In mid-1994 Jerry Yang and David Filo were spending inordinate amounts of time surfing the emerging Web. Part of this surfing involved building a list of hyperlinks to Internet sites which they classified by subject. The classification system was a traditional hierarchical one that reflected the personalities and interests of the compilers. The system carried a number of different names, ranging from ‘Jerry’s Fast Track to Mosaic’ to ‘Jerry and Dave’s Guide to the World Wide Web’. Eventually they settled on the name ‘Yahoo!’, which was subsequently underpinned by the somewhat contorted explanation: ‘Yet Another Hierarchical Officious Oracle’. As the classification system grew, it quickly became apparent that simply being presented with all the classifications left too laborious a task for users to sift and find what they were interested in. To simplify the process, Yang and Filo added keyword software that allowed users to search for keywords contained in hyperlinks. Suddenly there was a powerful tool for finding information on the burgeoning Web. Yang and Filo’s website was receiving more than 100,000 page views a day in 1994.¹⁰⁰ Simultaneously, the venture capital industry was beginning to wake up to the possibilities of the Internet and some interest was expressed in the work being conducted by Yang and Filo. Companies such as Netscape and AOL also initiated contact with the pair and each made an offer of $1m for the site.
10.15 – Yahoo IPO: climbing a wall of worry
Source: Montage – sources in art itself.
Netscape’s interest was sufficiently strong that in 1995 Marc Andreessen convinced Yang and Filo to move their site to be housed on Netscape equipment. Eventually the pair faced a decision of whether to accept one of the acquisition offers or to take the funding being offered by Sequoia Capital. Much to Andreessen’s chagrin, Sequoia’s offer was accepted. As a consequence, Yahoo was ejected from Netscape’s premises. Netscape did, however, retain the link to the Yahoo site and it was to be some time before Netscape woke up to the fact that this link itself had a value as it directed customers to Yahoo free of charge. The initial valuation implied in Sequoia’s funding was $4m.
Subsequent funds were raised from selling a limited amount of shares to Softbank of Japan, and by this time Yahoo had appointed professional management in the form of Tim Koogle and been successful in raising advertising revenues. In April 1996 the stock went public, priced at $13, raising just over $30m. Although the stock moved sharply upwards after the IPO, within a relatively short period it had returned to the IPO price, earning it the sobriquet ‘Yet Another Highly Overpriced Offering’ from the press. This was to prove the bottom. From there, the share price moved not just upwards but out of sight. The question for the future was the extent to which the company’s proprietary information, extensive classification system and management ability would be able to act as a barrier to entry and protect its advertising revenue stream.
Yahoo
The financial history of Yahoo was so short at the time of this book’s previous edition in 2001 that it was difficult to say then whether it displayed the same financial characteristics as earlier technology companies. Companies such as Marconi waited a long time for profits to appear, but when they did, they rose sharply – reflecting the impact of rapid growth in usage on an asset base the cost of which was largely sunk. Unlike Marconi, Yahoo never had large sunk costs, meaning that its profit ratios looked exceptional from an early stage. The downside was that without sunk costs, barriers to entry were low, making the preservation and strengthening of the brand the most vital competitive weapon that the company could deploy. It was obvious that, as with all brand names, substantial expenditure would be needed to maintain its value, since at some point well-funded established companies would undoubtedly renew their assault. Given that it had a strong balance sheet, Yahoo was never at risk of losing its position overnight, and certainly not as a result of the kind of cash flow pressures that many other early movers in a new technology had to face.
For many analysts, being cash-generative meant that Yahoo could be valued with the help of discounted cash flow techniques. Since the valuation looked stretched on that basis, in order to justify the share price it was necessary to push out the time horizon over which cash flows were discounted. Discounting cash flow over 15 years, however unrealistic, was not uncommon. The problem was that with a market capitalisation in excess of $100bn at its peak, and sales revenues of less than $1bn, no fundamental valuation technique at the time could arrive at a positive answer. This is not to say that Yahoo was an unsuccessful or poorly managed company. The opposite was the case. It was, after all, one of the few companies which at the height of the bubble produced both positive cash flow and profits.
10.16 – Yahoo: one-time race leader later lapped
Source: Thomson Reuters Datastream. Yahoo annual reports.
No doubt in part for that very reason, from 2002 to 2005 the share price of Yahoo rose by 300%, but that was as far as the company could take its market valuation. Since then the history has been one of struggle to compete against the new competitor Google. The latter was able to take advantage of the absence of financial barriers to entry that existed while Web search was still nascent. This rapidly eroded any hopes Yahoo might have had of sustaining a market leadership position and the accompanying protection that scale would have provided. Many opinions have been offered to explain why Google was able to supplant Yahoo in the search engine space. For some it is about the systems architecture of Yahoo, which powered its early rapid expansion but has since been too inflexible to allow the company to adapt as quickly. For others it is about the brand management. The most persuasive evidence suggests that Yahoo lost ground by paying insufficient attention to the development of search – and search advertising – technology, and then proved unable to create, or buy in and integrate, a competitive offering.
