We’re In The Money…

An ‘attention’-getting view of online business information websites.

Submitted by Robert Press

For

Prof. Shelley Evenson

Design Seminar 51-702B

May 8, 1998

Crisp description: Using the tenets of attention, so-called dynamically interactive business information websites are revealed to be not so dynamic. They fail to provide, gather or track individualized, real information that is relevant to a potential customer. In addition, they fail structurally to address many of the issues that can improve a site’s attention. I propose a dialogically based system whereby user inputs are tracked and correlated with a sophisticated filtering program that responds to user’s financial goals.

The world of money management is vast and continually expanding. In an increasingly uncertain, increasingly complex world, people want to feel that their money not only appreciates, but also actually works for them. Thus, people have become more interested in knowing specifically how their money works. At the same time, however, the products and services available to people into which they can deposit their savings have exponentially multiplied. This is especially true in the world of stock portfolios. And it is even more true in the virtual world of online stock information services. Almost borne overnight, the websites of the online stock information world, (which include online trading services, stock picking information sites, and opinionated business news sites, as well as traditional business news content provider sites such as CNN and The Wall Street Journal), contain a myriad of hints, picks, news, tables and links to every possible conceivable bit of information about publicly traded companies. All these sites purport to help you, the "independent investor" (read: man who can’t afford a broker), access the information required to make logical, even advantageous, decisions about how to manage your money and in what stocks to invest. However, what these sites actually have in common is a confusing tangle of cross-links to otherwise unfiltered encylclopaedic, arcane financial information which often have no bearing on the real reasons on why a stock’s price moves a particular way. Yet these sites bill themselves as dynamically interactive.

How do these sites work? What do they actually tell you, and what don’t they tell you, or are not willing to tell you? What is the line between publicly available information and information for which it is suddenly appropriate to pay? This paper will conduct a survey of a variety of dynamic interfaces of many popular internet online trading and business information sites. I will try to extract a set of features which are unique to these interfaces. As will be shown below, these features center about an idea of the "public face" of a company: its financials, analyst recommendations, earnings forecasts. Yet the sites also have a common set of features which they don’t talk about, such as rumors or industry fashion trends. From these sets of features, I will try to enumerate how these sites generally fail on their own terms to be dynamically active regarding the language of "attention", as described in class. Finally, I will attempt to propose a set of features which could more definitively create a positive user experience in these sites. Ironically, it is not that these sites don’t dynamically give you information to make a stock pick, it’s that they often give you information that doesn’t matter, combined with an almost total ignorance of what your specific investing situation is.

First, a few comments about assumptions. The appendix lists all of the websites which I visited for data for this paper. I have chosen to concentrate on stock-picking websites, not only because I am more financially familiar with the stock market than, say, the bond or commodities market, but also because there are a sufficient number of sites to make some general conclusions. However, ‘managing one’s money’, the meta-subject to which I’m really referring here, does not merely include stocks, but also mutual funds, CD’s and numerous other investment instruments. Any dynamic website that claims to help you pick a stock is really claiming to help you manage your money. The conclusions drawn in this paper can be readily applied to any larger set of investment instrument websites because the nature of the interaction will remain the same. Only the object mediating the interaction will have changed.

That being said, what do these sites have in common? On a content level, they strangely share almost identical information, which I refer to as the "public interface" to the Company. Whether it is a online trading site, such as Quick & Reilly’s QuickwayNet, a website of a traditional business media outlet, such as CNN.com or CNBC.com, or a subscription-based information network, such as the Business Channel of AOL, each site will allow you to enter a stock ticker symbol (or help you look it up if you don’t know it) and then give you some current information. This information is the same across all sites: the last traded price, the change from yesterday’s close, the daily trading range, the current dividend and price/earnings (P/E) ratio, the volume of shares traded and the stock’s 52-week high and low. I refer to this as an ‘interface’ because it acts much like an avatar to the stock: a public image of what is currently happening, yet not very correlative about the in-depth character of the issue.

