Started by Gio Wiederhold,
18 Feb 2000.
"Bioinformatics: Converting Data to Knowledge"; Introduction and Closing Observations for a meeting at the National Research Council, Washington DC, February 2000.
B2B has total volume 3-5 x retail community.
Handout: E-business Busters.
The assumption is that broad dissemination of knowledge will make all customers (that use the web) equally powerful.
In the past, mainly used for unique, valuable (non-fungible) items, where common distribution mechanisms fail. The auctioneer provides links, marketing, escrow services etc., and lives of a cut of the sales price (3-10%?) Most auctions are local (for goods of modestly high value, that are unlikely to be shipped far, unless purchased by dealers). There are sume large auction houses (Christie's etc) that have specialized in very valuable items, and provide services (telephone) for remote bidders. All are entering the on-line market. One (Butterfield's) has been acquired by eBay.
Various models, but generally the highest bid at closing time gets the goods for that price. In eBay, when lots of identical are auctioned, all pay the same price, which is the lowest of the winning prices. eBay does not provide escrow -- the seller and buyer must contact each other to arrange pay and delivery, but bad (and good) experiences with sellers (and buyers?) are posted and easily available, providing some degree of protection.
The on-line model has been attractive to a much broader range of customers than traditional auctions. It has allowed auctioning of relativley small value items (Pez-=dispensers intiated eBay). It is also used by vendors of fungible items, who gain broad exposure. For about $100 a vendor can place him/herself on a weekly? list of `hot' auctions and gain yet more visibility.
The popularity of eBay is in part due that it provides a rapid and broad means for comparison shopping. Although the selling model differs, having an item of interest available on eBay and seeing the bidding provides a base indication of the market, and can be used as a base for conventional shopping expectations as well. It has been noted [] that purchase prices at auctions do not indicate low, or rational prices, since the final price is the highest one, amd may well be affected by the excitement of the game or similar emotional factors.
Income versus Cost
Cost = Operations + Interest on Investment
Bernardo A. Huberman, Xerox PARC
Topic: The Laws of
the Web.
Abstract:
The past decade has witnessed the birth and explosive growth of the
World Wide Web, both in terms of content and its user population.
>From a few thousand pages in 1994 the web has grown to several
hundred million pages today. The number of users online has been
growing exponentially as well, with the number of internet users
expected to be 320 million by the year 2000.
There is hidden order behind the apparent arbitrariness of the growth
of Web and the preferences of its users. And this hidden order can be
explained from dynamical organizing principles. One observed
pattern is that there are many small elements contained within the
web, but few large ones. A few sites consist of millions of pages, but
millions of sites only contain a handful of pages. Also, few pages
contain millions of links, but many pages have one or two. In
addition, millions of users flock to a few select sites, giving little
attention to millions of others. And within those sites users surf
according to a mathematical law, while generating predictable
congestion patterns.
Used Georgia Tech (200 000 clicks), Univ.of Michigan, Xerox, and
Yahoo (10 000 000 clicks) data.
Takes clicks as proxy for economic activity -- best one can do. Residence time
could be a better metric, but messy with people going to sleep, etc.
Distribution of Clicks per session is highly skewed: many short sessions, few
very long ones. Validates Pareto economic principle.
Mode for Yahoo was 1.5 clicks.
The usage distribution is also Zipfian (close to a a negative exponential) with exponent 1.5, matching abstract economic predictors.
Economic issue for operators of web sites: to make experience pleasant for visitors, make it few clicks to get results, but to sell, and focus, more resience and more clicks will be needed. Not addressed is how often are there few clicks because site found was completely inappropriate and users gave up.
Market share was analized for 120 000 active sites, with visits as proxy for economic activity. All vists to the same IP address were summed. Here 0.1% (120) of the sites captured 32.6% of all business. The top 1% covered about 90%. The University of Michigan was the most active academic site. Unclear why. Good links to sex sites? Poor alternative communication? many similar departments with cross traffic?
There are differences by domains. Sex-related traffic still amounts to 66%, showed the same distribution curve. (Data from others showed a lower, and lowering fraction, but this analysis was done on historical data.) Of those 40% were intiated by women, as reported by Metaspy (also conflicts with other data -- how are females identified? Many people use avatars -- sex-less. Were those subtracted first? Do more males use avatars? Do some males use female handles? )
The imbalance of usage is common, and explained through `socially-mediated search': people tell friends or create links, greatly amplifying trends towards products that are otherwise indistinguishable (Laking for birds [?]). That leming-like behavior minimizes the cost of search. Relates to importance of advertising. (Figure from another source: Now (2000) it takes $40M of advertising to establish a web presence in the consumer market.) That pattern even applies to sex sites, who are not advertised much.
A good audience question [Bill Sharp?]: How will tools like Google, who alread direct uers to the most-linked site further excerbate the `Winner takes all' tendency?.
Why do investors view Internet companies as being different. Metcalf's law: Network size has a squared advantage. Higher ROIC - low capital costs AOL growth in customers, days to add 1M: 12-13M= 133, 13-14 = 77, 14-15 48, 15-16++ 41 days.
Valuation is volatile because all positve cashflows are way out. Only 3 % of Market cap today is in 3%, while demand is about 10% in peoples portofolio. Some due to investment bankers presuures. Portofolio managers have to buy them to justify their pay, because otherwise they perform less than others, and they get no bonuses. But losses don't affect them as much. 60% all new capital is flowing into technology. Devaluation of underwriting standards. IPOs can become rapidly an orphan stock: example TheGlow. A shakeout will come.
Third circle companies that support the Internet are good bets.
E2B is hard (CommerceOne convinced GM) (but only one deal for GM) Vertical net is a mile wide, one inch deep, good selling, and no can acquire stuff. Also done by CISCO, acquires companies started by its engineers that have quit. `Dream has become a belief'
Frustrating for conservative CEOs.
International presence differs. Softbank supports incubators all over, using it to replicate US ideas. QXL knocked of EBay that way in Great Britain. Internationalizing 180-flowers was hard due to local gift preferences, perishable. Requires more corporate work, not only local partnership. [Bill Burnham].
See www.ecommerce.gov
Business not ready
See also the references.