It was not that management failed to perceive the threat, as the purchases of Inktomi (search) and Overture (search advertising) demonstrated. However, the acquisitions took too long to integrate and the technology took longer to upgrade than was necessary to maintain parity with Google. A series of changes in the management were made, but each iteration proved unable to resuscitate the company, and in 2017 the Internet businesses were acquired by Verizon, with the rump of the business, including its valuable shareholdings in Alibaba and Yahoo Japan, being retained by the successor company Altaba. On a like-for-like basis the value of Yahoo’s search business had diminished by more than 95% from its bubble peak, albeit the value of the two minority investments mitigated the decline significantly.
Unfortunately for the company, the view that it was protected by barriers to entry proved to be misplaced. Faced with a new and fierce competitor in the shape of Google, Yahoo slowly but surely lost its market leadership to a more determined rival. Depending upon their background, commentators have ascribed the decline of Yahoo to a number of different factors. For some it was down to branding, with a view that “Yahoo has lacked a definitive brand purpose”, while “Google has mastered brand strategy and management”.¹⁰¹ For others it was about the design of infrastructure: Yahoo rapidly increased capacity through adding NetApp storage devices, whereas Google developed its Google File System using “commodity servers to support a flexible and resilient architecture that could solve scalability and accelerating the future rollout of a wide range of web-scale applications, from maps to cloud storage”.¹⁰²
The most likely explanation is also the simplest. From the start, Google was entirely focused on creating the fastest and most accurate search engine. Yahoo was not. As a consequence, Google was able to leapfrog Yahoo to such an extent that Yahoo sought to purchase its rival, only to baulk at the potential $5bn price. Instead it decided to buy in search and search advertising technology. In abstract, the purchase of search engine company Inktomi (for $257m) and the search advertising company Overture (for $1.4bn) might well have seemed the more sensible alternative. However, the process of integrating these two acquisitions proved troublesome, riven by internal
politics and compromised by external customer relations, most notably those with Microsoft, now awake to the potential power of search and an important client of Overture.
Google – so much for first-mover advantage!
Given the history of innovation and the competitive nature of emergent technologies it is not a surprise that ‘search’ followed the same pattern as many previous innovations, with a process of fierce competitive struggle ending with the emergence of a dominant player. Archie, Veronica, Jughead, the World Wide Wanderer, Aliweb, Excite, Galaxy, Dmoz, LookSmart, WebCrawler, Lycos, Infoseek, Inktomi, Ask Jeeves, alltheweb and Altavista were just some of the directory or search-related applications or companies that emerged over the period. Eventually they failed, were subsumed within competitors, or became specialist niche players in a world dominated by Google.
10.17 – Search not hard to find
The history of Google is almost a caricature of how a business is created in Silicon Valley. It was conceived in academia, nurtured in Stanford, given initial shape in a garage, early-stage funded with private money, turned down by possible partners and/or acquirers and then later funded by experienced tech venture capitalists before it had found a clear revenue model. Google’s fast and seemingly irresistible rise to dominance in search has been chronicled in many studies, but although well known, some of the key elements are worth highlighting in the context of this book. The story begins with Sergey Brin and Larry Page, two graduate students in computer science at Stanford. Both had backgrounds in mathematics and computer science and both had academic parents in the sciences. A Stanford facility, ironically funded by Bill Gates, became the early home for the work that was to form Google. It was an environment specifically designed to encourage innovation and collaborative thought, and the initial driving focus, as far as Brin and Page were concerned, was intellectual rather than commercial.
For the two graduate students, the roadmap to a PhD required original research, with the starting point being the particular task or theory to be addressed. Given the emergence of the Internet and the growth of information residing on it, designing a tool to find the defined information represented both a logical topic for study and one that might have commercial potential. It was also an area in which numerous academics had dabbled, if only to aid them in their own research. The most striking element of the Google story is that the founders’ entire focus in the early stages was not on the business opportunity, but rather the challenge of creating the fastest search engine possible. Design was therefore uniformly directed to this end and not subverted or undermined by incorporating features to enhance revenue. Indeed, as far as one can see from the outside, potential revenue formed little part of the early decision-making process.