These trading sites then allow you, through various links to other sites, to ferret out the ‘second level’ public face of the stock. This would include detailed financial filings of the company with the SEC, any news articles which have appeared on the business news wires (such as Reuters or PR Newswire), and most importantly, analyst recommendations. The links may take many forms, depending on the proprietary nature and complexity of the information. Generally, however, this information is handled by a few clearing sites, which are hyperlinked to the website in question. For example, SEC filings in themselves could take up whole libraries, and thus are not appropriate to be maintained on The Wall Street Journal.com. Therefore, to find out about filings, the Journal’s webpage sends you to the SEC’s database. Interestingly, so do all the other websites. Another example of this is with company research reports. The major brokerage houses keep voluminous amounts of information on each stock, including forecasts for future performance, and this information changes almost daily. Thus, to track this information within one business website would also be untenable, and so a few separate websites have evolved from the traditional companies that tracked this ‘consensus’ pre-internet. The companies handling this include First Call, and Zack’s Research. Almost every stock trading site and business news site (including subscription sites like AOL) eventually send you to these research sites if your information request gets too detailed. As will be explained below, however, these ‘research’ sites only provide a finer grained interface. They may give forecasts, but almost never say how these forecasts were reached and, more importantly, which analysts made the decision. Thus, a crucial "fashion" element is excluded from these reports, which often have a major contribution to stock price movement. Stock prices often move because the company’s industry suddenly becomes in vogue, with no real correlation to traditional earnings results.

Each site also provides access to some current and catalogued news stories on the companies. These news stories, again, however, tend to be generalized summaries straight from the news wires and thus come ultimately from a handful of sources (either CNN, CNBC/WSJ, Reuters or Bloomberg). Thus, by reading the news from one site, one can often access all the news on one company that one could read by canvassing every other site.

On the deepest level, one reaches the "opinion sites." These are locations, such as AOL’s Motley Fool or The Street.com, that purport to make a qualitative analysis of news regarding stocks, and then give an opinionated "pick" of what to do. These sites uniformly charge for their services. However, on closer inspection, one notes that the sites are very unidirectional. They lecture at you as to what stocks to pick and do not take into account user needs. A prime example is The Motley Fool, touted as the most visited business news opinion site on the web. This site maintains a ‘Fool’s Index’ of stocks to watch, which were chosen on criteria that the site deemed appropriate. Generally, these picks have succeeded on monetary terms, but it is interesting to note that these stocks may not be right for every person. In particular, Motley Fool subscribes to a "Dogs of the Dow" theory, whereby one buys the three lowest performing stocks in the Dow Jones Index and waits for the eventual turnaround in 6-12 months. This strategy may be too risky, or not risky enough, for a large number of investors, which makes Motley Fool picks useless. Additionally, there is no available input channel, where you could, say, ask The Motley Fool what it thinks of a particular stock. If they don’t talk about it, then you’re out of luck. A similar problem exists with their reporting of the Wall Street rumor mill. They choose the news that is important and often don’t say how or why they reached a particular conclusion. Then the authors of these sites, like The Street.com’s James Cramer, will appear on other sites like CNBC, as a commentator to create a fashion trend surrounding his opinions. Thus, these sites ultimately function much like gurus or delphic oracles: they make pronouncements, and rely on their own sheer publicity or shmoozing power with other media outlets to lend credibility to their choices.

Thus, the net result of these websites, from a content point of view, is that there is a fixed body of public information out there which can be accessed from almost any outlet. Each business website acts mainly as a ‘branding’ tool of the information, so that one may like Microsoft Investor’s look and feel of the news better than CNN’s. However, at the end of the day, you haven’t learned anything different at one site than from another. And more importantly, you haven’t necessarily learned the information that you need to make a wise investment decision. This occurs because these sites lack two crucial pieces of information: the unspoken ‘fashion trends’ surrounding the company, and, most importantly, what goals you as investor have regarding your money. Since these sites have not tried to find out anything about your goals, they have very little they can offer you in terms of insight beyond the "public interface" to the stock.

An analysis of these sites from the perspective of "attention", as defined in class, sheds more light on these deficits. Within an interaction language, ‘attention’ describes a multidimensional reflection of user interaction needs. These needs have been defined as connection, attraction, orientation, appropriate experience, extension, retention and social reputation. If each dimension in the experience were accounted for and addressed, the user would have a more fulfilling experience with the site, and presumably, enter into a more symbiotic relationship with it. Underlying attention is the assumption of a dialogue. If the site is not adapting to the user’s needs, let alone trying to find out about them, the user has little reason to continue interacting with the system. The system becomes a monologue and ultimately static. It is remarkable how most business websites fail on these terms.

