished I had the powerful search technology supplied by CyberAlert, but since my client only wanted me to check this twice within 1 week, I had no recourse but to check manually. Nor could regular Web search engines help because, for the most part, spidering technology and index software can take months to include sites and can prove useless for tracking the most current information. The next best thing for this type of research, in addition to the sources I mention above, is BullsEye 2 Pro [http://www.intelliseek.com], or some equivalent, which contains close to 200 news sources last time I checked. My new, updated version is on its way, but these solutions don’t seem to come close to the proprietary Web monitoring services offered by CyberAlert and others.
 

eWatch
Now let’s move on to eWatch [http://www.ewatch.com], a service recently acquired by PR Newswire. eWatch provides clips from “thousands” of Web publications that are continually added, as well as over 63,000 Usenet groups, Electronic Mailing Lists, and Online Service Forums. eWatch monitors “hundreds” of public discussion areas on AOL and CompuServe, as well as Web Bulletin Boards such as finance/investor bulletin boards on Yahoo!, Motley Fool, and Silicon Investor.

The WebWatch service alerts you when changes appear on sites you have selected, line-by-line. This is similar to two of the free services mentioned in my previous column, JaveElink [http://www.javelink.com] and NetMind (Mind-it) [http://www.netmind.com]. Unlike CyberAlert, that won’t publicly reveal the sources it monitors, eWatch generously provides a list of Web publications at its public Web site for viewing.

Similar to CyberAlert, eWatch provides a daily alert and Web access, but offers fewer features and options by comparison. The e-mail alerts do not hotlink to the URL cited. One must log onto to a specified Web site provided by eWatch to view the whole item. Once logged on, your results list is immediately available [see Figure 4]. Unlike CyberAlert, which saves your clips folder indefinitely, eWatch saves results for only 2 weeks.

Although the e-mail instructions I received to assist me with my test described a Search capability, I could not locate it on the Web site. When I spoke to a customer service representative she explained that the search function is being revised at this time and is currently unavailable. The search capability is expected to be better than ever when it returns, perhaps a few months down the road. Although I couldn’t test the search capability at this time, it is described as allowing you to search:

• Any of the words: Articles that contain any of the words in the search window will be pulled from past reports.
• All of the words: Articles that contain all of the words within the text will be retrieved from past reports.
• Boolean logic: Changes the search into a Boolean expression.
• Exact Phrase: Will pull articles from past reports that contain the exact phrase entered.

• Tools that changes the report into a text file. This facilitates copying eWatch reports into word-processing software for use in reports, memos, pressreleases, and other documents. (eWatch instructions do point out that you may need to acquire redistribution rights.)
• Human experts who cull through the Usenet results to sift out hits relevant to the user.

So Who Wins?
Overall, I found CyberAlert more useful than eWatch, based on the cumulative database approach it took and the greater number of options for working with the information. eWatch, however, does offer access to some services, the online forums on AOL and CompuServe, for example, not provided by CyberAlert, and these may be important to some clients. My test could not analyze the two systems in depth, but my impression from a quick view of the daily results is that these systems looked very different from each other. Part of this could stem from the different way that each set up searches initially or to the different sources each covers.

Sounds like traditional online, right? Put five vendors in a room and you get both unique and duplicate results from each.

My criticisms are minor. Neither provides a copy of the search terms or strategy I requested, either at the Web site or on the e-mail alert. Since I have so much on my mind, it would be a nice added feature to include this and refresh my memory. It could help with incentive for making changes, either by reformulating the current strategy (the techs do this for you) or by asking them to add or delete terms. And, the same old problem of Internet reliability reared its ugly head. Several of my attempts to use CyberAlert during my testing phase went unmet when a response was returned that the server was down or unavailable.
 