The genesis of how Google searches the Internet lies in the roots of the academic system – in the frequently used epithet ‘publish or perish’. Advancement in academia is heavily driven by the output of papers in refereed academic journals. As importantly, there is a hierarchy of journals (quality) which is accompanied by the number of citations received (volume) from other academic papers. The more prestigious the journal and the more citations received, the more important an academic publication is deemed to be. The peak of this particular mountain is reached when a publication comes to be described as ‘seminal’. The point is that adding up and weighting the number of citations has always been viewed in academia as independent measure of the worth of a publication.
This was precisely the methodology adopted by Google for search, with one important adjustment. As Brin and Page pointed out in their paper,¹⁰³ “unlike academic papers which are scrupulously reviewed, Web pages proliferate free of quality control or publishing costs”. Moreover, since the Web is a commercial venture, participants are incentivised to dupe search engines. A substitute method was therefore required to replace the quality control of academic refereeing. Like all true insights, the answer was straightforward in principle: volume of citations weighted by quality of citation. The proxy for ‘quality of citation’ was taken to be the number of links to the particular website. As an example, Brin and Page noted that because the Yahoo website had so many backlinks, the implication was that it was important and therefore should carry a higher weight. Reflecting Larry Page’s insight, it was named the ‘PageRank’.
This is not to say that translating this into practical use was easy. Web content was growing exponentially and, like all true-life situations, the data would be messy and incomplete, requiring all sorts of innovative solutions and workarounds. On top of this, downloading the entire Web so that it could be subjected to their Web crawling and indexing system was a non-trivial task. The volume of data this required was what led Page and Brin to choose the name ‘Google’, a misspelling of the mathematical term ‘googol’, meaning one followed by 100 zeroes. Nevertheless, as their published work revealed, even in the early stages the Google search engine, was distinguished by both its accuracy and its speed. The comparison between Google and Altavista, another search engine, was detailed in the paper and was as striking as it was intended to be.
What then followed had historic precedent and might be termed the ‘Decca effect’, after the recording company of that name which famously turned down the opportunity to sign the Beatles. Other precedents include Western Union turning down Bell and Marconi turning down Logie Baird. Google sought to raise funding to continue development and approached the search company whose results it had analysed and found second-best, Altavista, in the hope of licensing its technology. Perhaps DEC, Altavista’s parent, was unwilling to use external technology, or perhaps its own impending merger with Compaq diverted attention, but either way Google was spurned. This was repeated by other search engines – including Yahoo, for which the Google engine would have been a natural enhancement to its own directory-based structure.
The market developed differently
One of the reasons there appeared a lack of excitement for Google’s new search engine was that the initial hopes of commercialisation appeared to revolve around licensing the technology to others. The prevailing wisdom was that revenue generation followed from retention of potential customers on your portal, rather than sending them elsewhere. For this reason the accepted advertising business model revolved around product placement and paid-for advertising. This was in direct contradiction to search, which was about comparison and price discovery. If search was not overwhelmingly important in its own right, there seemed to be no need to outsource the function to a third party. In fact, there was a ‘mini mania’ for a period, during which real estate on portals became critical and ‘stickiness’ the most valued attribute.
On numerous occasions the founders of Google pointed to the inherent conflicts of interest in existing search models, where paid-for advertising could skew the results of search towards the advertiser. What Google had in its favour was the evidence both of the superior nature of its search engine and the traction it was gaining through exponential growth in users and traffic. These two features, combined with the powerful personalities of the founders, persuaded influential investors to back the venture. Through Stanford faculty member David Chilton, they were introduced to the founder of Sun Microsystems, Andy Bechtolsheim. As only a wealthy and successful individual can, Bechtolsheim backed the venture almost instantly and this set in motion a chain of events that brought in other notables from the Valley. The company ended with a $25m joint investment from Sequoia Capital and Kleiner, Perkins, Caufield and Byers at a valuation of $100m.
The ultimate pricing model solution for Google was already in train. Elsewhere, serial entrepreneur Bill Gross understood that the mass of traffic on the Internet had to be both filtered and specific to ultimately be of value to advertisers. He further understood that specific traffic could be purchased from ad networks/online banners and made available to interested parties on a price-per-click basis. The interested parties would bid on a keyword and then be rewarded by traffic with a high degree of relevance. This wou
ld initiate the volume needed to make the model operational and then search would eventually take over the provision of specific traffic. In common with existing general advertising practices, the higher the price the advertiser was willing to pay per click, the more prominent the placement they would receive.
For Internet true believers, the direct link between search results and advertising payments was anathema. However, this did not stop Bill Gross’s site GoTo.com growing rapidly. Such was its success that it was also able to syndicate its service to other companies, most notably AOL. The success of syndication in generating traffic, and hence revenues, was such there was internal concern about potential conflicts between the internal site and the syndicated sites. As a consequence GoTo.com was renamed Overture and the internal destination site was de-emphasised.
Engines That Move Markets (2nd Ed) Page 54