The sites, on the surface, appear to be dynamically interactive, because they obey rules of connection, attraction and retention. Clearly the market is open and businesses do business every weekday. Thus, information on stock prices and company performance is always in a state of flux and most business websites capitalize on that fact. One has to check the sites precisely because the information, by definition, is out of date as soon as it is gathered. Also, the sites are easily accessible on the web, relying on branded names for easy recognition. And once one has reached the site, one finds oneself absorbed in the information much like a visit to the library: its volume creates awe (yet also confusion). The stock market clearly has an addictive aspect to it and this addictive substance drives much of the attractive and retentive nature of these sites. ‘I just have to see what the price of my stock is right now,’ people say, or, ‘it’s so easy to update my screen and see what the market is doing’. People often minimize their stock tracking sites and check them throughout the day. They can even be aurally notified by the site of any breaking news on the company. This repeat engagement ironically fosters retention, but not loyalty, because the information can be had from any source. Indeed, I worked once at a law firm that blocked Quicken.com from its Netscape browser, yet I could get the same information from CNBC.com. Finally, because the stock prices have a correlation to actual money, the sites satisfy the attraction requirement of ‘worthwhile’ and ‘curiosity.’

As for the other qualities of attention, however, the sites almost uniformly fail. The encyclopaedic nature of the information, with its numerous cross-links to governmental archives and independent rating agencies violates the idea of orientation. It is almost impossible to gain one’s bearings in these sites after any substantial time in them. For example, to research a particular company on AOL’s Business Channel required me to have no less than 10 screens open simultaneously, in order to have all the cross-references regarding quarterly results, analyst recommendations, SEC filings, news of the day and historical performance. Since any one sub-site is cross listed with all the others, there is an exponential web of cross-links which is immediately stultifying and ultimately off-putting. And God forbid I accidentally closed a screen because I could often not find its content again. Thus, one may initially feel one can explore what’s possible; yet scratching the surface only reveals a bottomless chasm. After an initial search, one often gets tired, does not know where one has ultimately ended up, and does not know if one has reached the adequate end of the search. There is no mapping within these sites.

The sites also fail because they do not engage in a dialogue and do not really question the user of his/her investment goals or needs. Thus the sites fail in the realms of appropriate experience and extension. The sites do not manage the interaction at all. The user is always prompted to ask the question: "tell me all you know about X. The site then responds, but with no filters and with no analysis. Thus user frustration is poorly handled. For example, a stock may be having a particularly bad run over the past few weeks and all the research in the world often cannot explain why. The reasons often tend to be fashion considerations or it concerns more closely held information. The sites need to prompt the user for more information about himself: what his financial situation is, what his investment goals are, what level of risk he feels comfortable with. Similarly, the sites need to address user questions. Nowhere have I ever seen even a simple online dictionary to business terms in these sites. I have not encountered a simple financial planner within these sites. By defaulting over managing the experience, the user tends to feel lost and meaningless within these sites.

The lack of appropriate experience generates a corresponding lack of extension. The user is denied a sustained participation in the sites because the site really does not answer the pertinent question of what to do next. Similarly, there can be no rising expectation by the user of the site because the site is not truly probing the user’s needs or synthesizing conclusions. This explains why these sites tend to look all the same after a while. If the sites could engage the user directly on what he was looking for, then the user could say "I need to go to site X because it is telling me something I need to hear." With this process absent, the user has no expectations of the site beyond that of reportage. And with no feedback mechanism, the site cannot experience the user’s frustration. The structural exception to this comes in the "opinion" subscription sites, like The Motley Fool. However, at those sites the delphic nature of their pronouncements often leads to the same effect. In practice, one checks The Motley Fool each day to see what it has to say. The Fool’s daily pronouncements may or may not affect the market. Yet it may only tangentially affect the user’s situation if the site happens to talk about the user’s stocks or those stock’s industries. Thus in practice, the sites contain no ability to leverage capability. One cannot ultimately make a decision on whether to buy a stock given these site’s processes, because the site does not know enough about the user to make that kind of decision.

Finally, the sites fail to promote a social reputation. This follows logically from the lack of experience and extension. The sites do not learn about the user over time, past what the user tells it to remember ("send me news about company Z"). Thus, the user forms no bond with the sites as an investment advisor in the way one does with a live broker. The user is then not motivated to recommend to others one business website over another. The sites then do not have any sustained activation growth beyond its own generated publicity.

Hence upon closer investigation, these seemingly dynamically interactive sites generally fail to achieve the increasing returns benefit of attention, precisely because these sites were not really dynamic to begin with. The question then remains what are the definitive set of features required to make for a positive user experience of these sites. Clearly the sites serve a function or they would not be visited; however, we seek solutions that make the sites dynamically adapt and design themselves to the user’s criteria, thus enriching the site and user.