Cyveillance
As we went to press last issue, I learned about another service that sounded interesting. When I first contacted the company, Cyveillance [http://www.cyveillance.com], representatives weren’t too sure that they wanted to talk to me because they consider themselves very different from the other Web monitoring companies I’ve been describing. Cyveillance calls itself “the pioneer and leader in the field of “Extra-Site e-Business Intelligence” and explains that its goal is to:

…help top corporations understand how activity occurring on sites across the Internet — outside of their own Web site — is impacting their e-Business success. We provide expert strategic analysis backed up by critical market feedback direct from the Net. The underpinning of Cyveillance’s e-Business Intelligence Services is its patent-pending NetSapien Technology, industrial-strength software that functions like a human…but moves at the speed of the Net…. Simply put, [NetSapien] searches like a human would, actually looking inside Web sites, pages, and newsgroup postings to provide in-depth information about the intent, content, and meaning of each site, page, and posting. NetSapien Technology identifies a variety of client-related content, including text, logos, graphics, music, and video on sites that are potentially diverting eyeballs and revenue. It also drills down into the site to identify pertinent references in meta-tags, links, and hidden/visible text and other tactics used by rogue sites to leverage popular brands and products.

Once the relevant criteria are set and programmed, NetSapien Technology continuously scours the Internet, pulling down raw data, filtering it through the client-driven criteria, and prioritizing the information to provide the most relevant market feedback. Once issues and objectives are defined, with the help of our e-Business Strategists, the technology extracts increasingly relevant, actionable performance metrics. Clients access their prioritized market feedback and performance metrics through our proprietary tools via our Virtual e-Business Strategy Center on the Web, through our secure, online interface. Client metrics are refreshed on a regular basis and prioritized based on potential business impact to the client.

Diane Perlman, director of Marketing, underscored how Cyveillance helps companies to know what’s happening on the rest of the Internet that could impact their business and to identify opportunities and risks. “There may be tactics, for example, that divert traffic from your site to another or black market or gray market types who are selling your products illegally,” she said. With Cyveillance, companies can guard against illegal or unlicensed sales by other firms, for example.

Cyveillance business strategists are recruited from the “Big Four” consulting firms where they have worked on eBusiness strategies and understand both eBusiness and traditional business and technology. Unlike a consulting firm, the Cyveillance recommendations and strategies are leveraged from the Internet through the NetSapien technology. Due to shortage of time, Cyveillance would not give me a demonstration account, but Perlman explained that a client can access NetSapien, review results, and manipulate the information with tools available for viewing and archiving the information retrieved. This sounds very similar to both CyberAlert and eWatch functionality. Similar to CyberAlert, too, Cyveillance archives information collected for clients for the life of an account. Pricing is based on an annual contract and begins at $80,000.
 

The Last Last Word
The three companies described here all provide Web intelligence, each with various features and value-added solutions. For those who need Web research occasionally on an ad hoc basis, or for a shorter period that 1 month, professional searchers might supply an effective alternative. For example, independent information professionals, many of whom belong to the Association of Independent Information Professional (me included), can provide some of these services, though usually minus the advantage of higher-level automated technologies. Although searchers usually use manual tracking tools that require them to go from site to site or meta-site to meta-site, they can conduct searches for current information on an “as needed” basis without a subscription. Moreover, experienced searchers can track not only the Web, but also commercial and other sources, and they can provide a deliverable tailored to each request, either as a compilation or with analysis or summaries. One example from my own firm involved tracking five brand names for the CEO and corporate communications department of a Fortune 500 company on a daily basis for 4 months. During that period of time, two events occurred that involved crisis control and substantial public relations effort — a trademark infringement situation and criticism of the company’s manufacturing process.

Should you decide to employ someone for ongoing tracking and monitoring, I suggest you consider one of these Web monitoring tools for mining the Web and leave the more complex analytical work to an information professional. Indeed, information professionals are continuing to grow into higher-level positions as we enter the 21st century and to bring greater value to organizations than ever before by contributing corporate intelligence, not just data.

Amelia Kassel is the president of MarketingBASE Associates. Her e-mail address is amelia@marketingbase.com