I believe a potential solution lies in looking at the traditional relationships that these sites seek to mimic: that of the traditional broker. Nowadays reserved for those who can afford brokerage companies’ large trading commissions, a personal stockbroker not only gives suggestions about what to buy, but more importantly, makes those suggestions after gathering knowledge about the financial needs and imperatives of the customer. The broker then refines this knowledge through an ongoing trading relationship with the customer. The broker can then see what makes the customer antsy and what the customer likes. He then tailors his advice accordingly, and as an effect, spends most of his time not telling the customer what he does not need to hear. A broker has access to volumes of information, which he could relate to a customer, but the broker simply knows better. He knows when information is relevant to the customer and when it isn’t (in an idealized world). An experienced broker also has access to and familiarity with the unspoken market trends of fashionability and rumors. A broker can tell you when to listen to an analyst recommendation because that analyst may or may not command respect within the financial community. Similarly, he is familiar with the herd mentalities that characterize the local buying and selling trends of large-scale markets. Interestingly, all this information could now be organized within a complex filtering program, so the answer to why this process is not electronically implemented must revolve around a lack of attention to user input.

Thus I propose a solution to these stock websites that add two characteristics. First and foremost, the sites would require a detailed history regarding users’ needs. What are customers’ long and short term financial plans? What are comfortable levels of risk? How much money is available to invest? What is the expected rate of return? What industries are they most interested in supporting? These questions are not one-time inputs. They would have to be continually updated. However, I believe this would be easily achieved adding a second characteristic to these sites. The sites must keep track of where the user has gone and then correlate that path (and the information drawn in that path) with the user’s histories. This process actually describes a large, reflexive filtering device which basically tracks user input much like a neural net. Any action for new information would trigger an analysis of previous information and a synthesized reaction, so that the system grows with each user move. On the backend of this filter engine would be inputs from much the same sources available to human brokers: newswires, analyst recommendations, etc. However, the filtering engine would be able to reason about the inputs and make quantitative decisions regarding those inputs based on user histories.

An example could go as follows: after initially describing one’s financial goals to a stock picking website, the site would make available to you the information paths most readily needed at that moment. If you have stocks in mind, it would give you the information and a recommendation about those stocks. You would then indicate if you thought it was a good idea. If you instead wanted to research these stocks more, it would track your path through the research fields and, based on the information contained in those fields, ask you if you were still interested in the stock. Do you still like Phillip Morris after you read the news about the impending costs of National Tobacco Legislation? The system would then update its opinions about you and Phillip Morris, perhaps telling you about a positive analyst recommendation if you were still interested, or recommending another industry with similar profit potential if you were not. What is new about this system over the old websites is that it is centered about user’s opinions. Information is no longer transmitted monologically, but rather in dialogue.

This yields the benefits of attention in many ways. First of all, the site is now adequately mapped. Because the system knows the history of your inputs, it can tell where you’re going and how you may more easily get there. It can locate you if you’re lost and tell you where, if anywhere, there is left to go if you want more opinions. Similarly, there is now a more appropriate experience because the system manages your interaction as you have it. It could continually ask you whether certain screens should be remembered or whether some information was relevant to you. It would then update its database about you and make its next move accordingly. This would solve the endless circles of having to continually look up the research reports of companies because the system would know you are really only interested in certain things in those reports. Additionally, the managed interaction would foster a rising expectation in the customer, much like how a person feels when he calls his broker: "does he have something for me today?" Push technology can also reverse this so your system could also call you with alerts, much like real brokers do. Finally, a system that managed its customer’s interaction would naturally create its own social reputation which would be spread to other customers. If a particular website’s filtering algorithm was particularly effective, it would certainly be a cause célèbre. Because this structure is by definition opinionated, it would probably be allocated on a subscription basis, much like the delphic websites. However, one is now getting more relevant information for one’s money because it is tailored to your needs and not the pronouncements of some guru.

Thus, by using the tenets of attention, so-called dynamically interactive business information websites are revealed to be not so dynamic. They fail to provide, gather or track individualized, real information that is relevant to a potential customer. In addition, they fail structurally to address many of the issues that can improve a site’s attention, such as orientation or extension. This is done, however, in an environment that mimics a dynamic activity because it tangentially addresses issues of connection and attraction. However, I propose that unless these sites reorient themselves to identifying and tracking unique user actions, they are destined to become just one more tome on the encyclopaedic internet information heap, often confusing the issues rather providing clarity.

America Online Business Channel

The Motley Fool

CNBC.com

TheWallStreetJournal.com

Morningstar.net

Quick&Reilly.com

MyDiscountBroker.com

Microsoft Investor

CNN.com

E*Trade.com

Brown&Co.com

TheNewYorkTimes.com

Quicken.com

Standard&Poors.com

MerrillLynch.com

Dljdirect.com

Reuters.com

PRNewswire.com